| <?xml version="1.0" encoding="iso-8859-1"?> |
| <!DOCTYPE Meeting SYSTEM "http://www.icsi.berkeley.edu/Speech/mr/Meeting1.dtd"> |
|
|
| <Meeting Session="Bdb001" DateTimeStamp="2001-02-16-1000" Version="2" VersionDate="Jan 7 2004"> |
|
|
| <Preamble> |
| <Notes> |
| fe008 joined towards end of meeting. |
| </Notes> |
| <Channels> |
| <Channel Name="chan1" Mic="s1" Transmission="w1" Gain="8" AudioFile="chan1.sph"/> |
| <Channel Name="chan2" Mic="s1" Transmission="w1" Gain="8" AudioFile="chan2.sph"/> |
| <Channel Name="chan3" Mic="s1" Transmission="w1" Gain="8" AudioFile="chan3.sph"/> |
| <Channel Name="chan4" Mic="s1" Transmission="w1" Gain="8" AudioFile="chan4.sph"/> |
| <Channel Name="chan6" Mic="c2" Transmission="u1" Gain="8" AudioFile="chan6.sph"/> |
| <Channel Name="chan7" Mic="c2" Transmission="u1" Gain="8" AudioFile="chan7.sph"/> |
| <Channel Name="chan8" Mic="s1" Transmission="j1" Gain="11" AudioFile="chan8.sph"/> |
| <Channel Name="chanB" Mic="s1" Transmission="j1" Gain="11" AudioFile="chanB.sph"/> |
| <Channel Name="chanC" Mic="u3" Transmission="j1" Gain="12" AudioFile="chanC.sph"/> |
| <Channel Name="chanD" Mic="u3" Transmission="j1" Gain="12" AudioFile="chanD.sph"/> |
| <Channel Name="chanE" Mic="c2" Transmission="u1" Gain="8" AudioFile="chanE.sph"/> |
| <Channel Name="chanF" Mic="c2" Transmission="u1" Gain="8" AudioFile="chanF.sph"/> |
| </Channels> |
| <Participants DB="icsi1.spk"> |
| <Participant Name="mn017" Channel="chan1"/> |
| <Participant Name="me001" Channel="chan2"/> |
| <Participant Name="me011" Channel="chan3"/> |
| <Participant Name="me018" Channel="chan4"/> |
| <Participant Name="fe008" Channel="chan8"/> |
| <Participant Name="fe016" Channel="chanB"/> |
| </Participants> |
| </Preamble> |
|
|
|
|
| <Transcript StartTime="0.0" EndTime="3021.968"> |
|
|
| <Segment StartTime="0.000" EndTime="3.956" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="0.056" EndTime="1.861" Participant="me011"> |
| Yeah, we had a long discussion about |
| </Segment> |
| <Segment StartTime="2.674" EndTime="6.315" Participant="me011"> |
| how much w- how easy we want to make it for people to bleep things out. So - |
| </Segment> |
| <Segment StartTime="5.100" EndTime="7.881" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="6.260" EndTime="6.864" Participant="fe016" CloseMic="false"> |
| Right. |
| </Segment> |
| <Segment StartTime="6.315" EndTime="7.061" Participant="me011"> |
| <VocalSound Description="breath"/> <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="7.516" EndTime="8.757" Participant="fe016" CloseMic="false"> |
| O_K. So this is - |
| </Segment> |
| <Segment StartTime="8.169" EndTime="11.100" Participant="me011"> |
| Morgan wants to make it hard. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="8.320" EndTime="13.210" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="9.751" EndTime="13.549" Participant="fe016" CloseMic="false"> |
| The, uh, counter is not <Pause/> moving again. It - |
| </Segment> |
| <Segment StartTime="9.798" EndTime="11.268" Participant="mn017"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="13.210" EndTime="14.476" Participant="me018"> |
| It - it doesn't - |
| </Segment> |
| <Segment StartTime="14.457" EndTime="15.844" Participant="fe016" CloseMic="false"> |
| It doesn't like me, but - |
| </Segment> |
| <Segment StartTime="14.476" EndTime="17.937" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="16.383" EndTime="19.490" Participant="me011"> |
| Did - did - did it - ? I didn't even <Emphasis> check </Emphasis> yesterday whether it was moving. |
| </Segment> |
| <Segment StartTime="17.937" EndTime="20.021" Participant="me018"> |
| It didn't move yesterday either when I started it. |
| </Segment> |
| <Segment StartTime="19.536" EndTime="20.461" Participant="fe016" CloseMic="false"> |
| Oh, it didn't? O_K. |
| </Segment> |
| <Segment StartTime="20.260" EndTime="20.738" Participant="me011"> |
| So. |
| </Segment> |
| <Segment StartTime="20.480" EndTime="21.926" Participant="me018"> |
| So I don't know if it doesn't like both of us - |
| </Segment> |
| <Segment StartTime="20.716" EndTime="25.543" Participant="fe016" CloseMic="false"> |
| It didn't move on me on one other meeting and it did on one and both recorded the same <Uncertain> so </Uncertain>. |
| </Segment> |
| <Segment StartTime="21.356" EndTime="21.992" Participant="me011"> |
| Channel three? |
| </Segment> |
| <Segment StartTime="21.926" EndTime="25.840" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="22.525" EndTime="23.410" Participant="me011"> |
| Channel three? |
| </Segment> |
| <Segment StartTime="23.823" EndTime="28.189" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="25.840" EndTime="29.211" Participant="me018"> |
| You know, I discovered something yesterday on these, um, wireless ones. <Comment Description="rising intonation"/> |
| </Segment> |
| <Segment StartTime="26.353" EndTime="30.527" Participant="fe016"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="27.563" EndTime="28.333" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="28.333" EndTime="30.496" Participant="me001"> |
| Channel two. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="28.859" EndTime="29.472" Participant="me011"> |
| Mm-hmm? |
| </Segment> |
| <Segment StartTime="29.818" EndTime="33.430" Participant="me018"> |
| You can tell if it's picking up <Pause/> breath noise and stuff. |
| </Segment> |
| <Segment StartTime="32.513" EndTime="33.379" Participant="me001"> |
| <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="32.823" EndTime="35.530" Participant="me011"> |
| Yeah, it has a little indicator on it - on the A_F. |
| </Segment> |
| <Segment StartTime="33.957" EndTime="34.639" Participant="me018"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="34.187" EndTime="34.943" Participant="fe016" CloseMic="false"> |
| Oh, really? |
| </Segment> |
| <Segment StartTime="35.410" EndTime="36.100" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="35.993" EndTime="40.297" Participant="fe016" CloseMic="false"> |
| Oh that's - the gains I've left here as the sort of normal ones, but - |
| </Segment> |
| <Segment StartTime="36.669" EndTime="44.968" Participant="me018"> |
| So if you - yeah, if you breathe under - breathe and then you see A_F go off, then you know <Pause/> it's p- picking up your mouth noise. |
| </Segment> |
| <Segment StartTime="40.872" EndTime="44.734" Participant="fe016"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="44.734" EndTime="47.514" Participant="fe016"> |
| Oh, that's good. Cuz we have a lot of breath noises. |
| </Segment> |
| <Segment StartTime="45.165" EndTime="47.382" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="46.488" EndTime="47.395" Participant="me011"> |
| Yep. Test. |
| </Segment> |
| <Segment StartTime="48.600" EndTime="55.714" Participant="fe016"> |
| In fact, if you listen to just the channels of people not talking, it's like "<Uncertain NumSyllables="4"> @@ </Uncertain>". <Comment Description="imitates and exaggerates heavy breathing"/> <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="50.512" EndTime="61.981" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> <VocalSound Description="laugh"/> <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="52.547" EndTime="56.129" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="55.324" EndTime="57.835" Participant="mn017"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="55.714" EndTime="57.412" Participant="fe016"> |
| It's very disgust- Sorry. <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="56.129" EndTime="59.360" Participant="me011"> |
| <VocalSound Description="inbreath"/> What? Did you see Hannibal recently or something? <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="57.948" EndTime="61.094" Participant="fe016"> |
| Exactly. <Comment Description="while laughing"/> It's very disconcerting. O_K. |
| </Segment> |
| <Segment StartTime="59.788" EndTime="61.569" Participant="me011"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="61.094" EndTime="67.030" Participant="fe016"> |
| So, um, <VocalSound Description="mouth"/> I was gonna try to get out of here, like, in half an hour, um, cuz I |
| </Segment> |
| <Segment StartTime="62.011" EndTime="62.586" Participant="me011"> |
| <Uncertain NumSyllables="1"> @@ </Uncertain> <Comment Description="uncodeable vocal sound, possible vocal gesture"/> |
| </Segment> |
| <Segment StartTime="65.547" EndTime="67.292" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="65.787" EndTime="71.673" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="67.490" EndTime="72.152" Participant="fe016"> |
| really appreciate people coming, and <VocalSound Description="breath"/> the main thing that I was |
| </Segment> |
| <Segment StartTime="73.111" EndTime="77.303" Participant="fe016"> |
| gonna ask people to help with today is <Pause/> to |
| </Segment> |
| <Segment StartTime="73.663" EndTime="74.239" CloseMic="false"> |
| <NonVocalSound Description="clinking noise"/> |
| </Segment> |
| <Segment StartTime="75.328" EndTime="75.664" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="76.610" EndTime="77.290" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="78.017" EndTime="83.943" Participant="fe016"> |
| give input on what kinds of database format we should <Pause/> use in |
| </Segment> |
| <Segment StartTime="82.401" EndTime="83.350" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="84.244" EndTime="90.022" Participant="fe016"> |
| starting to link up things like word transcripts and annotations of word transcripts, so anything that |
| </Segment> |
| <Segment StartTime="90.022" EndTime="94.158" Participant="fe016"> |
| transcribers or discourse coders or whatever put in the signal, <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="94.102" EndTime="100.374" Participant="me001"> |
| <NonVocalSound Description="pages shuffling"/> |
| </Segment> |
| <Segment StartTime="94.158" EndTime="100.561" Participant="fe016"> |
| with time-marks for, like, words and phone boundaries and all the stuff we get out of the forced alignments and the recognizer. |
| </Segment> |
| <Segment StartTime="96.570" EndTime="97.924" Participant="mn017"> |
| <VocalSound Description="sniff-breath"/> |
| </Segment> |
| <Segment StartTime="101.129" EndTime="103.861" Participant="fe016"> |
| So, we have this, um - |
| </Segment> |
| <Segment StartTime="103.861" EndTime="112.257" Participant="fe016"> |
| <VocalSound Description="mouth"/> I think a starting point is clearly the - the channelized <Pause/> output of Dave Gelbart's program, which Don brought a copy of, <Comment Description="rising intonation"/> |
| </Segment> |
| <Segment StartTime="110.526" EndTime="112.019" Participant="me011"> |
| <VocalSound Description="inbreath"/> Yeah. <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="112.019" EndTime="116.860" Participant="me011"> |
| Yeah, I'm - I'm familiar with that. I mean, we - I sort of already have developed an X_M_L format |
| </Segment> |
| <Segment StartTime="113.179" EndTime="114.700" Participant="fe016"> |
| um, which - |
| </Segment> |
| <Segment StartTime="114.408" EndTime="115.049" Participant="me018"> |
| Can I see it? |
| </Segment> |
| <Segment StartTime="117.340" EndTime="120.182" Participant="me011"> |
| for this sort of stuff. And so the only question - |
| </Segment> |
| <Segment StartTime="120.636" EndTime="123.369" Participant="me011"> |
| is it the sort of thing that you want to use or not? |
| </Segment> |
| <Segment StartTime="123.597" EndTime="125.709" Participant="me011"> |
| Have you looked at that? I mean, I had a web page up. |
| </Segment> |
| <Segment StartTime="125.474" EndTime="126.553" Participant="fe016"> |
| <VocalSound Description="inbreath"/> Right. So, |
| </Segment> |
| <Segment StartTime="126.553" EndTime="127.472" Participant="fe016"> |
| <VocalSound Description="faint creaky vocal vibrations"/> |
| </Segment> |
| <Segment StartTime="127.472" EndTime="131.596" Participant="fe016"> |
| I actually mostly need to be able to link up, or - |
| </Segment> |
| <Segment StartTime="127.834" EndTime="128.201" Participant="me011"> |
| So - <Comment Description="shortened"/> |
| </Segment> |
| <Segment StartTime="131.563" EndTime="132.117" Participant="me011"> |
| You mean, this - |
| </Segment> |
| <Segment StartTime="131.596" EndTime="134.662" Participant="fe016"> |
| I- it's - it's a question both of what the representation is and - |
| </Segment> |
| <Segment StartTime="133.679" EndTime="136.370" Participant="me011"> |
| I guess I <Emphasis> am </Emphasis> gonna be standing up and drawing on the board. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="135.136" EndTime="137.226" Participant="fe016"> |
| O_K, yeah. So you <Emphasis> should, </Emphasis> definitely. |
| </Segment> |
| <Segment StartTime="136.923" EndTime="141.477" Participant="me011"> |
| Um, so - so it definitely had that as a concept. So tha- it has a single time-line, |
| </Segment> |
| <Segment StartTime="140.268" EndTime="140.545" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="141.259" EndTime="141.855" Participant="fe016"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="141.895" EndTime="148.940" Participant="me011"> |
| and then you can have lots of different sections, each of which have I_Ds attached to it, and then you can refer from other sections to those I_Ds, |
| </Segment> |
| <Segment StartTime="149.237" EndTime="150.122" Participant="me011"> |
| if you <Emphasis> want </Emphasis> to. |
| </Segment> |
| <Segment StartTime="150.701" EndTime="151.886" Participant="me011"> |
| So that, um - |
| </Segment> |
| <Segment StartTime="152.860" EndTime="155.211" Participant="me011"> |
| so that you start with - with a <Pronounce Pronunciation=""time-lime""> time-line </Pronounce> tag. |
| </Segment> |
| <Segment StartTime="155.211" EndTime="157.700" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="157.700" EndTime="158.998" Participant="me011"> |
| "Time-line". <Comment Description="while writing on the whiteboard"/> |
| </Segment> |
| <Segment StartTime="158.998" EndTime="159.348" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="159.348" EndTime="161.496" Participant="me011"> |
| And then you have a bunch of times. |
| </Segment> |
| <Segment StartTime="161.496" EndTime="164.673" Participant="me011"> |
| I don't e- I don't remember <Emphasis> exactly </Emphasis> what my notation was, but it - |
| </Segment> |
| <Segment StartTime="163.804" EndTime="166.450" Participant="mn017"> |
| <VocalSound Description="mouth"/> Oh, I remember seeing an example of this. Yeah. |
| </Segment> |
| <Segment StartTime="164.687" EndTime="166.230" Participant="fe016"> |
| Right, right. |
| </Segment> |
| <Segment StartTime="166.317" EndTime="168.818" Participant="me011"> |
| Yeah, "T_ equals one point |
| </Segment> |
| <Segment StartTime="169.801" EndTime="171.142" Participant="me011"> |
| three two", uh - |
| </Segment> |
| <Segment StartTime="171.142" EndTime="174.098" Participant="me011"> |
| And then I - I also had optional things like accuracy, |
| </Segment> |
| <Segment StartTime="174.449" EndTime="180.318" Participant="me011"> |
| and then "I_D equals T_ one, uh, one seven". <Comment Description="while writing on the whiteboard"/> And then, |
| </Segment> |
| <Segment StartTime="180.318" EndTime="181.402" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="181.402" EndTime="187.510" Participant="me011"> |
| I also wanted to - to be i- to be able to not specify specifically what the time was and just have a stamp. |
| </Segment> |
| <Segment StartTime="186.014" EndTime="186.845" Participant="fe016"> |
| Right. |
| </Segment> |
| <Segment StartTime="187.510" EndTime="189.839" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="188.065" EndTime="189.367" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="189.839" EndTime="194.543" Participant="me011"> |
| Yeah, so these are arbitrary, assigned by a program, not - not by a user. So you have a whole bunch of those. |
| </Segment> |
| <Segment StartTime="194.543" EndTime="198.571" Participant="me011"> |
| <VocalSound Description="inbreath"/> And then somewhere la- further down you might have something like an utterance tag |
| </Segment> |
| <Segment StartTime="199.281" EndTime="200.910" Participant="me011"> |
| which has "start |
| </Segment> |
| <Segment StartTime="201.980" EndTime="206.185" Participant="me011"> |
| equals T_seventeen, end equals T_eighteen". |
| </Segment> |
| <Segment StartTime="206.185" EndTime="210.830" Participant="me011"> |
| So what <Emphasis> that's </Emphasis> saying is, we know it starts at this particular time. We don't <Emphasis> know </Emphasis> when it <Emphasis> ends. </Emphasis> |
| </Segment> |
| <Segment StartTime="210.680" EndTime="211.535" Participant="fe016"> |
| <VocalSound Description="inbreath"/> O_K. |
| </Segment> |
| <Segment StartTime="211.358" EndTime="217.266" Participant="me011"> |
| Right? But it ends at this T_eighteen, which may be somewhere else. We say there's another utterance. |
| </Segment> |
| <Segment StartTime="218.244" EndTime="222.890" Participant="me011"> |
| We don't know what the t- <Emphasis> time </Emphasis> actually is but we know that it's the same time as this end time. |
| </Segment> |
| <Segment StartTime="223.798" EndTime="225.757" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="225.265" EndTime="225.644" Participant="mn017"> |
| Mmm. |
| </Segment> |
| <Segment StartTime="225.757" EndTime="228.186" Participant="me011"> |
| <Uncertain> You know </Uncertain>, thirty-eight, whatever you want. |
| </Segment> |
| <Segment StartTime="227.948" EndTime="230.553" Participant="mn017"> |
| <VocalSound Description="inbreath"/> So <Pronounce> you're </Pronounce> essentially defining a lattice. |
| </Segment> |
| <Segment StartTime="228.806" EndTime="229.700" Participant="me011"> |
| O_K. |
| </Segment> |
| <Segment StartTime="231.170" EndTime="232.109" Participant="me011"> |
| Yes, exactly. |
| </Segment> |
| <Segment StartTime="232.124" EndTime="232.620" Participant="mn017"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="233.516" EndTime="237.050" Participant="me011"> |
| And then, uh - and then these <Emphasis> also </Emphasis> have I_Ds. |
| </Segment> |
| <Segment StartTime="237.050" EndTime="240.261" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="240.261" EndTime="247.430" Participant="me011"> |
| Right? So you could - you could have some sort of other - other tag later in the file that would be something like, um, |
| </Segment> |
| <Segment StartTime="241.833" EndTime="242.306" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="248.569" EndTime="250.360" Participant="me011"> |
| oh, I don't know, <Comment Description="embedded"/> uh, |
| </Segment> |
| <Segment StartTime="250.360" EndTime="253.494" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="253.494" EndTime="255.840" Participant="me011"> |
| "noise-type equals |
| </Segment> |
| <Segment StartTime="255.840" EndTime="257.739" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="257.739" EndTime="259.367" Participant="me011"> |
| door-slam". You know? <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="259.367" EndTime="260.847" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="260.847" EndTime="261.953" Participant="me011"> |
| And then, uh, |
| </Segment> |
| <Segment StartTime="261.953" EndTime="262.359" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="262.359" EndTime="267.790" Participant="me011"> |
| you could either say "time equals a particular time-mark" or you could do other sorts of references. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="267.790" EndTime="269.548" Participant="me011"> |
| So - or - or you might have a prosody - |
| </Segment> |
| <Segment StartTime="269.548" EndTime="270.521" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="270.521" EndTime="273.980" Participant="me011"> |
| "Prosody" <Comment Description="slower tempo, pronounced slowing while writing on whiteboard, with stress on D_"/> right? D_? T_? <Comment Description="checking spelling of "prosody""/> <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="272.563" EndTime="273.037" Participant="fe016"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="273.440" EndTime="276.299" Participant="fe016"> |
| It's an O_ instead of an I_, but the D_ is good. |
| </Segment> |
| <Segment StartTime="275.981" EndTime="276.554" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="276.188" EndTime="276.834" Participant="me001"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="276.554" EndTime="278.056" Participant="me011"> |
| You like the D_? That's a good D_. <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="276.614" EndTime="277.784" Participant="fe016"> |
| <VocalSound Description="breath-laugh"/> Yeah. |
| </Segment> |
| <Segment StartTime="277.050" EndTime="278.716" Participant="me018"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="278.056" EndTime="278.943" Participant="me011"> |
| <VocalSound Description="inbreath"/> Um, |
| </Segment> |
| <Segment StartTime="279.305" EndTime="281.402" Participant="me011"> |
| you know, so you could have some sort of <Emphasis> type </Emphasis> here, |
| </Segment> |
| <Segment StartTime="281.402" EndTime="282.173" Participant="me011"> |
| <NonVocalSound Description="tapping or writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="282.173" EndTime="284.349" Participant="me011"> |
| and then you could have, um - |
| </Segment> |
| <Segment StartTime="284.349" EndTime="288.780" Participant="me011"> |
| the utterance that it's <Emphasis> referring </Emphasis> to could be U_seventeen or something like that. |
| </Segment> |
| <Segment StartTime="285.919" EndTime="287.768" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="289.419" EndTime="291.943" Participant="fe016"> |
| O_K. So, I mean, that seems - |
| </Segment> |
| <Segment StartTime="291.866" EndTime="292.539" Participant="me011"> |
| <NonVocalSound Description="click noise, as cap is put back on marker"/> |
| </Segment> |
| <Segment StartTime="292.738" EndTime="293.648" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="292.916" EndTime="293.695" Participant="fe016"> |
| that seems g- |
| </Segment> |
| <Segment StartTime="294.445" EndTime="296.052" Participant="fe016"> |
| great for all of the |
| </Segment> |
| <Segment StartTime="294.981" EndTime="295.535" Participant="me011"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="297.284" EndTime="297.816" Participant="me011"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="297.500" EndTime="300.310" Participant="fe016"> |
| encoding of things with time and, |
| </Segment> |
| <Segment StartTime="300.044" EndTime="301.090" Participant="me011"> |
| Oh, well. |
| </Segment> |
| <Segment StartTime="300.710" EndTime="301.123" Participant="fe016"> |
| um - |
| </Segment> |
| <Segment StartTime="301.657" EndTime="304.228" Participant="fe016"> |
| <VocalSound Description="inbreath"/> I - I guess my question is more, |
| </Segment> |
| <Segment StartTime="305.027" EndTime="311.325" Participant="fe016"> |
| uh, what d- what do you do with, say, a forced alignment? <Uncertain> I mean </Uncertain> you've got all these phone labels, and what do you do if you - |
| </Segment> |
| <Segment StartTime="305.259" EndTime="306.043" Participant="mn017"> |
| How- how- |
| </Segment> |
| <Segment StartTime="311.797" EndTime="318.324" Participant="fe016"> |
| just conceptually, if you get, um, transcriptions where the words are staying |
| </Segment> |
| <Segment StartTime="318.606" EndTime="324.537" Participant="fe016"> |
| but the time boundaries are changing, cuz you've got a new recognition output, or s- sort of - what's the, um, |
| </Segment> |
| <Segment StartTime="324.905" EndTime="325.276" Participant="mn017"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="325.169" EndTime="327.930" Participant="fe016"> |
| sequence of going from the |
| </Segment> |
| <Segment StartTime="328.346" EndTime="333.434" Participant="fe016"> |
| waveforms that stay the same, the transcripts that may or may not change, and then the |
| </Segment> |
| <Segment StartTime="333.916" EndTime="337.780" Participant="fe016"> |
| utterance which - <Uncertain> where </Uncertain> the time boundaries that may or may not change - ? |
| </Segment> |
| <Segment StartTime="334.041" EndTime="335.243" Participant="mn017"> |
| <VocalSound Description="mouth"/> Oh, that's - |
| </Segment> |
| <Segment StartTime="336.996" EndTime="342.309" Participant="mn017"> |
| That's actually very nicely handled here because you could - you could - all you'd have to change is the, <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="338.247" EndTime="338.912" Participant="fe016"> |
| Um. |
| </Segment> |
| <Segment StartTime="342.309" EndTime="348.423" Participant="mn017"> |
| um, time-stamps in the time-line without - without, uh, changing the I_Ds. |
| </Segment> |
| <Segment StartTime="345.709" EndTime="347.684" Participant="fe016"> |
| And you'd be able to propagate all of the - |
| </Segment> |
| <Segment StartTime="347.779" EndTime="351.360" Participant="me011"> |
| Right. That's, the who- that's <Emphasis> why </Emphasis> you do that <Emphasis> extra </Emphasis> level of indirection. |
| </Segment> |
| <Segment StartTime="348.099" EndTime="349.340" Participant="fe016"> |
| the information? |
| </Segment> |
| <Segment StartTime="353.230" EndTime="355.340" Participant="me011"> |
| So that you can just change the <Emphasis> time-line. </Emphasis> |
| </Segment> |
| <Segment StartTime="355.680" EndTime="362.220" Participant="mn017"> |
| <VocalSound Description="inbreath"/> Except the <Emphasis> time-line </Emphasis> is gonna be <Emphasis> huge. </Emphasis> If you say - suppose you have a phone-level alignment. You'd have - you'd have - |
| </Segment> |
| <Segment StartTime="356.053" EndTime="356.586" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="357.746" EndTime="358.438" Participant="me011"> |
| Yes. |
| </Segment> |
| <Segment StartTime="357.876" EndTime="362.760" Participant="fe016"> |
| Yeah, yeah, especially at the <Emphasis> phone-level. </Emphasis> The - we - we have phone-level backtraces. <Comment Description="rising intonation"/> |
| </Segment> |
| <Segment StartTime="361.184" EndTime="363.602" Participant="me011"> |
| Yeah, this - I don't think I would do this for phone-level. |
| </Segment> |
| <Segment StartTime="363.513" EndTime="364.370" Participant="fe016"> |
| Um - |
| </Segment> |
| <Segment StartTime="363.602" EndTime="368.335" Participant="me011"> |
| I think for phone-level you want to use some sort of binary representation because it'll be too dense otherwise. |
| </Segment> |
| <Segment StartTime="367.409" EndTime="375.974" Participant="fe016"> |
| O_K. So, if you were doing that and you had this sort of companion, uh, thing that gets called up for phone-level, uh, what would that look like? How would you - ? |
| </Segment> |
| <Segment StartTime="375.230" EndTime="377.116" Participant="me011"> |
| <VocalSound Description="inbreath"/> I would use just an existing - |
| </Segment> |
| <Segment StartTime="375.429" EndTime="376.427" Participant="mn017"> |
| Why- |
| </Segment> |
| <Segment StartTime="378.191" EndTime="384.146" Participant="mn017"> |
| Mmm. But - but why <Emphasis> not </Emphasis> use it for phone-level? It's just a matter of - it's just a matter of it being bigger. But if you have - |
| </Segment> |
| <Segment StartTime="378.349" EndTime="379.241" Participant="fe016"> |
| H- h- |
| </Segment> |
| <Segment StartTime="378.373" EndTime="379.707" Participant="me011"> |
| an existing way of doing it. |
| </Segment> |
| <Segment StartTime="384.146" EndTime="384.918" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="384.918" EndTime="388.930" Participant="mn017"> |
| you know, barring <Emphasis> memory </Emphasis> limitations, or- uh - I- w- I mean this is still the m- |
| </Segment> |
| <Segment StartTime="387.566" EndTime="394.006" Participant="me011"> |
| It's <Emphasis> parsing </Emphasis> limitations. I don't want to have this text file that you have to read in the whole thing to do something very simple for. |
| </Segment> |
| <Segment StartTime="394.524" EndTime="401.240" Participant="mn017"> |
| Oh, no. You would use it <Emphasis> only </Emphasis> <Pause/> for <Pause/> purposes where you actually <Emphasis> want </Emphasis> the phone-level information, I'd imagine. |
| </Segment> |
| <Segment StartTime="400.638" EndTime="406.919" Participant="fe016"> |
| So you could have some file that configures how much information you want in your - in your X_M_L or something. |
| </Segment> |
| <Segment StartTime="406.818" EndTime="412.388" Participant="me011"> |
| Right. I mean, you'd - y- I - I am imagining you'd have multiple <Emphasis> versions </Emphasis> of this depending on the information that you <Emphasis> want. </Emphasis> |
| </Segment> |
| <Segment StartTime="407.438" EndTime="408.535" Participant="mn017"> |
| <VocalSound Description="inbreath"/> You - |
| </Segment> |
| <Segment StartTime="407.470" EndTime="410.153" Participant="fe016"> |
| Um, cuz th- it does get very bush- with - |
| </Segment> |
| <Segment StartTime="410.852" EndTime="411.477" Participant="fe016"> |
| Right. |
| </Segment> |
| <Segment StartTime="412.388" EndTime="417.948" Participant="me011"> |
| <VocalSound Description="breath"/> Um, I'm just - what I'm wondering is whether - I think for <Emphasis> word-level, </Emphasis> this would be O_K. |
| </Segment> |
| <Segment StartTime="417.125" EndTime="419.342" Participant="fe016"> |
| <VocalSound Description="inbreath"/> Yeah. Yeah. Definitely. |
| </Segment> |
| <Segment StartTime="418.194" EndTime="419.670" Participant="me011"> |
| For word-level, it's alright. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="419.254" EndTime="419.757" Participant="mn017"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="419.670" EndTime="424.951" Participant="me011"> |
| For <Emphasis> lower </Emphasis> than word-level, you're talking about <Emphasis> so </Emphasis> much data that I just - <Emphasis> I </Emphasis> don't know. I don't know if that - |
| </Segment> |
| <Segment StartTime="424.930" EndTime="429.541" Participant="fe016"> |
| <VocalSound Description="inbreath"/> I mean, we actually have - So, one thing that Don is doing, is we're - we're running - |
| </Segment> |
| <Segment StartTime="427.383" EndTime="428.628" Participant="me018"> |
| Lattices are big, too. |
| </Segment> |
| <Segment StartTime="429.541" EndTime="435.059" Participant="fe016"> |
| For every frame, you get a pitch value, and not only one pitch value but different kinds of pitch values depending on - |
| </Segment> |
| <Segment StartTime="434.294" EndTime="440.128" Participant="me011"> |
| <VocalSound Description="inbreath"/> Yeah, I mean, for something like that I would use P_file or - or any frame-level stuff I would use P_file. |
| </Segment> |
| <Segment StartTime="435.585" EndTime="436.108" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="440.231" EndTime="441.201" Participant="fe016"> |
| Meaning - ? <Comment Description="leading question"/> |
| </Segment> |
| <Segment StartTime="441.128" EndTime="447.319" Participant="me011"> |
| Uh, that's a - well, or something like it. It's ICS- uh, ICSI has a format for frame-level representation of features. |
| </Segment> |
| <Segment StartTime="447.748" EndTime="452.811" Participant="fe016"> |
| O_K. That you could call - that you would tie into this representation with like an I_D. |
| </Segment> |
| <Segment StartTime="447.762" EndTime="448.121" Participant="me018"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="448.130" EndTime="448.980" Participant="me011"> |
| Um. |
| </Segment> |
| <Segment StartTime="451.077" EndTime="451.565" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="452.645" EndTime="457.695" Participant="me011"> |
| Right. Or - or there's a - there's a particular way in X_M_L to refer to external resources. |
| </Segment> |
| <Segment StartTime="453.662" EndTime="454.255" Participant="fe016"> |
| And - |
| </Segment> |
| <Segment StartTime="457.759" EndTime="458.313" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="458.363" EndTime="461.828" Participant="me011"> |
| So you would say "refer to this external file". |
| </Segment> |
| <Segment StartTime="462.464" EndTime="465.227" Participant="me011"> |
| Um, so that external file wouldn't be in - |
| </Segment> |
| <Segment StartTime="463.347" EndTime="464.830" Participant="fe016"> |
| So that might - that might work. |
| </Segment> |
| <Segment StartTime="464.964" EndTime="469.414" Participant="me018"> |
| But what - what's the advantage of doing that versus just putting it into this format? |
| </Segment> |
| <Segment StartTime="465.501" EndTime="466.606" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="468.103" EndTime="473.690" Participant="me011"> |
| <VocalSound Description="inbreath"/> <Uncertain NumSyllables="1"> @@ </Uncertain> More compact, which I think is - is better. I mean, if you did it at this - |
| </Segment> |
| <Segment StartTime="470.374" EndTime="470.786" Participant="me018"> |
| Uh-huh. |
| </Segment> |
| <Segment StartTime="472.191" EndTime="475.838" Participant="fe016"> |
| <Uncertain> I mean </Uncertain> these are long meetings and with - for every frame, um - |
| </Segment> |
| <Segment StartTime="474.250" EndTime="479.540" Participant="me011"> |
| You don't want to do it with that - Anything at frame-level you had better encode binary or it's gonna be really painful. |
| </Segment> |
| <Segment StartTime="478.498" EndTime="484.531" Participant="mn017"> |
| <VocalSound Description="mouth"/> Or you just compre- I mean, I like text formats. Um, <Uncertain> b- </Uncertain> you can always, uh, G_zip them, <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="484.531" EndTime="489.345" Participant="mn017"> |
| and, um, you know, c- decompress them on the fly if y- if space is really a concern. |
| </Segment> |
| <Segment StartTime="485.461" EndTime="486.450" Participant="me018"> |
| <VocalSound Description="inbreath"/> Yeah, I was thi- |
| </Segment> |
| <Segment StartTime="488.507" EndTime="491.592" Participant="me018"> |
| I was thinking the advantage is that we can <Emphasis> share </Emphasis> this with other people. |
| </Segment> |
| <Segment StartTime="489.048" EndTime="492.199" Participant="me011"> |
| <VocalSound Description="inbreath"/> Well, but if you're talking about one per frame, |
| </Segment> |
| <Segment StartTime="492.808" EndTime="499.118" Participant="me011"> |
| you're talking about gigabyte-size files. You're gonna actually run out of space in your <Emphasis> filesystem </Emphasis> for one file. |
| </Segment> |
| <Segment StartTime="496.570" EndTime="499.123" Participant="fe016"> |
| These are big files. These are really - I mean - |
| </Segment> |
| <Segment StartTime="499.620" EndTime="502.850" Participant="me011"> |
| Right? Because you have a two-gigabyte limit on most O_Ss. |
| </Segment> |
| <Segment StartTime="499.887" EndTime="506.988" Participant="mn017"> |
| <VocalSound Description="mouth"/> Right, O_K. I would say - O_K, so frame-level is probably <Emphasis> not </Emphasis> a good idea. But for <Emphasis> phone-level </Emphasis> stuff it's perfectly - |
| </Segment> |
| <Segment StartTime="504.185" EndTime="505.324" Participant="fe016"> |
| <VocalSound Description="inbreath"/> And th- it's - |
| </Segment> |
| <Segment StartTime="505.523" EndTime="506.335" Participant="me001"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="508.360" EndTime="510.966" Participant="mn017"> |
| Like phones, or syllables, or anything like that. |
| </Segment> |
| <Segment StartTime="509.429" EndTime="511.848" Participant="fe016"> |
| Phones are every <Pronounce Pronunciation=""frive""> five </Pronounce> <Emphasis> frames </Emphasis> though, so. |
| </Segment> |
| <Segment StartTime="512.466" EndTime="513.455" Participant="fe016"> |
| Or <Emphasis> something </Emphasis> like that. |
| </Segment> |
| <Segment StartTime="512.983" EndTime="517.078" Participant="mn017"> |
| But - but - but <Emphasis> most </Emphasis> of the frames are actually not <Emphasis> speech. </Emphasis> <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="514.263" EndTime="516.018" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="517.540" EndTime="519.869" Participant="mn017"> |
| So, you know, people don't - |
| </Segment> |
| <Segment StartTime="518.281" EndTime="518.994" Participant="fe016"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="519.320" EndTime="521.006" Participant="fe016"> |
| Yeah, but we actually - |
| </Segment> |
| <Segment StartTime="519.869" EndTime="526.133" Participant="mn017"> |
| v- Look at it, words times the average - The average number of phones in an English word is, I don't know, <Comment Description="embedded"/> <Emphasis> five </Emphasis> maybe? |
| </Segment> |
| <Segment StartTime="521.396" EndTime="522.260" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="524.772" EndTime="525.642" Participant="me011"> |
| <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="526.133" EndTime="526.810" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="526.810" EndTime="530.612" Participant="mn017"> |
| So, look at it, t- number of words times <Emphasis> five. </Emphasis> That's not - that not - |
| </Segment> |
| <Segment StartTime="528.869" EndTime="532.509" Participant="fe016"> |
| Oh, so you mean <Emphasis> pause </Emphasis> phones take up a lot of the - long pause phones. |
| </Segment> |
| <Segment StartTime="532.460" EndTime="532.863" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="532.673" EndTime="533.980" Participant="mn017"> |
| Exactly. Yeah. |
| </Segment> |
| <Segment StartTime="532.825" EndTime="536.906" Participant="fe016"> |
| Yeah. O_K. That's true. But you do have to keep them in there. Y- yeah. O_K. |
| </Segment> |
| <Segment StartTime="535.557" EndTime="542.503" Participant="me011"> |
| <VocalSound Description="breath"/> So I think it - it's debatable whether you want to do phone-level in the same thing. But I think, a- anything at frame-level, even <Emphasis> P_file </Emphasis>, |
| </Segment> |
| <Segment StartTime="539.653" EndTime="540.275" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="541.503" EndTime="542.156" Participant="fe016"> |
| So - |
| </Segment> |
| <Segment StartTime="542.893" EndTime="546.125" Participant="me011"> |
| is too verbose. I would use something <Emphasis> tighter </Emphasis> than P_files. |
| </Segment> |
| <Segment StartTime="545.973" EndTime="549.375" Participant="fe016"> |
| Do you - Are you familiar with it? I haven't seen this particular format, but - |
| </Segment> |
| <Segment StartTime="546.772" EndTime="547.207" Participant="me011"> |
| So. |
| </Segment> |
| <Segment StartTime="548.212" EndTime="552.500" Participant="mn017"> |
| I mean, I've - I've <Emphasis> used </Emphasis> them. I don't know what their structure is. I've forgot what the str- |
| </Segment> |
| <Segment StartTime="549.936" EndTime="550.830" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="551.698" EndTime="557.547" Participant="me018"> |
| <VocalSound Description="inbreath"/> But, wait a minute, P_file for each frame is storing a vector of cepstral or P_L_P values, right? |
| </Segment> |
| <Segment StartTime="557.180" EndTime="561.550" Participant="me011"> |
| It's whatever you want, actually. So that - what's nice about the P_file - It - i- |
| </Segment> |
| <Segment StartTime="558.042" EndTime="558.576" Participant="me018"> |
| Right. |
| </Segment> |
| <Segment StartTime="562.145" EndTime="567.921" Participant="me011"> |
| Built into it is the concept of <Pause/> frames, utterances, sentences, that sort of thing, that structure. <Comment Description="rising intonation"/> |
| </Segment> |
| <Segment StartTime="565.631" EndTime="566.746" Participant="mn017"> |
| <NonVocalSound Description="pages turning"/> |
| </Segment> |
| <Segment StartTime="567.458" EndTime="569.787" Participant="mn017"> |
| <NonVocalSound Description="pages turning"/> |
| </Segment> |
| <Segment StartTime="567.921" EndTime="568.277" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="568.277" EndTime="573.043" Participant="me011"> |
| And then also <Emphasis> attached </Emphasis> to it is an arbitrary vector of values. |
| </Segment> |
| <Segment StartTime="573.318" EndTime="573.907" Participant="fe016"> |
| Oh. |
| </Segment> |
| <Segment StartTime="573.461" EndTime="580.191" Participant="me011"> |
| And it can take different types. So it - th- they don't all have to be floats. You know, you can have integers and you can have doubles, and all that sort of stuff. |
| </Segment> |
| <Segment StartTime="574.468" EndTime="575.573" Participant="mn017"> |
| <NonVocalSound Description="pages turning"/> |
| </Segment> |
| <Segment StartTime="576.006" EndTime="577.486" Participant="mn017"> |
| <NonVocalSound Description="pages turning"/> |
| </Segment> |
| <Segment StartTime="578.494" EndTime="581.759" Participant="fe016"> |
| So that - that sounds - that sounds about what I w- |
| </Segment> |
| <Segment StartTime="580.705" EndTime="581.218" Participant="me011"> |
| Um. |
| </Segment> |
| <Segment StartTime="581.821" EndTime="587.724" Participant="me011"> |
| Right? And it has a header - it has a header format that <Pause/> describes it <Pause/> to some extent. So, <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="587.951" EndTime="595.019" Participant="me011"> |
| the only <Emphasis> problem </Emphasis> with it is it's actually storing the <Pause/> utterance numbers and the <Pause/> frame numbers in the file, <Comment Description="rising intonation"/> |
| </Segment> |
| <Segment StartTime="595.182" EndTime="599.075" Participant="me011"> |
| even though they're always sequential. <Comment Description="while laughing"/> And so it does waste a lot of space. |
| </Segment> |
| <Segment StartTime="599.681" EndTime="600.071" Participant="mn017"> |
| Hmm. |
| </Segment> |
| <Segment StartTime="599.697" EndTime="604.379" Participant="me011"> |
| But it's still a <Emphasis> lot </Emphasis> tighter than - than ASCII. <Emphasis> And </Emphasis> we have a lot of tools already to deal with it. |
| </Segment> |
| <Segment StartTime="603.685" EndTime="609.200" Participant="fe016"> |
| You do? O_K. Is there some documentation on this somewhere? O_K, great. So, |
| </Segment> |
| <Segment StartTime="606.650" EndTime="610.816" Participant="me011"> |
| Yeah, there's a ton of it. Man-pages and, uh, source code, and |
| </Segment> |
| <Segment StartTime="611.229" EndTime="617.218" Participant="fe016"> |
| I mean, that sounds good. I - I was just looking for something - I'm not a database person, <Comment Description="embedded"/> but something sort of standard enough that, |
| </Segment> |
| <Segment StartTime="611.328" EndTime="612.519" Participant="me011"> |
| me. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="612.062" EndTime="612.544" Participant="me018"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="617.732" EndTime="622.600" Participant="fe016"> |
| you know, if we start using this we can give it out, other people can work on it, or - <Comment Description="rising intonation"/> Is it - ? |
| </Segment> |
| <Segment StartTime="620.261" EndTime="623.732" Participant="me011"> |
| <VocalSound Description="inbreath"/> Yeah, it's not standard. I mean, it's something that we developed at ICSI. |
| </Segment> |
| <Segment StartTime="621.618" EndTime="622.448" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="624.602" EndTime="625.790" Participant="me011"> |
| But, uh - |
| </Segment> |
| <Segment StartTime="625.929" EndTime="629.610" Participant="fe016"> |
| But it's <Pause/> been used here and people've - |
| </Segment> |
| <Segment StartTime="627.799" EndTime="634.800" Participant="me011"> |
| But it's been used here and - and, you know, we have a <Pause/> well-configured system that you can distribute for free, and - |
| </Segment> |
| <Segment StartTime="633.882" EndTime="640.116" Participant="me018"> |
| <VocalSound Description="inbreath"/> I mean, it must be the equivalent of whatever <Emphasis> you </Emphasis> guys used to store feat- your computed features in, right? |
| </Segment> |
| <Segment StartTime="635.207" EndTime="636.238" Participant="fe016"> |
| O_K. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="640.100" EndTime="641.155" Participant="mn017"> |
| Yeah, th- we have - |
| </Segment> |
| <Segment StartTime="641.669" EndTime="646.753" Participant="mn017"> |
| Actually, we - we use a generalization of the - the Sphere format. <Comment Description="rising intonation"/> <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="645.434" EndTime="646.138" Participant="me018"> |
| Mmm. |
| </Segment> |
| <Segment StartTime="647.287" EndTime="649.326" Participant="mn017"> |
| Um, <VocalSound Description="mouth"/> but - |
| </Segment> |
| <Segment StartTime="650.341" EndTime="654.490" Participant="mn017"> |
| Yeah, so there is something like that but it's, um, probably not as sophist- |
| </Segment> |
| <Segment StartTime="652.753" EndTime="653.778" Participant="me018"> |
| And I think there's - |
| </Segment> |
| <Segment StartTime="653.224" EndTime="657.372" Participant="me011"> |
| Well, what does H_T_<Emphasis> K </Emphasis> do for features? Or does it even <Emphasis> have </Emphasis> a concept of features? |
| </Segment> |
| <Segment StartTime="655.842" EndTime="657.451" Participant="mn017"> |
| They ha- it has its own - |
| </Segment> |
| <Segment StartTime="657.451" EndTime="664.422" Participant="mn017"> |
| I mean, Entropic has their own feature format that's called, like, S_- S_D or some so- S_F or something like that. |
| </Segment> |
| <Segment StartTime="662.076" EndTime="662.807" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="663.980" EndTime="664.582" Participant="me018"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="664.370" EndTime="667.036" Participant="me011"> |
| I'm just wondering, would it be worth while to use that <Emphasis> instead? </Emphasis> |
| </Segment> |
| <Segment StartTime="665.114" EndTime="665.609" Participant="mn017"> |
| <Uncertain> Hmm? </Uncertain> |
| </Segment> |
| <Segment StartTime="665.632" EndTime="677.223" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="666.882" EndTime="670.610" Participant="fe016"> |
| Yeah. Th- this is exactly the kind of decision - It's just whatever - |
| </Segment> |
| <Segment StartTime="668.018" EndTime="673.275" Participant="me018"> |
| <VocalSound Description="inbreath"/> But, I mean, people don't typically <Emphasis> share </Emphasis> <Emphasis> this </Emphasis> kind of stuff, right? I mean - |
| </Segment> |
| <Segment StartTime="672.575" EndTime="673.057" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="672.627" EndTime="673.181" Participant="fe016"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="673.174" EndTime="674.630" Participant="me011"> |
| They generate their own. <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="674.073" EndTime="675.584" Participant="fe016"> |
| Actually, I - |
| </Segment> |
| <Segment StartTime="674.225" EndTime="675.110" Participant="me018"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="675.960" EndTime="679.910" Participant="fe016"> |
| I just - you know, we - we've done this stuff on <Emphasis> prosodics </Emphasis> and |
| </Segment> |
| <Segment StartTime="679.655" EndTime="680.416" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="680.167" EndTime="684.163" Participant="fe016"> |
| three or four places have asked for those prosodic <Emphasis> files, </Emphasis> and we just have an ASCII, |
| </Segment> |
| <Segment StartTime="684.163" EndTime="687.088" Participant="fe016"> |
| uh, output of frame-by-frame. Which is fine, |
| </Segment> |
| <Segment StartTime="684.527" EndTime="685.224" Participant="me011"> |
| Ah, right. |
| </Segment> |
| <Segment StartTime="687.505" EndTime="692.660" Participant="fe016"> |
| but it gets unwieldy to go in and - and query these files with really huge files. |
| </Segment> |
| <Segment StartTime="692.676" EndTime="693.309" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="693.060" EndTime="697.815" Participant="fe016"> |
| I mean, we could <Emphasis> do </Emphasis> it. I was just thinking if there's something that - where <Emphasis> all </Emphasis> the frame values are - Hmm? |
| </Segment> |
| <Segment StartTime="695.211" EndTime="696.916" Participant="me011"> |
| And a- and again, if you have a - |
| </Segment> |
| <Segment StartTime="697.510" EndTime="701.656" Participant="me011"> |
| if you have a two-hour-long meeting, that's gonna - |
| </Segment> |
| <Segment StartTime="701.129" EndTime="707.240" Participant="fe016"> |
| They're - they're fair- they're quite large. And these are for ten-minute Switchboard conversations, and - <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="702.642" EndTime="704.346" Participant="me011"> |
| Yeah, I mean, they'd be emo- enormous. |
| </Segment> |
| <Segment StartTime="705.940" EndTime="706.443" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="707.240" EndTime="713.161" Participant="fe016"> |
| So it's doable, it's just that you can only store a feature vector at frame-by-frame and it doesn't have any kind of, |
| </Segment> |
| <Segment StartTime="714.510" EndTime="715.370" Participant="fe016"> |
| um - |
| </Segment> |
| <Segment StartTime="714.829" EndTime="722.290" Participant="me018"> |
| Is - is the <Emphasis> sharing </Emphasis> part of this a pretty important <Pause/> consideration or does that just sort of, uh - a nice thing to have? |
| </Segment> |
| <Segment StartTime="718.684" EndTime="721.989" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="722.947" EndTime="730.304" Participant="fe016"> |
| I - I don't know enough about what we're gonna do with the data. <Comment Description="rising intonation"/> But I thought it would be good to get something that |
| </Segment> |
| <Segment StartTime="724.650" EndTime="725.283" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="726.190" EndTime="728.635" Participant="mn017"> |
| <VocalSound Description="breath"/> <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="729.424" EndTime="730.285" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="731.241" EndTime="735.277" Participant="fe016"> |
| we can - that other people can use or adopt for <Pronounce> their </Pronounce> own kinds of |
| </Segment> |
| <Segment StartTime="735.886" EndTime="742.470" Participant="fe016"> |
| encoding. And just, <Uncertain> I mean </Uncertain> we have to use some- we have to make <Emphasis> some </Emphasis> decision about what to do. And especially for the prosody work, what - |
| </Segment> |
| <Segment StartTime="738.018" EndTime="739.084" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="739.681" EndTime="740.136" Participant="me011"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="742.760" EndTime="744.299" Participant="fe016"> |
| what it ends up being is you get |
| </Segment> |
| <Segment StartTime="744.828" EndTime="750.570" Participant="fe016"> |
| features from the signal, and of course those change every time your alignments change. So you re-run a recognizer, you want to recompute |
| </Segment> |
| <Segment StartTime="751.219" EndTime="756.930" Participant="fe016"> |
| your features, um, and then keep the database up to date. Or you change a word, or you change a <VocalSound Description="mouth"/> |
| </Segment> |
| <Segment StartTime="754.757" EndTime="755.256" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="757.170" EndTime="760.195" Participant="fe016"> |
| utterance boundary segment, which is gonna happen a lot. |
| </Segment> |
| <Segment StartTime="758.064" EndTime="759.315" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="760.697" EndTime="766.077" Participant="fe016"> |
| And so I wanted something where <Pause/> all of this can be done in a elegant way and that if somebody |
| </Segment> |
| <Segment StartTime="766.887" EndTime="770.845" Participant="fe016"> |
| wants to <Emphasis> try </Emphasis> something or compute something <Emphasis> else, </Emphasis> that it can be done flexibly. |
| </Segment> |
| <Segment StartTime="771.335" EndTime="774.670" Participant="fe016"> |
| Um, it doesn't have to be <Emphasis> pretty, </Emphasis> it just has to be, |
| </Segment> |
| <Segment StartTime="775.403" EndTime="776.815" Participant="fe016"> |
| you know, easy to use, and - |
| </Segment> |
| <Segment StartTime="777.090" EndTime="778.392" Participant="me011"> |
| <VocalSound Description="breath"/> Yeah, the other thing - |
| </Segment> |
| <Segment StartTime="778.966" EndTime="782.004" Participant="me011"> |
| We should look at ATLAS, the NIST thing, |
| </Segment> |
| <Segment StartTime="779.568" EndTime="780.725" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="781.213" EndTime="781.579" Participant="fe016"> |
| <Uncertain> Oh. </Uncertain> |
| </Segment> |
| <Segment StartTime="781.899" EndTime="782.347" Participant="mn017"> |
| Mmm. |
| </Segment> |
| <Segment StartTime="782.156" EndTime="783.441" Participant="fe016"> |
| <VocalSound Description="inbreath"/> Uh - <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="782.818" EndTime="788.732" Participant="me011"> |
| and see if they have anything at that level. I mean, I'm not sure what to do about <Emphasis> this </Emphasis> with ATLAS, because they chose a different route. |
| </Segment> |
| <Segment StartTime="788.732" EndTime="789.116" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="789.116" EndTime="793.399" Participant="me011"> |
| I chose something that - Th- there are sort of two choices. Your - your file format |
| </Segment> |
| <Segment StartTime="791.507" EndTime="791.863" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="794.033" EndTime="795.123" Participant="me011"> |
| can know about - |
| </Segment> |
| <Segment StartTime="795.795" EndTime="800.201" Participant="me011"> |
| know that you're talking about language <Pause/> and speech, which is what <Emphasis> I </Emphasis> chose, and time, |
| </Segment> |
| <Segment StartTime="800.201" EndTime="803.233" Participant="me011"> |
| <VocalSound Description="inbreath"/> or your file format can just be a graph representation. |
| </Segment> |
| <Segment StartTime="802.117" EndTime="806.135" Participant="me018"> |
| <VocalSound Description="yawn"/> |
| </Segment> |
| <Segment StartTime="803.728" EndTime="806.327" Participant="me011"> |
| And then the application has to impose the structure on top. |
| </Segment> |
| <Segment StartTime="806.327" EndTime="806.844" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="806.844" EndTime="811.449" Participant="me011"> |
| So what it looked like ATLAS chose is, <Emphasis> they </Emphasis> chose the <Emphasis> other </Emphasis> way, which was their file format is just |
| </Segment> |
| <Segment StartTime="811.908" EndTime="815.983" Participant="me011"> |
| nodes and links, and you have to interpret what they mean yourself. |
| </Segment> |
| <Segment StartTime="816.727" EndTime="819.890" Participant="fe016"> |
| And why did you not choose that type of approach? |
| </Segment> |
| <Segment StartTime="818.161" EndTime="821.565" Participant="me011"> |
| <VocalSound Description="inbreath"/> Uh, because I knew that we were doing speech, and I thought it was <Emphasis> better </Emphasis> |
| </Segment> |
| <Segment StartTime="820.618" EndTime="821.224" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="822.231" EndTime="828.660" Participant="me011"> |
| if you're looking at a raw file to be - t- for the tags to say "it's an utterance", as opposed to the tag to say "it's a link". |
| </Segment> |
| <Segment StartTime="823.732" EndTime="824.168" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="828.658" EndTime="829.184" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="829.492" EndTime="832.310" Participant="fe016"> |
| But other than that, are they compatible? I mean, you could sort of - |
| </Segment> |
| <Segment StartTime="829.851" EndTime="830.345" Participant="me011"> |
| So, but - |
| </Segment> |
| <Segment StartTime="832.130" EndTime="833.772" Participant="me011"> |
| Yeah, they're reasonably compatible. |
| </Segment> |
| <Segment StartTime="832.914" EndTime="834.480" Participant="fe016"> |
| I mean, you - you could - |
| </Segment> |
| <Segment StartTime="833.745" EndTime="835.550" Participant="me018"> |
| You could probably translate between <Pronounce Pronunciation=""em""> them. </Pronounce> |
| </Segment> |
| <Segment StartTime="835.530" EndTime="837.590" Participant="fe016"> |
| Yeah, that's w- So, O_K. |
| </Segment> |
| <Segment StartTime="835.739" EndTime="836.086" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="836.422" EndTime="838.598" Participant="me011"> |
| <VocalSound Description="inbreath"/> So, well, the other thing <Emphasis> is </Emphasis> |
| </Segment> |
| <Segment StartTime="838.598" EndTime="845.164" Participant="me011"> |
| if we choose to use <Emphasis> ATLAS, </Emphasis> which maybe we should just <Emphasis> do, </Emphasis> we should just throw this <Emphasis> out </Emphasis> before we invest a lot of <Emphasis> time </Emphasis> in it. |
| </Segment> |
| <Segment StartTime="842.347" EndTime="842.759" Participant="fe016"> |
| I don't - |
| </Segment> |
| <Segment StartTime="845.059" EndTime="852.231" Participant="fe016"> |
| So this is what the meeting's about, just sort of how to - Um, cuz we need to come up with a database like this just to do our work. And I actually don't |
| </Segment> |
| <Segment StartTime="846.570" EndTime="847.075" Participant="me011"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="852.733" EndTime="856.650" Participant="fe016"> |
| care, as long as it's something useful to other people, what we <Emphasis> choose. </Emphasis> So maybe it's - |
| </Segment> |
| <Segment StartTime="855.018" EndTime="855.533" Participant="me011"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="857.071" EndTime="860.818" Participant="fe016"> |
| maybe oth- you know, if - if you have any idea of how to choose, cuz I don't. |
| </Segment> |
| <Segment StartTime="858.005" EndTime="859.093" Participant="me011"> |
| <VocalSound Description="breath"/> The only thing - |
| </Segment> |
| <Segment StartTime="860.418" EndTime="861.469" Participant="me011"> |
| <VocalSound Description="breath"/> Yeah. |
| </Segment> |
| <Segment StartTime="861.372" EndTime="863.080" Participant="mn017"> |
| Do they already have <Emphasis> tools? </Emphasis> |
| </Segment> |
| <Segment StartTime="861.469" EndTime="865.845" Participant="me011"> |
| <VocalSound Description="breath"/> I mean, I - I <Emphasis> chose </Emphasis> this for a couple <Emphasis> reasons. </Emphasis> One of <Pronounce Pronunciation=""em""> them </Pronounce> is that it's easy to <Emphasis> parse. </Emphasis> |
| </Segment> |
| <Segment StartTime="864.761" EndTime="870.866" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="866.458" EndTime="871.635" Participant="me011"> |
| You don't need a full X_M_L parser. It's very easy to just write a Perl script <Pause/> to parse it. |
| </Segment> |
| <Segment StartTime="871.657" EndTime="876.138" Participant="mn017"> |
| As long as <Uncertain> uh </Uncertain> each tag is on one line. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="874.040" EndTime="877.220" Participant="me011"> |
| Exactly. Exactly. Which I always do. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="875.239" EndTime="876.105" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="876.365" EndTime="879.957" Participant="fe016"> |
| And you can have as much information in the tag as you want, right? |
| </Segment> |
| <Segment StartTime="877.966" EndTime="878.574" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="880.182" EndTime="886.333" Participant="me011"> |
| <VocalSound Description="breath"/> Well, I have it structured. Right? So each type tag has only particular items that it can take. |
| </Segment> |
| <Segment StartTime="886.490" EndTime="889.412" Participant="fe016"> |
| Can you - But you can add to those structures |
| </Segment> |
| <Segment StartTime="889.192" EndTime="889.845" Participant="me011"> |
| Sure. |
| </Segment> |
| <Segment StartTime="889.788" EndTime="890.875" Participant="fe016"> |
| if you - |
| </Segment> |
| <Segment StartTime="890.980" EndTime="892.203" Participant="me011"> |
| If you have more information. <Comment Description="rising intonation"/> <Comment Description="completing [fe016]'s utterance"/> So what - |
| </Segment> |
| <Segment StartTime="891.872" EndTime="893.380" Participant="fe016"> |
| Yeah. So - |
| </Segment> |
| <Segment StartTime="892.203" EndTime="896.056" Participant="me011"> |
| What <Emphasis> NIST </Emphasis> would say is that instead of doing <Emphasis> this, </Emphasis> you would say something like "link |
| </Segment> |
| <Segment StartTime="896.056" EndTime="898.004" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="898.004" EndTime="905.220" Participant="me011"> |
| start equals, um, you know, some node I_D, end equals some other node I_D", |
| </Segment> |
| <Segment StartTime="905.685" EndTime="909.480" Participant="me011"> |
| and then "type" would be "utterance". <Comment Description="while writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="909.480" EndTime="911.061" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="910.401" EndTime="910.745" Participant="mn017"> |
| Hmm. |
| </Segment> |
| <Segment StartTime="910.557" EndTime="910.893" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="911.061" EndTime="912.740" Participant="me011"> |
| You know, so it's very similar. |
| </Segment> |
| <Segment StartTime="913.369" EndTime="918.915" Participant="fe016"> |
| So why would it be a - a <Emphasis> waste </Emphasis> to <Emphasis> do </Emphasis> it this way if it's <Emphasis> similar </Emphasis> enough that we can always <Emphasis> translate </Emphasis> it? |
| </Segment> |
| <Segment StartTime="913.900" EndTime="920.032" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="918.212" EndTime="922.270" Participant="me018"> |
| It probably wouldn't be a <Emphasis> waste. </Emphasis> It would mean that at some point if we wanted to switch, we'd just have to |
| </Segment> |
| <Segment StartTime="922.127" EndTime="923.496" Participant="me011"> |
| Write a translator. But it se- |
| </Segment> |
| <Segment StartTime="922.940" EndTime="924.000" Participant="me018"> |
| translate everything. |
| </Segment> |
| <Segment StartTime="923.820" EndTime="925.264" Participant="fe016"> |
| But it - but that sounds - |
| </Segment> |
| <Segment StartTime="923.910" EndTime="925.793" Participant="me011"> |
| Since they are developing a big - |
| </Segment> |
| <Segment StartTime="924.935" EndTime="926.655" Participant="me018"> |
| But that's - I don't think that's a big deal. |
| </Segment> |
| <Segment StartTime="925.778" EndTime="927.162" Participant="fe016"> |
| As long as it is - |
| </Segment> |
| <Segment StartTime="926.993" EndTime="930.316" Participant="me011"> |
| they're developing a big infrastructure. And so it seems to me that if - |
| </Segment> |
| <Segment StartTime="930.316" EndTime="935.129" Participant="me011"> |
| <Emphasis> if </Emphasis> we want to use that, we might as well go <Emphasis> directly </Emphasis> to what they're doing, rather than - |
| </Segment> |
| <Segment StartTime="930.614" EndTime="931.571" Participant="mn017"> |
| If we want to - |
| </Segment> |
| <Segment StartTime="933.927" EndTime="938.066" Participant="mn017"> |
| Do they already <Emphasis> have </Emphasis> something that's - that would be useful for us in place? |
| </Segment> |
| <Segment StartTime="936.500" EndTime="939.627" Participant="me018"> |
| Yeah. See, that's the question. I mean, how stable is their - |
| </Segment> |
| <Segment StartTime="940.201" EndTime="941.605" Participant="me018"> |
| Are they ready to go, or - ? <Comment Description="leading question"/> |
| </Segment> |
| <Segment StartTime="940.841" EndTime="945.347" Participant="me011"> |
| The - I looked at it - The last time I looked at it was a <Emphasis> while </Emphasis> ago, probably a <Emphasis> year </Emphasis> ago, |
| </Segment> |
| <Segment StartTime="945.422" EndTime="945.765" Participant="me018"> |
| Hmm. |
| </Segment> |
| <Segment StartTime="945.679" EndTime="949.091" Participant="me011"> |
| uh, when we first started talking about this. And at <Emphasis> that </Emphasis> time at least |
| </Segment> |
| <Segment StartTime="949.091" EndTime="949.586" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="949.586" EndTime="951.976" Participant="me011"> |
| it was still not very <Pause/> complete. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="951.976" EndTime="956.855" Participant="me011"> |
| And so, specifically they didn't have any external format representation at that time. |
| </Segment> |
| <Segment StartTime="956.855" EndTime="957.069" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="957.069" EndTime="960.614" Participant="me011"> |
| They just had the sort of conceptual <Pause/> node - |
| </Segment> |
| <Segment StartTime="960.614" EndTime="965.447" Participant="me011"> |
| uh, annotated transcription graph, which I really <Emphasis> liked. </Emphasis> And that's exactly what this stuff is based on. |
| </Segment> |
| <Segment StartTime="965.447" EndTime="965.816" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="965.816" EndTime="968.624" Participant="me011"> |
| <Emphasis> Since </Emphasis> then, they've developed their own external file format, |
| </Segment> |
| <Segment StartTime="966.862" EndTime="967.447" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="968.624" EndTime="969.099" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="969.099" EndTime="972.136" Participant="me011"> |
| which is, uh, you know, this sort of s- <Emphasis> this </Emphasis> sort of thing. |
| </Segment> |
| <Segment StartTime="972.136" EndTime="972.561" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="972.561" EndTime="976.578" Participant="me011"> |
| Um, and apparently they've also developed a lot of <Emphasis> tools, </Emphasis> but I haven't looked at them. Maybe I should. |
| </Segment> |
| <Segment StartTime="976.205" EndTime="977.997" Participant="mn017"> |
| <Uncertain NumSyllables="1"> @@ </Uncertain> We should - we should find out. |
| </Segment> |
| <Segment StartTime="976.822" EndTime="981.548" Participant="fe016"> |
| I mean, would the tools - would the tools run on something like <Emphasis> this, </Emphasis> if you can translate them anyway? |
| </Segment> |
| <Segment StartTime="979.356" EndTime="980.144" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="979.686" EndTime="980.121" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="980.515" EndTime="982.116" Participant="me001"> |
| <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="981.686" EndTime="988.192" Participant="me011"> |
| <VocalSound Description="mouth"/> Um, th- what would - would - would - what would <Emphasis> worry </Emphasis> me is that maybe we might miss a little detail |
| </Segment> |
| <Segment StartTime="982.577" EndTime="984.630" Participant="fe016"> |
| I mean, that - I guess it's a question that - <Uncertain> uh, yeah </Uncertain>. |
| </Segment> |
| <Segment StartTime="982.673" EndTime="983.884" Participant="mn017"> |
| It's a hassle |
| </Segment> |
| <Segment StartTime="984.690" EndTime="985.239" Participant="mn017"> |
| if - <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="988.192" EndTime="991.123" Participant="me011"> |
| that would make it very difficult to translate from one to the other. |
| </Segment> |
| <Segment StartTime="988.553" EndTime="989.164" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="990.073" EndTime="998.857" Participant="mn017"> |
| I - I think if it's conceptually <Emphasis> close, </Emphasis> and they already have or will have tools that everybody <Emphasis> else </Emphasis> will be using, I mean, |
| </Segment> |
| <Segment StartTime="990.860" EndTime="991.405" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="998.846" EndTime="999.953" Participant="me011"> |
| Yeah, we might as well. |
| </Segment> |
| <Segment StartTime="998.857" EndTime="999.229" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="999.229" EndTime="1003.664" Participant="mn017"> |
| it would be crazy to do something s- you know, separate that - |
| </Segment> |
| <Segment StartTime="1003.276" EndTime="1003.692" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1004.031" EndTime="1004.486" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1004.602" EndTime="1006.718" Participant="me011"> |
| <VocalSound Description="inbreath"/> So I'll - I'll take a closer look at it. |
| </Segment> |
| <Segment StartTime="1005.830" EndTime="1012.041" Participant="fe016"> |
| Actually, so it's - that - that would really be the question, is just what you would feel is in the long run the best thing. Cuz <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1008.201" EndTime="1008.755" Participant="me011"> |
| And - |
| </Segment> |
| <Segment StartTime="1009.764" EndTime="1010.398" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1011.782" EndTime="1012.276" Participant="me011"> |
| The - |
| </Segment> |
| <Segment StartTime="1012.041" EndTime="1019.481" Participant="fe016"> |
| once we start, sort of, doing this I don't - we don't actually have enough time to probably have to rehash it out again and - |
| </Segment> |
| <Segment StartTime="1014.910" EndTime="1015.330" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1018.492" EndTime="1019.382" Participant="me011"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1018.915" EndTime="1022.906" Participant="me018"> |
| <VocalSound Description="yawn"/> |
| </Segment> |
| <Segment StartTime="1019.382" EndTime="1026.553" Participant="me011"> |
| The other thing - the other way that I sort of established <Emphasis> this </Emphasis> was as easy translation to and from the <Emphasis> Transcriber </Emphasis> format. |
| </Segment> |
| <Segment StartTime="1020.189" EndTime="1021.197" Participant="fe016"> |
| s- <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1026.593" EndTime="1028.400" Participant="fe016"> |
| Right. Right. |
| </Segment> |
| <Segment StartTime="1027.470" EndTime="1029.193" Participant="me011"> |
| Um, but - |
| </Segment> |
| <Segment StartTime="1029.174" EndTime="1035.367" Participant="fe016"> |
| I mean, I <Emphasis> like </Emphasis> this. This is sort of intuitively easy to actually r- read, as easy it could - as it could be. But, |
| </Segment> |
| <Segment StartTime="1032.004" EndTime="1032.382" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1035.829" EndTime="1040.274" Participant="fe016"> |
| I suppose that <Pause/> as long as they have a type here that specifies "utt", |
| </Segment> |
| <Segment StartTime="1040.591" EndTime="1041.145" Participant="fe016"> |
| um, |
| </Segment> |
| <Segment StartTime="1041.165" EndTime="1047.761" Participant="me011"> |
| It's <Emphasis> almost </Emphasis> the same. The - the - the - the <Emphasis> point </Emphasis> is - with <Emphasis> this, </Emphasis> though, is that you can't really add any supplementary information. |
| </Segment> |
| <Segment StartTime="1042.081" EndTime="1043.637" Participant="fe016"> |
| it's - yeah, close enough that - |
| </Segment> |
| <Segment StartTime="1042.895" EndTime="1044.000" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1043.065" EndTime="1048.810" Participant="mn017"> |
| <VocalSound Description="yawn"/> |
| </Segment> |
| <Segment StartTime="1048.492" EndTime="1050.641" Participant="me011"> |
| Right? So if you suddenly decide that you want - |
| </Segment> |
| <Segment StartTime="1049.472" EndTime="1050.358" Participant="fe016"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1050.358" EndTime="1052.109" Participant="fe016"> |
| You have to make a different type. <Comment Description="rising intonation"/> <Comment Description="completing [me011]'s utterance"/> |
| </Segment> |
| <Segment StartTime="1051.670" EndTime="1053.846" Participant="me011"> |
| Yeah. You'd have to make a different type. |
| </Segment> |
| <Segment StartTime="1053.463" EndTime="1055.250" Participant="fe016"> |
| So - Well, if you look at it and - |
| </Segment> |
| <Segment StartTime="1054.162" EndTime="1056.535" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1055.720" EndTime="1059.443" Participant="fe016"> |
| Um, I guess in my mind I don't know enough - Jane would know better, <Comment Description="embedded"/> about |
| </Segment> |
| <Segment StartTime="1060.200" EndTime="1063.228" Participant="fe016"> |
| the <Pause/> <Emphasis> types </Emphasis> of <Emphasis> annotations </Emphasis> and - and - |
| </Segment> |
| <Segment StartTime="1063.228" EndTime="1068.150" Participant="fe016"> |
| But I imagine that those are things that would - well, <Emphasis> you </Emphasis> guys mentioned this, <Comment Description="embedded"/> that could span any - |
| </Segment> |
| <Segment StartTime="1063.616" EndTime="1064.408" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1064.764" EndTime="1065.997" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1068.610" EndTime="1069.263" Participant="mn017"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1068.748" EndTime="1076.504" Participant="fe016"> |
| it could be in its own channel, it could span time boundaries of any type, it could be instantaneous, things like that. Um, |
| </Segment> |
| <Segment StartTime="1072.464" EndTime="1073.137" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1076.847" EndTime="1079.005" Participant="fe016"> |
| and then from the <Emphasis> recognition </Emphasis> side we have |
| </Segment> |
| <Segment StartTime="1079.973" EndTime="1084.800" Participant="fe016"> |
| backtraces at the phone-level. If - if it can handle <Emphasis> that, </Emphasis> it could handle states or whatever. <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1081.629" EndTime="1082.434" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1084.800" EndTime="1087.253" Participant="fe016"> |
| And then at the <Emphasis> prosody-level </Emphasis> we have |
| </Segment> |
| <Segment StartTime="1087.827" EndTime="1089.844" Participant="fe016"> |
| frame - sort of like cepstral |
| </Segment> |
| <Segment StartTime="1089.085" EndTime="1089.711" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1090.330" EndTime="1095.397" Participant="fe016"> |
| feature files, uh, like these P_files or anything like that. And that's sort of the world of things that I - |
| </Segment> |
| <Segment StartTime="1095.788" EndTime="1098.990" Participant="fe016"> |
| And then we have the aligned channels, of course, and - |
| </Segment> |
| <Segment StartTime="1097.523" EndTime="1098.067" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1098.503" EndTime="1101.666" Participant="mn017"> |
| It seems to me you want to keep the frame-level stuff separate. |
| </Segment> |
| <Segment StartTime="1101.753" EndTime="1109.533" Participant="fe016"> |
| Yeah. I - I definitely agree and I wanted to find actually a f- a nicer format or a - maybe a more compact format than what we used before. Just cuz you've got |
| </Segment> |
| <Segment StartTime="1102.178" EndTime="1103.015" Participant="mn017"> |
| And then - |
| </Segment> |
| <Segment StartTime="1104.612" EndTime="1105.695" Participant="mn017"> |
| <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1107.614" EndTime="1108.189" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1109.533" EndTime="1109.922" Participant="fe016"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1109.922" EndTime="1115.071" Participant="fe016"> |
| ten channels or whatever and two hours of a meeting. It's - it's a lot of - |
| </Segment> |
| <Segment StartTime="1112.086" EndTime="1115.268" Participant="mn017"> |
| <VocalSound Description="breath"/> <VocalSound Description="mouth"/> Now - now how would you - |
| </Segment> |
| <Segment StartTime="1113.903" EndTime="1114.441" Participant="me011"> |
| Huge. <Comment Description="completing [fe016]'s utterance"/> |
| </Segment> |
| <Segment StartTime="1115.603" EndTime="1123.283" Participant="mn017"> |
| how would you represent, um, multiple speakers in this framework? Were - You would just represent them as - <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1121.451" EndTime="1122.280" Participant="me011"> |
| Um, |
| </Segment> |
| <Segment StartTime="1123.283" EndTime="1125.790" Participant="mn017"> |
| You would have like a speaker tag or something? |
| </Segment> |
| <Segment StartTime="1124.932" EndTime="1128.679" Participant="me011"> |
| <VocalSound Description="inbreath"/> there's a spea- speaker tag up at the top which identifies them and then each utt- |
| </Segment> |
| <Segment StartTime="1128.679" EndTime="1135.740" Participant="me011"> |
| the way I had it is each turn or each utterance, <Comment Description="rising intonation"/> I don't even remember now, had a speaker I_D tag attached to it. |
| </Segment> |
| <Segment StartTime="1129.685" EndTime="1130.215" Participant="mn017"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="1136.140" EndTime="1136.782" Participant="mn017"> |
| O_K. |
| </Segment> |
| <Segment StartTime="1136.156" EndTime="1139.530" Participant="me011"> |
| <VocalSound Description="inbreath"/> And in <Emphasis> this </Emphasis> format you would have a different <Emphasis> tag, </Emphasis> |
| </Segment> |
| <Segment StartTime="1140.120" EndTime="1143.049" Participant="me011"> |
| which - which would, uh, be linked to the link. |
| </Segment> |
| <Segment StartTime="1143.049" EndTime="1143.959" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1143.823" EndTime="1144.337" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1143.959" EndTime="1148.530" Participant="me011"> |
| So - so somewhere else you would have another thing <Pause/> that would be, um - |
| </Segment> |
| <Segment StartTime="1150.243" EndTime="1152.656" Participant="me011"> |
| Let's see, would it be a node or a link? Um - |
| </Segment> |
| <Segment StartTime="1151.075" EndTime="1156.491" Participant="mn017"> |
| <VocalSound Description="mouth"/> <VocalSound Description="yawn"/> |
| </Segment> |
| <Segment StartTime="1153.519" EndTime="1158.202" Participant="me011"> |
| And so - so this one would have, um, an I_D is link - <Comment Description="while writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="1158.881" EndTime="1164.855" Participant="me011"> |
| link seventy-four or something like that. And then somewhere up here you would have a link that - that, uh, |
| </Segment> |
| <Segment StartTime="1160.933" EndTime="1161.522" Participant="mn017"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="1166.048" EndTime="1172.847" Participant="me011"> |
| you know, was referencing L_seventy-four and had speaker Adam. <Comment Description="while writing on the whiteboard"/> |
| </Segment> |
| <Segment StartTime="1173.250" EndTime="1174.116" Participant="mn017"> |
| <VocalSound Description="breath"/> Is i- ? |
| </Segment> |
| <Segment StartTime="1173.408" EndTime="1175.578" Participant="fe016"> |
| Actually, it's the <Emphasis> channel, </Emphasis> I think, that - |
| </Segment> |
| <Segment StartTime="1173.578" EndTime="1175.280" Participant="me011"> |
| You know, or something like that. |
| </Segment> |
| <Segment StartTime="1176.131" EndTime="1179.658" Participant="fe016"> |
| I mean, w- yeah, channel is what the channelized output out- |
| </Segment> |
| <Segment StartTime="1176.133" EndTime="1178.590" Participant="mn017"> |
| Well, channel or speaker or whatever. It doesn't - |
| </Segment> |
| <Segment StartTime="1179.068" EndTime="1181.460" Participant="me011"> |
| <Emphasis> This </Emphasis> isn't quite right. I have to look at it again. |
| </Segment> |
| <Segment StartTime="1180.154" EndTime="1184.294" Participant="mn017"> |
| Right. <VocalSound Description="breath"/> But - but - so how in the NIST format do we express <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1180.212" EndTime="1181.241" Participant="fe016"> |
| Yeah, but - |
| </Segment> |
| <Segment StartTime="1183.622" EndTime="1184.087" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1183.950" EndTime="1184.920" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1184.294" EndTime="1190.445" Participant="mn017"> |
| a hierarchical relationship between, um, say, an utterance and the words within it? |
| </Segment> |
| <Segment StartTime="1190.445" EndTime="1196.540" Participant="mn017"> |
| So how do you <Pause/> tell <Pause/> that <Pause/> these are the words that belong to that utterance? |
| </Segment> |
| <Segment StartTime="1196.821" EndTime="1203.990" Participant="me011"> |
| Um, you would have another structure lower down than this that would be saying they're all belonging to this I_D. |
| </Segment> |
| <Segment StartTime="1197.046" EndTime="1199.834" Participant="me018"> |
| <VocalSound Description="yawn"/> |
| </Segment> |
| <Segment StartTime="1200.250" EndTime="1200.949" Participant="mn017"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="1203.541" EndTime="1204.159" Participant="me018"> |
| <NonVocalSound Description="squeak noise"/> |
| </Segment> |
| <Segment StartTime="1205.462" EndTime="1209.682" Participant="me018"> |
| So each thing refers to the <Pause/> utterance that it belongs to. |
| </Segment> |
| <Segment StartTime="1209.371" EndTime="1209.958" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1210.878" EndTime="1216.160" Participant="me011"> |
| And then each utterance could refer to a turn, and each turn could refer to something higher up. |
| </Segment> |
| <Segment StartTime="1210.970" EndTime="1213.299" Participant="me018"> |
| So it's - it's not hi- it's sort of bottom-up. |
| </Segment> |
| <Segment StartTime="1215.828" EndTime="1217.675" Participant="fe016"> |
| <VocalSound Description="inbreath"/> And what if you actually have - |
| </Segment> |
| <Segment StartTime="1217.935" EndTime="1223.672" Participant="fe016"> |
| So right now what <Emphasis> you </Emphasis> have as <Emphasis> utterance, </Emphasis> um, the closest thing that comes out of the <Emphasis> channelized </Emphasis> is the |
| </Segment> |
| <Segment StartTime="1223.973" EndTime="1228.664" Participant="fe016"> |
| stuff between the segment boundaries that the transcribers put in or that Thilo put in, |
| </Segment> |
| <Segment StartTime="1229.173" EndTime="1236.008" Participant="fe016"> |
| which may or may not actually be, like, a s- it's usually <Emphasis> not </Emphasis> - um, the beginning and end of a sentence, say. |
| </Segment> |
| <Segment StartTime="1235.910" EndTime="1238.932" Participant="me011"> |
| Well, that's why I didn't call it "sentence". <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1236.604" EndTime="1240.553" Participant="fe016"> |
| So, right. <VocalSound Description="laugh"/> Um, so it's like a segment or something. So, |
| </Segment> |
| <Segment StartTime="1239.327" EndTime="1239.762" Participant="me011"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1241.261" EndTime="1247.103" Participant="fe016"> |
| I mean, I assume this is possible, that if you have - someone annotates the punctuation or whatever when they transcribe, |
| </Segment> |
| <Segment StartTime="1243.587" EndTime="1243.924" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1247.513" EndTime="1248.997" Participant="fe016"> |
| you can say, you know, from - |
| </Segment> |
| <Segment StartTime="1248.675" EndTime="1249.704" Participant="mn017"> |
| <VocalSound Description="yawn"/> |
| </Segment> |
| <Segment StartTime="1249.480" EndTime="1249.916" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1249.551" EndTime="1256.110" Participant="fe016"> |
| for - from the c- beginning of the sentence to the end of the sentence, from the annotations, this is a unit, even though it <Emphasis> never </Emphasis> actually - |
| </Segment> |
| <Segment StartTime="1256.827" EndTime="1260.862" Participant="fe016"> |
| i- It's only a unit by virtue of the annotations <Pause/> at the <Emphasis> word-level. </Emphasis> |
| </Segment> |
| <Segment StartTime="1260.898" EndTime="1263.473" Participant="me011"> |
| Sure. I mean, so you would - you would have yet another tag. |
| </Segment> |
| <Segment StartTime="1261.460" EndTime="1264.801" Participant="fe016"> |
| And then that would get a tag somehow. O_K. |
| </Segment> |
| <Segment StartTime="1264.128" EndTime="1266.817" Participant="me011"> |
| You'd have another tag which says this is of type "sentence". |
| </Segment> |
| <Segment StartTime="1266.561" EndTime="1267.273" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="1267.300" EndTime="1268.953" Participant="me011"> |
| And, what - <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1267.955" EndTime="1270.467" Participant="fe016"> |
| But it's just not overtly in the - |
| </Segment> |
| <Segment StartTime="1269.104" EndTime="1270.473" Participant="mn017"> |
| <VocalSound Description="inbreath"/> O_K. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="1271.931" EndTime="1273.770" Participant="fe016"> |
| Um, cuz this is exactly the kind of - |
| </Segment> |
| <Segment StartTime="1273.058" EndTime="1273.925" Participant="mn017"> |
| So - |
| </Segment> |
| <Segment StartTime="1274.977" EndTime="1277.696" Participant="fe016"> |
| I think that should be <Pause/> <Emphasis> possible </Emphasis> as long as the - |
| </Segment> |
| <Segment StartTime="1278.056" EndTime="1283.680" Participant="fe016"> |
| But, uh, what I don't understand is where the - where in this type of file <Pause/> that would be expressed. |
| </Segment> |
| <Segment StartTime="1283.820" EndTime="1284.597" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1284.720" EndTime="1286.707" Participant="me011"> |
| Right. You would have another <Emphasis> tag </Emphasis> somewhere. |
| </Segment> |
| <Segment StartTime="1286.707" EndTime="1289.300" Participant="me011"> |
| <VocalSound Description="breath"/> It's - well, there're <Emphasis> two </Emphasis> ways of doing it. |
| </Segment> |
| <Segment StartTime="1287.221" EndTime="1292.320" Participant="fe016"> |
| S- so it would just be floating before the sentence or floating after the sentence without a time-mark. |
| </Segment> |
| <Segment StartTime="1290.642" EndTime="1292.928" Participant="me011"> |
| You could have some sort of link type - |
| </Segment> |
| <Segment StartTime="1292.928" EndTime="1293.956" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="1293.956" EndTime="1295.361" Participant="me011"> |
| type equals "sentence", |
| </Segment> |
| <Segment StartTime="1295.361" EndTime="1296.798" Participant="me011"> |
| <NonVocalSound Description="writing on whiteboard"/> |
| </Segment> |
| <Segment StartTime="1296.798" EndTime="1300.121" Participant="me011"> |
| <VocalSound Description="inbreath"/> and I_D is "S_whatever". |
| </Segment> |
| <Segment StartTime="1300.121" EndTime="1305.345" Participant="me011"> |
| And then lower down you could have an utterance. So the type is "utterance" - |
| </Segment> |
| <Segment StartTime="1306.471" EndTime="1308.770" Participant="me011"> |
| equals "utt". And you could either say that - |
| </Segment> |
| <Segment StartTime="1309.502" EndTime="1311.460" Participant="me011"> |
| No. I don't know - I take that back. |
| </Segment> |
| <Segment StartTime="1310.382" EndTime="1312.627" Participant="mn017"> |
| <VocalSound Description="mouth"/> So here's the thing. Um - |
| </Segment> |
| <Segment StartTime="1312.901" EndTime="1313.860" Participant="fe016"> |
| See, cuz it's - |
| </Segment> |
| <Segment StartTime="1313.183" EndTime="1316.130" Participant="me011"> |
| Can you - can you say that this is |
| </Segment> |
| <Segment StartTime="1315.172" EndTime="1316.094" Participant="mn017"> |
| Hhh. <Comment Description="vocal gesture"/> |
| </Segment> |
| <Segment StartTime="1315.681" EndTime="1316.744" Participant="fe016"> |
| <Uncertain> it's - </Uncertain> <Comment Description="whispered"/> |
| </Segment> |
| <Segment StartTime="1316.286" EndTime="1317.907" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1316.733" EndTime="1317.714" Participant="me018"> |
| You would just have a r- |
| </Segment> |
| <Segment StartTime="1317.179" EndTime="1319.789" Participant="me011"> |
| part of this, or do you say this is part of this? I think - |
| </Segment> |
| <Segment StartTime="1317.372" EndTime="1318.055" Participant="fe016"> |
| S- |
| </Segment> |
| <Segment StartTime="1319.804" EndTime="1320.433" Participant="fe016"> |
| But they're - |
| </Segment> |
| <Segment StartTime="1319.881" EndTime="1321.590" Participant="me018"> |
| You would refer up to the sentence. |
| </Segment> |
| <Segment StartTime="1320.722" EndTime="1321.640" Participant="mn017"> |
| Well, the thing - |
| </Segment> |
| <Segment StartTime="1321.087" EndTime="1323.660" Participant="fe016"> |
| they're actually overlapping each other, sort of. |
| </Segment> |
| <Segment StartTime="1321.728" EndTime="1322.242" Participant="me011"> |
| So - |
| </Segment> |
| <Segment StartTime="1321.986" EndTime="1322.403" Participant="me018"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1323.042" EndTime="1326.181" Participant="mn017"> |
| the thing is that some- something may be a part of |
| </Segment> |
| <Segment StartTime="1323.310" EndTime="1325.013" Participant="me011"> |
| Right. <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1326.687" EndTime="1329.627" Participant="mn017"> |
| one thing for one purpose and another thing of another purpose. |
| </Segment> |
| <Segment StartTime="1330.019" EndTime="1332.195" Participant="fe016"> |
| You have to have another type then, I guess. |
| </Segment> |
| <Segment StartTime="1330.205" EndTime="1330.713" Participant="mn017"> |
| So f- |
| </Segment> |
| <Segment StartTime="1331.428" EndTime="1331.934" Participant="mn017"> |
| s- |
| </Segment> |
| <Segment StartTime="1332.388" EndTime="1335.580" Participant="mn017"> |
| Um, well, <Comment Description="rising intonation"/> s- let's - let's ta- so let's - |
| </Segment> |
| <Segment StartTime="1333.225" EndTime="1337.537" Participant="me011"> |
| <VocalSound Description="mouth"/> Well, I think I'm - I think w- I had better look at it again because I - I'm - |
| </Segment> |
| <Segment StartTime="1333.805" EndTime="1334.450" Participant="fe016"> |
| Yeah. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="1336.171" EndTime="1337.476" Participant="fe016"> |
| O_K. O_K. |
| </Segment> |
| <Segment StartTime="1336.392" EndTime="1337.150" Participant="mn017"> |
| so - |
| </Segment> |
| <Segment StartTime="1337.743" EndTime="1340.167" Participant="mn017"> |
| y- So for instance <Uncertain NumSyllables="2"> @@ </Uncertain> <Comment Description="uncodeable vocal sounds"/> sup- |
| </Segment> |
| <Segment StartTime="1337.814" EndTime="1341.889" Participant="me011"> |
| <VocalSound Description="breath"/> There's one level - there's one more level of indirection that I'm forgetting. |
| </Segment> |
| <Segment StartTime="1340.975" EndTime="1349.690" Participant="mn017"> |
| Suppose you have a word sequence and you have two different segmentations of that same word sequence. f- Say, one segmentation is in terms of, |
| </Segment> |
| <Segment StartTime="1344.675" EndTime="1349.814" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1350.926" EndTime="1351.710" Participant="mn017"> |
| um, |
| </Segment> |
| <Segment StartTime="1352.482" EndTime="1359.539" Participant="mn017"> |
| you know, uh, <Emphasis> sentences. </Emphasis> And another segmentation is in terms of, um, <VocalSound Description="mouth"/> I don't know, <Comment Description="embedded"/> prosodic <Emphasis> phrases. </Emphasis> |
| </Segment> |
| <Segment StartTime="1360.113" EndTime="1365.440" Participant="mn017"> |
| And let's say that they don't <Pause/> nest. So, you know, a prosodic phrase may cross |
| </Segment> |
| <Segment StartTime="1365.096" EndTime="1365.650" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1366.470" EndTime="1369.932" Participant="mn017"> |
| two sentences or something. I don't know if that's true or not but <VocalSound Description="breath"/> let's as- |
| </Segment> |
| <Segment StartTime="1368.801" EndTime="1374.754" Participant="fe016"> |
| Well, it's definitely true with the segment. That's what I - exactly what I meant by the utterances versus the sentence could be sort of - |
| </Segment> |
| <Segment StartTime="1370.575" EndTime="1370.914" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="1371.237" EndTime="1371.900" Participant="mn017"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1373.221" EndTime="1380.310" Participant="mn017"> |
| So, you want to be s- you want to say this - this <Emphasis> word </Emphasis> is part of that sentence and this prosodic phrase. |
| </Segment> |
| <Segment StartTime="1375.209" EndTime="1375.819" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1380.944" EndTime="1381.299" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1381.028" EndTime="1384.850" Participant="mn017"> |
| But the phrase is not part of the sentence and neither is the sentence part of the phrase. |
| </Segment> |
| <Segment StartTime="1383.576" EndTime="1384.189" Participant="fe016"> |
| Right. |
| </Segment> |
| <Segment StartTime="1384.107" EndTime="1388.480" Participant="me011"> |
| I- I'm pretty sure that you can do that, but I'm forgetting the exact level of nesting. |
| </Segment> |
| <Segment StartTime="1386.282" EndTime="1392.870" Participant="mn017"> |
| So, you would have to have <VocalSound Description="breath"/> two different pointers from the word up - one level up, one to the sent- |
| </Segment> |
| <Segment StartTime="1391.180" EndTime="1396.412" Participant="me011"> |
| <VocalSound Description="inbreath"/> So - so what you would end up having is a tag saying "here's a word, and it starts here and it ends here". <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1394.959" EndTime="1396.282" Participant="mn017"> |
| <Uncertain NumSyllables="1"> @@ </Uncertain> Right. |
| </Segment> |
| <Segment StartTime="1396.412" EndTime="1400.409" Participant="me011"> |
| And then lower down you would say "here's a prosodic boundary and it has these words in it". |
| </Segment> |
| <Segment StartTime="1400.920" EndTime="1404.120" Participant="me011"> |
| And lower down you'd have "here's a <Emphasis> sentence, </Emphasis> and it has these words in it". |
| </Segment> |
| <Segment StartTime="1401.399" EndTime="1402.494" Participant="fe016"> |
| An- Right. |
| </Segment> |
| <Segment StartTime="1401.761" EndTime="1402.483" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="1403.016" EndTime="1408.698" Participant="fe016"> |
| So <Emphasis> you </Emphasis> would be able to go in and say, you know, "give me all the words in the bound- in the prosodic phrase and give me all the words in the -" <Comment Description="abandoned as clear"/> Yeah. |
| </Segment> |
| <Segment StartTime="1407.821" EndTime="1408.403" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1409.179" EndTime="1410.382" Participant="mn017"> |
| <VocalSound Description="inbreath"/> <Comment Description="as if to speak"/> <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1409.324" EndTime="1411.810" Participant="me011"> |
| So I think that's - that would wor- Let me look at it again. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1409.432" EndTime="1410.578" Participant="fe016"> |
| Um, O_K. |
| </Segment> |
| <Segment StartTime="1411.137" EndTime="1411.928" Participant="mn017"> |
| Mm-hmm. <Comment Description="quietly"/> <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1411.507" EndTime="1412.279" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="1411.928" EndTime="1418.497" Participant="mn017"> |
| The - the o- the <Emphasis> other </Emphasis> issue that <Emphasis> you </Emphasis> had was, how do you actually efficiently extract, um - |
| </Segment> |
| <Segment StartTime="1412.150" EndTime="1412.699" Participant="me011"> |
| So. |
| </Segment> |
| <Segment StartTime="1413.278" EndTime="1414.170" Participant="fe016"> |
| That's good. <Comment Description="fading"/> |
| </Segment> |
| <Segment StartTime="1419.557" EndTime="1424.060" Participant="mn017"> |
| find and extract information in a structure of this type? |
| </Segment> |
| <Segment StartTime="1424.508" EndTime="1427.150" Participant="mn017"> |
| So you gave some examples like - |
| </Segment> |
| <Segment StartTime="1426.884" EndTime="1430.004" Participant="fe016"> |
| <VocalSound Description="inbreath"/> Well, uh, and, I mean, you guys might - |
| </Segment> |
| <Segment StartTime="1430.833" EndTime="1438.346" Participant="fe016"> |
| I don't know if this is premature because I suppose once you get the representation you can <Emphasis> do </Emphasis> this, but the kinds of things I was worried about is, |
| </Segment> |
| <Segment StartTime="1439.126" EndTime="1443.128" Participant="mn017"> |
| No, that's not clear. I mean, yeah, you c- sure you can do it, but can you do it |
| </Segment> |
| <Segment StartTime="1439.150" EndTime="1439.782" Participant="fe016"> |
| uh - |
| </Segment> |
| <Segment StartTime="1440.217" EndTime="1442.240" Participant="fe016"> |
| Well, O_K. So i- if it - |
| </Segment> |
| <Segment StartTime="1443.310" EndTime="1446.793" Participant="fe016"> |
| I- I mean, <Emphasis> I </Emphasis> can't do it, but I can - <VocalSound Description="laugh"/> um, |
| </Segment> |
| <Segment StartTime="1443.833" EndTime="1445.838" Participant="mn017"> |
| sort of l- l- you know, |
| </Segment> |
| <Segment StartTime="1444.993" EndTime="1446.503" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1446.067" EndTime="1446.269" Participant="mn017"> |
| it - |
| </Segment> |
| <Segment StartTime="1446.721" EndTime="1450.877" Participant="mn017"> |
| y- y- you gotta - you gotta do this - you - you're gonna want to do this very quickly |
| </Segment> |
| <Segment StartTime="1447.074" EndTime="1447.588" Participant="me011"> |
| Well - |
| </Segment> |
| <Segment StartTime="1450.877" EndTime="1456.275" Participant="mn017"> |
| or else you'll spend all your time sort of searching through very <VocalSound Description="breath"/> complex data structures - |
| </Segment> |
| <Segment StartTime="1453.953" EndTime="1458.592" Participant="fe016"> |
| Right. You'd need a p- sort of a paradigm for how to do it. But an example would be |
| </Segment> |
| <Segment StartTime="1460.267" EndTime="1463.568" Participant="fe016"> |
| "find all the cases in which Adam |
| </Segment> |
| <Segment StartTime="1465.110" EndTime="1472.280" Participant="fe016"> |
| started to talk while Andreas was talking and his pitch was rising, <Emphasis> Andreas's </Emphasis> pitch". <Emphasis> That </Emphasis> kind of thing. |
| </Segment> |
| <Segment StartTime="1472.424" EndTime="1473.769" Participant="me011"> |
| <VocalSound Description="mouth"/> Right. I mean, that's gonna be - |
| </Segment> |
| <Segment StartTime="1474.224" EndTime="1478.932" Participant="me011"> |
| Is the rising pitch a <Pause/> <Emphasis> feature, </Emphasis> or is it gonna be in the same file? |
| </Segment> |
| <Segment StartTime="1478.585" EndTime="1486.820" Participant="fe016"> |
| Well, the rising pitch will never be <Pause/> hand-annotated. So the - all the prosodic features are going to be automatically - So they're gonna be in those - |
| </Segment> |
| <Segment StartTime="1484.140" EndTime="1485.072" Participant="me011"> |
| <VocalSound Description="mouth"/> But the - |
| </Segment> |
| <Segment StartTime="1485.428" EndTime="1492.449" Participant="me011"> |
| I mean, that's gonna be hard regardless, right? Because you're gonna have to write a program that goes through your feature file and looks for rising pitches. |
| </Segment> |
| <Segment StartTime="1489.624" EndTime="1490.138" Participant="mn017"> |
| <Pronounce Pronunciation=""n-yeah""> Yeah. </Pronounce> |
| </Segment> |
| <Segment StartTime="1491.097" EndTime="1494.462" Participant="fe016"> |
| <VocalSound Description="inbreath"/> So - Right. So normally what we would do is we would say |
| </Segment> |
| <Segment StartTime="1494.462" EndTime="1499.151" Participant="fe016"> |
| "what do we wanna assign rising pitch to?" Are we gonna assign it to <Emphasis> words? </Emphasis> Are we gonna just assign it to sort of - |
| </Segment> |
| <Segment StartTime="1499.466" EndTime="1504.763" Participant="fe016"> |
| when it's rising we have a begin-end <Emphasis> rise </Emphasis> representation? But suppose we dump out this file and we say, |
| </Segment> |
| <Segment StartTime="1499.927" EndTime="1500.876" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1504.763" EndTime="1509.589" Participant="fe016"> |
| uh, for every word we just classify it as, w- you know, rise or fall or neither? |
| </Segment> |
| <Segment StartTime="1509.552" EndTime="1513.270" Participant="me011"> |
| O_K. Well, in that case you would add that to this <Pause/> format r- |
| </Segment> |
| <Segment StartTime="1510.024" EndTime="1510.507" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="1512.736" EndTime="1514.713" Participant="fe016"> |
| So we would basically be sort of, um, |
| </Segment> |
| <Segment StartTime="1515.563" EndTime="1518.939" Participant="fe016"> |
| taking the format and enriching it with things that we wanna |
| </Segment> |
| <Segment StartTime="1518.990" EndTime="1519.487" Participant="mn017"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1519.330" EndTime="1524.375" Participant="fe016"> |
| query in relation to the words that are already in the file, and then querying it. O_K. |
| </Segment> |
| <Segment StartTime="1521.954" EndTime="1522.508" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1523.860" EndTime="1527.377" Participant="mn017"> |
| <VocalSound Description="inbreath"/> You want sort of a grep that's - that works at the structural - |
| </Segment> |
| <Segment StartTime="1527.302" EndTime="1528.311" Participant="me011"> |
| <VocalSound Description="mouth"/> You <Emphasis> have </Emphasis> that. |
| </Segment> |
| <Segment StartTime="1527.832" EndTime="1529.505" Participant="mn017"> |
| on the structural representation. <Comment Description="faster tempo on "representation""/> |
| </Segment> |
| <Segment StartTime="1528.780" EndTime="1538.940" Participant="me011"> |
| There's a <Pause/> standard again in X_M_L, specifically for searching X_M_L documents - structured X_- X_M_L documents, where you can specify both the content and the structural position. |
| </Segment> |
| <Segment StartTime="1539.572" EndTime="1542.166" Participant="mn017"> |
| <VocalSound Description="inbreath"/> Yeah, but it's - it's not clear that that's - |
| </Segment> |
| <Segment StartTime="1542.490" EndTime="1543.200" Participant="fe016"> |
| <VocalSound Description="inbreath"/> If - |
| </Segment> |
| <Segment StartTime="1542.885" EndTime="1548.340" Participant="mn017"> |
| That's relative to the structure of the X_M_L document, not to the structure of what you're <Emphasis> representing </Emphasis> in the document. |
| </Segment> |
| <Segment StartTime="1545.289" EndTime="1546.930" Participant="me011"> |
| You use it as a tool. |
| </Segment> |
| <Segment StartTime="1546.930" EndTime="1554.110" Participant="me011"> |
| You use it as a tool, not an end-user. It's not an end-user thing. It's - it's - you would use that to build your tool to do that sort of search. |
| </Segment> |
| <Segment StartTime="1549.741" EndTime="1550.379" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="1552.427" EndTime="1552.771" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="1553.414" EndTime="1553.849" Participant="mn017"> |
| <Uncertain> Be- </Uncertain> |
| </Segment> |
| <Segment StartTime="1554.015" EndTime="1554.950" Participant="fe016"> |
| <VocalSound Description="inbreath"/> Uh - |
| </Segment> |
| <Segment StartTime="1554.182" EndTime="1558.041" Participant="mn017"> |
| Because <Emphasis> here </Emphasis> you're specifying a <Emphasis> lattice. </Emphasis> So the <Pronounce Pronunciation=""underslying""> underlying </Pronounce> - |
| </Segment> |
| <Segment StartTime="1558.041" EndTime="1562.495" Participant="mn017"> |
| that's the underlying data structure. And you want to be able to search in that lattice. |
| </Segment> |
| <Segment StartTime="1558.589" EndTime="1559.143" Participant="fe016"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1559.580" EndTime="1561.222" Participant="fe016"> |
| But as long as the - |
| </Segment> |
| <Segment StartTime="1561.580" EndTime="1562.927" Participant="me011"> |
| It's a graph, but - |
| </Segment> |
| <Segment StartTime="1563.204" EndTime="1565.800" Participant="mn017"> |
| That's <Emphasis> different </Emphasis> from searching through the text. |
| </Segment> |
| <Segment StartTime="1564.367" EndTime="1566.938" Participant="fe016"> |
| <Uncertain> But </Uncertain> it seems like as long as the features that - |
| </Segment> |
| <Segment StartTime="1566.090" EndTime="1570.542" Participant="me011"> |
| Well, no, no, no. The whole <Emphasis> point </Emphasis> is that the text and the lattice are <Emphasis> isomorphic. </Emphasis> |
| </Segment> |
| <Segment StartTime="1570.970" EndTime="1573.726" Participant="me011"> |
| They <Pause/> represent each other <Pause/> <Emphasis> completely. </Emphasis> |
| </Segment> |
| <Segment StartTime="1573.889" EndTime="1574.661" Participant="mn017"> |
| Um - |
| </Segment> |
| <Segment StartTime="1575.073" EndTime="1575.868" Participant="me011"> |
| So that - |
| </Segment> |
| <Segment StartTime="1577.410" EndTime="1585.918" Participant="fe016"> |
| That's true if the features from your acoustics or whatever that are not explicitly <Emphasis> in </Emphasis> this are at the level of these types. |
| </Segment> |
| <Segment StartTime="1577.525" EndTime="1578.177" Participant="me011"> |
| I mean th- |
| </Segment> |
| <Segment StartTime="1578.551" EndTime="1579.750" Participant="mn017"> |
| Hhh. <Comment Description="vocal gesture"/> |
| </Segment> |
| <Segment StartTime="1586.235" EndTime="1587.402" Participant="fe016"> |
| That - that if you can |
| </Segment> |
| <Segment StartTime="1586.989" EndTime="1587.917" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1587.300" EndTime="1589.405" Participant="me011"> |
| Yeah, but that's gonna be the trouble no matter what. |
| </Segment> |
| <Segment StartTime="1588.288" EndTime="1589.004" Participant="fe016"> |
| do that - |
| </Segment> |
| <Segment StartTime="1589.741" EndTime="1594.028" Participant="me011"> |
| Right? No matter what format you choose, you're gonna have the trou- you're gonna have the difficulty of relating |
| </Segment> |
| <Segment StartTime="1594.020" EndTime="1600.606" Participant="fe016"> |
| That's right. That's true. That's why I was trying to figure out what's the best format for this representation. And it's <Emphasis> still </Emphasis> gonna be - |
| </Segment> |
| <Segment StartTime="1594.375" EndTime="1597.132" Participant="me011"> |
| the - the frame-level features - |
| </Segment> |
| <Segment StartTime="1599.167" EndTime="1599.662" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1600.358" EndTime="1601.819" CloseMic="false"> |
| <NonVocalSound Description="click noise, perhaps door closes"/> |
| </Segment> |
| <Segment StartTime="1601.931" EndTime="1604.726" Participant="fe016"> |
| it's still gonna be, uh, not direct. You know, it - |
| </Segment> |
| <Segment StartTime="1602.289" EndTime="1602.899" Participant="mn017"> |
| Hmm. |
| </Segment> |
| <Segment StartTime="1604.089" EndTime="1604.663" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1605.576" EndTime="1611.582" Participant="fe016"> |
| Or another example was, you know, uh, where in the language - where in the word sequence |
| </Segment> |
| <Segment StartTime="1606.640" EndTime="1607.528" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1607.761" EndTime="1608.047" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1608.227" EndTime="1608.550" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1610.035" EndTime="1611.093" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1612.291" EndTime="1614.790" Participant="fe016"> |
| are people interrupting? So, |
| </Segment> |
| <Segment StartTime="1615.799" EndTime="1617.282" Participant="fe016"> |
| I guess that one's actually easier. |
| </Segment> |
| <Segment StartTime="1618.385" EndTime="1621.207" Participant="me018"> |
| <VocalSound Description="inbreath"/> What about - what about, um, |
| </Segment> |
| <Segment StartTime="1621.784" EndTime="1632.352" Participant="me018"> |
| the idea of using a relational database to, uh, store the information from the X_M_L? So you would have - X_M_L basically would - |
| </Segment> |
| <Segment StartTime="1622.154" EndTime="1624.528" CloseMic="false"> |
| <NonVocalSound Description="clinking noise"/> |
| </Segment> |
| <Segment StartTime="1625.389" EndTime="1627.468" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1627.127" EndTime="1627.811" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1633.576" EndTime="1637.246" Participant="me018"> |
| Uh, you - you could use the X_M_L to put the data <Emphasis> in, </Emphasis> |
| </Segment> |
| <Segment StartTime="1637.246" EndTime="1640.916" Participant="me018"> |
| and then when you get data out, you put it back in X_M_L. So use X_M_L as sort of the - |
| </Segment> |
| <Segment StartTime="1640.235" EndTime="1641.048" Participant="me011"> |
| Transfer. <Comment Description="completing [me018]'s utterance"/> |
| </Segment> |
| <Segment StartTime="1640.916" EndTime="1644.707" Participant="me018"> |
| the transfer format, uh, but then you store the data in the database, which |
| </Segment> |
| <Segment StartTime="1645.369" EndTime="1646.360" Participant="me011"> |
| The, uh - |
| </Segment> |
| <Segment StartTime="1645.414" EndTime="1648.769" Participant="me018"> |
| allows you to do all kinds of <Pause/> good search things in there. |
| </Segment> |
| <Segment StartTime="1648.269" EndTime="1654.070" Participant="me011"> |
| One of the things that ATLAS is doing is they're trying to define an A_P_I which is independent of the back store, |
| </Segment> |
| <Segment StartTime="1649.452" EndTime="1649.940" Participant="fe016"> |
| Huh. |
| </Segment> |
| <Segment StartTime="1654.945" EndTime="1661.463" Participant="me011"> |
| so that, uh, you could define a single A_P_I and the - the storage could be flat X_M_L files or a database. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1658.967" EndTime="1660.587" Participant="me018"> |
| Mm-hmm. <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1661.463" EndTime="1669.079" Participant="me011"> |
| <Emphasis> My </Emphasis> opinion on that is for the s- sort of stuff that we're doing, <Comment Description="rising intonation"/> I suspect it's overkill to do a full relational database, that, um, |
| </Segment> |
| <Segment StartTime="1668.579" EndTime="1670.350" Participant="mn017"> |
| But - <Comment Description="quietly"/> <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1670.160" EndTime="1674.961" Participant="me011"> |
| just a flat file and, uh, search tools I <Emphasis> bet </Emphasis> will be enough. |
| </Segment> |
| <Segment StartTime="1675.776" EndTime="1678.388" Participant="me011"> |
| But that's the advantage of ATLAS, is that if we actually take - |
| </Segment> |
| <Segment StartTime="1679.026" EndTime="1686.010" Participant="me011"> |
| decide to go that route completely and we program to their A_P_I, then if we wanted to add a database later it would be pretty easy. |
| </Segment> |
| <Segment StartTime="1680.447" EndTime="1681.086" Participant="me018"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="1682.790" EndTime="1683.357" Participant="me018"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="1688.042" EndTime="1696.134" Participant="fe016"> |
| It seems like the kind of thing you'd do if - I don't know, if people start adding all kinds of s- bells and whistles to the data. And so that might be - |
| </Segment> |
| <Segment StartTime="1691.586" EndTime="1694.510" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1696.134" EndTime="1699.134" Participant="fe016"> |
| I mean, it'd be good for us to know - to use a format where |
| </Segment> |
| <Segment StartTime="1700.052" EndTime="1705.235" Participant="fe016"> |
| we know we can easily, um, input that to some database if other people are using it. |
| </Segment> |
| <Segment StartTime="1703.931" EndTime="1704.504" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="1705.551" EndTime="1706.590" Participant="fe016"> |
| Something like that. |
| </Segment> |
| <Segment StartTime="1706.264" EndTime="1711.634" Participant="me011"> |
| <VocalSound Description="inbreath"/> I guess I'm just a <Emphasis> little </Emphasis> hesitant to try to go whole hog on sort of the - |
| </Segment> |
| <Segment StartTime="1707.317" EndTime="1708.086" Participant="fe016"> |
| So - <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="1712.626" EndTime="1718.315" Participant="me011"> |
| the whole framework that - that NIST is talking about, with ATLAS <Emphasis> and </Emphasis> a database and all that sort of stuff, cuz it's a big learning curve, |
| </Segment> |
| <Segment StartTime="1718.599" EndTime="1719.055" Participant="me018"> |
| Hmm. |
| </Segment> |
| <Segment StartTime="1719.165" EndTime="1722.033" Participant="me011"> |
| just to get going. Whereas if we <Emphasis> just </Emphasis> do a flat file format, |
| </Segment> |
| <Segment StartTime="1719.248" EndTime="1719.913" Participant="mn017"> |
| Hmm. |
| </Segment> |
| <Segment StartTime="1722.986" EndTime="1727.741" Participant="me011"> |
| sure, it may not be as efficient but <Emphasis> everyone </Emphasis> can program in Perl and - and <Emphasis> use </Emphasis> it. |
| </Segment> |
| <Segment StartTime="1726.829" EndTime="1727.887" Participant="fe016"> |
| <Uncertain NumSyllables="1"> @@ </Uncertain> O_K. |
| </Segment> |
| <Segment StartTime="1727.741" EndTime="1729.125" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1729.125" EndTime="1731.140" Participant="me011"> |
| Right? So, as opposed to - |
| </Segment> |
| <Segment StartTime="1729.425" EndTime="1732.901" Participant="mn017"> |
| But this is - I - I'm still, um, <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="1732.901" EndTime="1741.210" Participant="mn017"> |
| not convinced that you can do much at <Emphasis> all </Emphasis> on the text - on the flat file that - that - you know, the text representation. |
| </Segment> |
| <Segment StartTime="1741.210" EndTime="1742.032" Participant="mn017"> |
| e- |
| </Segment> |
| <Segment StartTime="1742.032" EndTime="1744.361" Participant="mn017"> |
| Because the text representation is gonna be, |
| </Segment> |
| <Segment StartTime="1745.337" EndTime="1751.014" Participant="mn017"> |
| uh, not reflecting the structure of - of your words and annotations. |
| </Segment> |
| <Segment StartTime="1751.508" EndTime="1752.758" Participant="mn017"> |
| It's just - it's - |
| </Segment> |
| <Segment StartTime="1751.733" EndTime="1758.495" Participant="me011"> |
| Well, if it's not <Emphasis> representing </Emphasis> it, then how do you <Emphasis> recover </Emphasis> it? Of <Emphasis> course </Emphasis> it's representing it. That's the whole <Emphasis> point. </Emphasis> <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1754.831" EndTime="1758.374" Participant="mn017"> |
| No. You - you have to - what you have to do is you have to basically - <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1758.659" EndTime="1765.265" Participant="mn017"> |
| Y- yeah. You can use Perl to read it in and construct a internal representation that is essentially a lattice. |
| </Segment> |
| <Segment StartTime="1766.531" EndTime="1767.047" Participant="mn017"> |
| But, |
| </Segment> |
| <Segment StartTime="1767.354" EndTime="1768.561" Participant="me018"> |
| <Uncertain NumSyllables="2"> @@ </Uncertain> Yeah. |
| </Segment> |
| <Segment StartTime="1767.355" EndTime="1769.154" Participant="mn017"> |
| the - and then - |
| </Segment> |
| <Segment StartTime="1768.279" EndTime="1771.038" Participant="me011"> |
| O_K. Well, <Emphasis> that </Emphasis> was a different <Emphasis> point. </Emphasis> Right? So what I was <Emphasis> saying </Emphasis> |
| </Segment> |
| <Segment StartTime="1770.045" EndTime="1770.994" Participant="mn017"> |
| Right. <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1771.335" EndTime="1772.820" Participant="me011"> |
| is that - <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="1772.358" EndTime="1773.823" Participant="mn017"> |
| But <Emphasis> that's </Emphasis> what you'll have to <Emphasis> do. </Emphasis> |
| </Segment> |
| <Segment StartTime="1773.986" EndTime="1785.666" Participant="me011"> |
| For Perl - if you want to just do Perl. If you wanted to use the structured X_M_L query language, that's a different thing. And it's a set of tools <VocalSound Description="breath"/> that let you specify given the D_- D_D_T - D_T_D of the document, |
| </Segment> |
| <Segment StartTime="1774.273" EndTime="1774.767" Participant="mn017"> |
| Bec- be- |
| </Segment> |
| <Segment StartTime="1786.278" EndTime="1789.550" Participant="me011"> |
| um, what sorts of structural searches you want to do. |
| </Segment> |
| <Segment StartTime="1790.340" EndTime="1792.643" Participant="me011"> |
| So you want to say that, you know, you're looking for, |
| </Segment> |
| <Segment StartTime="1793.359" EndTime="1798.321" Participant="me011"> |
| um, a tag within a tag within a particular tag that has this particular text in it, |
| </Segment> |
| <Segment StartTime="1798.759" EndTime="1801.914" Participant="me011"> |
| um, and, uh, refers to a particular value. |
| </Segment> |
| <Segment StartTime="1802.794" EndTime="1811.427" Participant="me011"> |
| <VocalSound Description="breath"/> And so the point isn't that an end-user, who is looking for a query like you specified, wouldn't program it in this language. What you would do is, |
| </Segment> |
| <Segment StartTime="1811.427" EndTime="1814.632" Participant="me011"> |
| someone would build a <Emphasis> tool </Emphasis> that used that as a library. |
| </Segment> |
| <Segment StartTime="1816.320" EndTime="1820.603" Participant="me011"> |
| So that they - so that you wouldn't have to construct the internal representations yourself. |
| </Segment> |
| <Segment StartTime="1819.034" EndTime="1819.492" Participant="fe016"> |
| Is a - |
| </Segment> |
| <Segment StartTime="1821.425" EndTime="1822.330" Participant="fe016"> |
| <VocalSound Description="inbreath"/> <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1822.330" EndTime="1827.674" Participant="fe016"> |
| See, I think the kinds of questions, at least in the next - to the end of this year, are - |
| </Segment> |
| <Segment StartTime="1828.261" EndTime="1831.748" Participant="fe016"> |
| there may be a lot of different ones, but they'll all have a similar nature. They'll be |
| </Segment> |
| <Segment StartTime="1832.317" EndTime="1835.829" Participant="fe016"> |
| looking at either a word-level prosodic, |
| </Segment> |
| <Segment StartTime="1834.010" EndTime="1834.861" Participant="me018"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1836.856" EndTime="1841.837" Participant="fe016"> |
| uh, an - a <Emphasis> value, </Emphasis> like a continuous value, like the slope of something. But |
| </Segment> |
| <Segment StartTime="1839.020" EndTime="1839.491" Participant="me011"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="1842.333" EndTime="1844.259" Participant="fe016"> |
| you know, we'll do <Emphasis> something </Emphasis> where we - |
| </Segment> |
| <Segment StartTime="1844.754" EndTime="1850.239" Participant="fe016"> |
| some kind of data reduction where the prosodic features are sort o- uh, either at the word-level or at the |
| </Segment> |
| <Segment StartTime="1850.577" EndTime="1856.420" Participant="fe016"> |
| segment-level, or - or <Emphasis> something </Emphasis> like that. They're not gonna be at the <Emphasis> phone-level </Emphasis> and they're no- not gonna be at the <Emphasis> frame-level </Emphasis> when we get done with |
| </Segment> |
| <Segment StartTime="1850.587" EndTime="1851.239" Participant="mn017"> |
| <VocalSound Description="sniff"/> |
| </Segment> |
| <Segment StartTime="1851.433" EndTime="1851.921" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1853.867" EndTime="1855.272" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1856.923" EndTime="1859.487" Participant="fe016"> |
| sort of giving them simpler shapes and things. |
| </Segment> |
| <Segment StartTime="1860.219" EndTime="1862.958" Participant="fe016"> |
| And so the main thing is just being able - |
| </Segment> |
| <Segment StartTime="1864.930" EndTime="1867.761" Participant="fe016"> |
| Well, I guess, the <Emphasis> two </Emphasis> goals. Um, one that <Emphasis> Chuck </Emphasis> mentioned is |
| </Segment> |
| <Segment StartTime="1867.761" EndTime="1874.025" Participant="fe016"> |
| starting out with something that we don't have to start over, that we don't have to throw away if other people want to extend it for other kinds of questions, |
| </Segment> |
| <Segment StartTime="1872.479" EndTime="1873.033" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1874.025" EndTime="1878.172" Participant="fe016"> |
| <Emphasis> and </Emphasis> being able to at least get enough, uh, information out on - |
| </Segment> |
| <Segment StartTime="1879.770" EndTime="1886.855" Participant="fe016"> |
| where we condition the location of features on information that's in the kind of file that you <Pause/> put up <Emphasis> there. </Emphasis> |
| </Segment> |
| <Segment StartTime="1887.168" EndTime="1889.900" Participant="fe016"> |
| And that would - that would do it, I mean, for <Emphasis> me. </Emphasis> |
| </Segment> |
| <Segment StartTime="1888.733" EndTime="1895.370" Participant="me011"> |
| Yeah. I think that there are quick and dirty solutions, and then there are long-term, big-infrastructure solutions. And so <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1895.370" EndTime="1898.416" Participant="me011"> |
| we want to try to pick something that lets us do a little bit of both. |
| </Segment> |
| <Segment StartTime="1897.428" EndTime="1903.810" Participant="fe016"> |
| In the between, right. And especially that the representation doesn't have to be thrown away, even if your tools change. |
| </Segment> |
| <Segment StartTime="1898.950" EndTime="1899.910" Participant="me011"> |
| Um - |
| </Segment> |
| <Segment StartTime="1901.260" EndTime="1902.000" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="1903.559" EndTime="1907.944" Participant="me011"> |
| <VocalSound Description="inbreath"/> And so it seems to me that - I mean, I have to <Emphasis> look </Emphasis> at it again to see whether it can really do what we <Emphasis> want, </Emphasis> <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1905.222" EndTime="1906.579" Participant="fe016"> |
| <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="1907.944" EndTime="1912.060" Participant="me011"> |
| but if we use the ATLAS external file representation, |
| </Segment> |
| <Segment StartTime="1912.800" EndTime="1913.475" Participant="me011"> |
| um, |
| </Segment> |
| <Segment StartTime="1914.068" EndTime="1918.975" Participant="me011"> |
| it <Emphasis> seems </Emphasis> like it's rich enough that you could do <Emphasis> quick </Emphasis> tools just as I said in Perl, <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="1918.953" EndTime="1919.705" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1918.975" EndTime="1925.305" Participant="me011"> |
| and then later on if we choose to go up the learning curve, we can use the whole ATLAS inter- infrastructure, |
| </Segment> |
| <Segment StartTime="1922.848" EndTime="1925.359" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1925.342" EndTime="1927.501" Participant="fe016"> |
| I mean, that sounds good to me. I - I don't - |
| </Segment> |
| <Segment StartTime="1926.158" EndTime="1927.659" Participant="me011"> |
| which has all that built in. |
| </Segment> |
| <Segment StartTime="1927.963" EndTime="1931.526" Participant="fe016"> |
| So if - if you would l- look at that and let us know what you think. <Comment Description="rising intonation"/> I mean, |
| </Segment> |
| <Segment StartTime="1930.309" EndTime="1930.900" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1930.383" EndTime="1930.819" Participant="me011"> |
| Sure. |
| </Segment> |
| <Segment StartTime="1931.526" EndTime="1937.223" Participant="fe016"> |
| I think we're sort of guinea pigs, cuz I - I want to get the prosody work <Emphasis> done </Emphasis> but I don't want to waste time, you know, |
| </Segment> |
| <Segment StartTime="1936.504" EndTime="1937.222" Participant="mn017"> |
| Oh, maybe - |
| </Segment> |
| <Segment StartTime="1937.940" EndTime="1939.301" Participant="fe016"> |
| getting the - Yeah? |
| </Segment> |
| <Segment StartTime="1939.269" EndTime="1939.774" Participant="mn017"> |
| um - |
| </Segment> |
| <Segment StartTime="1939.847" EndTime="1943.565" Participant="me011"> |
| Well, I wouldn't wait for the formats, because anything you pick we'll be able to translate |
| </Segment> |
| <Segment StartTime="1941.364" EndTime="1942.577" Participant="mn017"> |
| Well - <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="1944.297" EndTime="1945.365" Participant="me011"> |
| to another form. |
| </Segment> |
| <Segment StartTime="1944.604" EndTime="1945.417" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="1945.417" EndTime="1950.436" Participant="mn017"> |
| <VocalSound Description="inbreath"/> Ma- well, maybe you should actually look at it yourself too to get a sense of |
| </Segment> |
| <Segment StartTime="1945.934" EndTime="1946.593" Participant="fe016"> |
| O_K. |
| </Segment> |
| <Segment StartTime="1951.113" EndTime="1953.784" Participant="mn017"> |
| what it is you'll - you'll be dealing with, because, |
| </Segment> |
| <Segment StartTime="1954.153" EndTime="1954.812" Participant="mn017"> |
| um, |
| </Segment> |
| <Segment StartTime="1955.821" EndTime="1959.114" Participant="mn017"> |
| you know, Adam might have <Emphasis> one </Emphasis> opinion but you might have <Emphasis> another, </Emphasis> so <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="1957.343" EndTime="1957.779" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1958.169" EndTime="1959.100" Participant="me001"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1958.747" EndTime="1959.578" Participant="fe016"> |
| Yeah, definitely. |
| </Segment> |
| <Segment StartTime="1959.595" EndTime="1960.510" Participant="me001"> |
| <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="1960.593" EndTime="1962.639" Participant="mn017"> |
| I think the more eyes look at this the better. |
| </Segment> |
| <Segment StartTime="1962.663" EndTime="1967.470" Participant="fe016"> |
| Especially if there's, <Uncertain> e- </Uncertain> um - you know, if someone can help with at least the - |
| </Segment> |
| <Segment StartTime="1968.340" EndTime="1970.324" Participant="fe016"> |
| the setup of the right - <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="1968.725" EndTime="1970.131" CloseMic="false"> |
| <NonVocalSound Description="door opens"/> |
| </Segment> |
| <Segment StartTime="1970.131" EndTime="1971.625" CloseMic="false"> |
| <Comment Description="[fe008] arrives"/> |
| </Segment> |
| <Segment StartTime="1970.324" EndTime="1972.039" Participant="fe016"> |
| Oh, hi. <VocalSound Description="laugh"/> <Comment Description="embedded"/> |
| </Segment> |
| <Segment StartTime="1970.388" EndTime="1971.120" Participant="me011"> |
| Hi, Jane. |
| </Segment> |
| <Segment StartTime="1971.558" EndTime="1972.051" Participant="mn017"> |
| Mmm. |
| </Segment> |
| <Segment StartTime="1974.509" EndTime="1982.736" Participant="fe016"> |
| the right representation, then, i- you know, I hope it won't - We don't actually need the whole full-blown thing to be <Emphasis> ready, </Emphasis> so. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1975.504" EndTime="1976.422" CloseMic="false"> |
| <NonVocalSound Description="door closes"/> |
| </Segment> |
| <Segment StartTime="1980.377" EndTime="1981.805" Participant="me011"> |
| Can you - Oh, well. |
| </Segment> |
| <Segment StartTime="1984.128" EndTime="1989.166" Participant="fe016"> |
| Um, so maybe if you guys can look at it and sort of see what, <Comment Description="rising intonation"/> |
| </Segment> |
| <Segment StartTime="1987.125" EndTime="1987.627" Participant="me011"> |
| Sure. |
| </Segment> |
| <Segment StartTime="1987.524" EndTime="1988.374" Participant="me001"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="1989.481" EndTime="1990.144" Participant="fe016"> |
| um - |
| </Segment> |
| <Segment StartTime="1990.968" EndTime="1995.377" Participant="fe016"> |
| I think we're - we're - <VocalSound Description="laugh"/> we're actually just - yeah, wrapping up, but, um - |
| </Segment> |
| <Segment StartTime="1991.365" EndTime="1994.646" Participant="me001"> |
| <VocalSound Description="clears throat"/> <VocalSound Description="breath"/> Hmm. |
| </Segment> |
| <Segment StartTime="1992.133" EndTime="1993.710" Participant="me011"> |
| We're about done. <Comment Description="while laughing"/> <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1993.302" EndTime="1996.328" Participant="me018"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="1994.849" EndTime="1995.699" Participant="fe008" CloseMic="false"> |
| Oh you are? Oh. |
| </Segment> |
| <Segment StartTime="1995.377" EndTime="1998.423" Participant="fe016"> |
| Yeah, sorry, it's a <Uncertain> uh </Uncertain> short meeting, but, um - |
| </Segment> |
| <Segment StartTime="1995.850" EndTime="1998.001" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="1998.890" EndTime="2000.581" Participant="fe016"> |
| Well, I don't know. Is there anything else, like - |
| </Segment> |
| <Segment StartTime="2001.115" EndTime="2002.690" Participant="fe016"> |
| <Uncertain> I mean </Uncertain> that helps <Emphasis> me </Emphasis> a lot, but - |
| </Segment> |
| <Segment StartTime="2001.489" EndTime="2009.917" Participant="me011"> |
| <VocalSound Description="inbreath"/> Well, I think the <Emphasis> other </Emphasis> thing we <Emphasis> might </Emphasis> want to look at is alternatives to P_file. I mean, th- the reason I like P_file is I'm already familiar with it, we have expertise <Emphasis> here, </Emphasis> |
| </Segment> |
| <Segment StartTime="2002.056" EndTime="2002.999" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2008.082" EndTime="2011.466" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2009.917" EndTime="2010.372" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2010.372" EndTime="2015.854" Participant="me011"> |
| and so if we pick something else, there's the learning-curve problem. But, I mean, it is just something we developed at ICSI. |
| </Segment> |
| <Segment StartTime="2013.213" EndTime="2013.787" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2016.210" EndTime="2016.942" Participant="me011"> |
| And so - |
| </Segment> |
| <Segment StartTime="2016.232" EndTime="2018.297" Participant="mn017"> |
| Is there an - is there an I_P- A_P_I? |
| </Segment> |
| <Segment StartTime="2019.059" EndTime="2021.512" Participant="me011"> |
| Yeah. There's an A_P_I for it. And, uh, |
| </Segment> |
| <Segment StartTime="2019.625" EndTime="2020.159" Participant="mn017"> |
| O_K. |
| </Segment> |
| <Segment StartTime="2021.663" EndTime="2024.568" Participant="mn017"> |
| <VocalSound Description="inbreath"/> There used to be a problem that they get too large, |
| </Segment> |
| <Segment StartTime="2022.086" EndTime="2024.221" Participant="me011"> |
| a bunch of libraries, P_file utilities. |
| </Segment> |
| <Segment StartTime="2023.847" EndTime="2026.909" Participant="me018"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2024.568" EndTime="2024.908" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2024.908" EndTime="2028.700" Participant="mn017"> |
| and so <Pause/> basically the - <Uncertain> uh </Uncertain> the filesystem |
| </Segment> |
| <Segment StartTime="2028.780" EndTime="2032.916" Participant="me011"> |
| <VocalSound Description="inbreath"/> Well, that's gonna be a problem no matter what. You have the two-gigabyte limit on the filesystem size. |
| </Segment> |
| <Segment StartTime="2029.100" EndTime="2029.752" Participant="mn017"> |
| wouldn't - |
| </Segment> |
| <Segment StartTime="2032.916" EndTime="2035.286" Participant="me011"> |
| And we definitely hit that with Broadcast News. |
| </Segment> |
| <Segment StartTime="2036.451" EndTime="2040.994" Participant="mn017"> |
| <VocalSound Description="inbreath"/> Maybe you could extend the A_P_I to, uh, support, uh, like splitting up, |
| </Segment> |
| <Segment StartTime="2040.513" EndTime="2044.607" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2041.615" EndTime="2047.154" Participant="mn017"> |
| you know, conceptually one file into smaller files on disk so that you can essentially, |
| </Segment> |
| <Segment StartTime="2047.436" EndTime="2047.812" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2047.787" EndTime="2049.900" Participant="mn017"> |
| you know, have arbitrarily long f- |
| </Segment> |
| <Segment StartTime="2048.188" EndTime="2050.724" Participant="me011"> |
| Most of the tools can handle that. So that |
| </Segment> |
| <Segment StartTime="2050.150" EndTime="2050.605" Participant="mn017"> |
| Yeah. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2051.160" EndTime="2059.128" Participant="me011"> |
| we didn't do it at the A_P_I-level. We did it at the t- tool-level. That - that - most - many of them can s- you can specify several P_files and they'll just be done sequentially. |
| </Segment> |
| <Segment StartTime="2059.466" EndTime="2060.320" Participant="mn017"> |
| O_K. |
| </Segment> |
| <Segment StartTime="2060.211" EndTime="2060.750" Participant="me011"> |
| So. |
| </Segment> |
| <Segment StartTime="2062.040" EndTime="2068.610" Participant="fe016"> |
| So, I guess, yeah, if - if you and Don can - if you can show him the P_file stuff and see. So this would be like for the F_zero - |
| </Segment> |
| <Segment StartTime="2065.899" EndTime="2066.386" Participant="me011"> |
| Sure. |
| </Segment> |
| <Segment StartTime="2066.386" EndTime="2070.322" Participant="me011"> |
| <VocalSound Description="inbreath"/> I mean, if you do "man P_file" or "apropos P_file", you'll see a lot. |
| </Segment> |
| <Segment StartTime="2066.865" EndTime="2067.886" Participant="me001"> |
| <Uncertain> True. </Uncertain> |
| </Segment> |
| <Segment StartTime="2068.867" EndTime="2073.517" Participant="me001"> |
| <VocalSound Description="inbreath"/> I've <Emphasis> used </Emphasis> the P_file, I think. I've <Emphasis> looked </Emphasis> at it at least, briefly, I think when we were doing |
| </Segment> |
| <Segment StartTime="2073.427" EndTime="2074.170" Participant="me011"> |
| <VocalSound Description="hiccup"/> |
| </Segment> |
| <Segment StartTime="2074.120" EndTime="2075.812" Participant="mn017"> |
| What does the P_ stand for anyway? |
| </Segment> |
| <Segment StartTime="2074.629" EndTime="2076.180" Participant="me001"> |
| s- something. |
| </Segment> |
| <Segment StartTime="2075.525" EndTime="2079.417" Participant="me011"> |
| I have no idea. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2076.698" EndTime="2077.623" Participant="fe016"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2077.728" EndTime="2078.970" Participant="me001"> |
| <Uncertain> Oh, in there. </Uncertain> |
| </Segment> |
| <Segment StartTime="2079.596" EndTime="2080.426" Participant="mn017"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2079.860" EndTime="2083.318" Participant="me011"> |
| <Emphasis> I </Emphasis> didn't de- I didn't develop it. You know, it was - I <Emphasis> think </Emphasis> it was Dave Johnson. |
| </Segment> |
| <Segment StartTime="2083.931" EndTime="2087.432" Participant="me011"> |
| So it's all part of the Quicknet library. It has all the utilities for it. |
| </Segment> |
| <Segment StartTime="2088.293" EndTime="2090.904" Participant="mn017"> |
| No, <Emphasis> P_files </Emphasis> were around <Emphasis> way </Emphasis> before <Emphasis> Quicknet. </Emphasis> <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2090.749" EndTime="2091.422" Participant="me011"> |
| Oh, were they? |
| </Segment> |
| <Segment StartTime="2090.904" EndTime="2094.840" Participant="mn017"> |
| P_files were - were around when - w- with, um, <VocalSound Description="mouth"/> RAP. |
| </Segment> |
| <Segment StartTime="2094.805" EndTime="2095.531" Participant="me018"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="2094.867" EndTime="2098.111" Participant="fe016"> |
| It's like the history of ICSI. Like - <Comment Description="while laughing"/> <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2095.372" EndTime="2096.007" Participant="mn017"> |
| Right? |
| </Segment> |
| <Segment StartTime="2096.167" EndTime="2097.591" Participant="me001"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2096.472" EndTime="2100.774" Participant="mn017"> |
| <Emphasis> You </Emphasis> worked with P_files. <Emphasis> I </Emphasis> worked with P_files. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2097.413" EndTime="2098.264" Participant="me018"> |
| No. |
| </Segment> |
| <Segment StartTime="2097.444" EndTime="2098.037" Participant="me011"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="2098.507" EndTime="2099.270" Participant="fe016"> |
| Yeah? |
| </Segment> |
| <Segment StartTime="2099.210" EndTime="2100.200" Participant="me001"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2100.973" EndTime="2102.615" Participant="me018"> |
| I don't remember what the "<Emphasis> P_ </Emphasis>" is, though. |
| </Segment> |
| <Segment StartTime="2102.161" EndTime="2103.670" Participant="fe008" CloseMic="false"> |
| Is it related to P_make? |
| </Segment> |
| <Segment StartTime="2103.366" EndTime="2103.900" Participant="me011"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2103.544" EndTime="2104.263" Participant="mn017"> |
| No. |
| </Segment> |
| <Segment StartTime="2105.010" EndTime="2106.278" Participant="me011"> |
| But there are ni- they're - |
| </Segment> |
| <Segment StartTime="2106.950" EndTime="2111.245" Participant="me011"> |
| The <Pause/> Quicknet library has a bunch of things in it to handle P_files, so it works pretty well. |
| </Segment> |
| <Segment StartTime="2109.509" EndTime="2109.964" Participant="mn017"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2111.253" EndTime="2111.645" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2112.139" EndTime="2117.383" Participant="fe016"> |
| And that isn't really, I guess, as important as the - the main - I don't know what you call it, the - |
| </Segment> |
| <Segment StartTime="2112.208" EndTime="2112.736" Participant="mn017"> |
| <Uncertain NumSyllables="1"> @@ </Uncertain> |
| </Segment> |
| <Segment StartTime="2115.685" EndTime="2116.357" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2117.383" EndTime="2119.070" Participant="fe016"> |
| the main sort of word-level - |
| </Segment> |
| <Segment StartTime="2117.548" EndTime="2118.304" Participant="me011"> |
| Neither do I. |
| </Segment> |
| <Segment StartTime="2118.188" EndTime="2119.691" Participant="me018"> |
| <Emphasis> Probably </Emphasis> stands for "Phil". |
| </Segment> |
| <Segment StartTime="2118.587" EndTime="2119.121" Participant="me011"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2120.916" EndTime="2121.743" Participant="me018"> |
| Phil Kohn. |
| </Segment> |
| <Segment StartTime="2121.664" EndTime="2122.288" Participant="mn017"> |
| <VocalSound Description="mouth"/> <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2122.938" EndTime="2124.540" Participant="me011"> |
| It's a Phil file? <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2124.005" EndTime="2124.417" Participant="me018"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2124.022" EndTime="2124.951" Participant="me001"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2124.161" EndTime="2125.848" Participant="fe016"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2125.836" EndTime="2127.603" Participant="me018"> |
| That's <Emphasis> my </Emphasis> guess. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2126.924" EndTime="2127.362" Participant="fe016"> |
| <Uncertain> Huh. </Uncertain> |
| </Segment> |
| <Segment StartTime="2128.608" EndTime="2135.220" Participant="fe016"> |
| O_K. Well, that's really useful. I mean, this is exactly the kind of thing that I wanted to settle. Um, so - |
| </Segment> |
| <Segment StartTime="2136.119" EndTime="2140.270" Participant="me011"> |
| Yeah, I've been meaning to look at the ATLAS stuff again anyway. So, just keep - |
| </Segment> |
| <Segment StartTime="2136.259" EndTime="2136.832" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2136.535" EndTime="2137.710" Participant="fe016"> |
| Great. |
| </Segment> |
| <Segment StartTime="2139.118" EndTime="2139.651" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2140.166" EndTime="2145.244" Participant="fe016"> |
| I guess it's also sort of a political deci- I mean, if - if you feel like that's a community that would |
| </Segment> |
| <Segment StartTime="2146.374" EndTime="2148.631" Participant="fe016"> |
| be good to tie into <Emphasis> anyway, </Emphasis> then it's - |
| </Segment> |
| <Segment StartTime="2149.521" EndTime="2150.965" Participant="fe016"> |
| sounds like it's worth doing. |
| </Segment> |
| <Segment StartTime="2149.858" EndTime="2150.768" Participant="me011"> |
| Yeah, I think it - it w- |
| </Segment> |
| <Segment StartTime="2150.566" EndTime="2151.990" Participant="mn017"> |
| <Uncertain> j- </Uncertain> I think there's - |
| </Segment> |
| <Segment StartTime="2150.992" EndTime="2157.598" Participant="me011"> |
| And, w- uh, as I said, I - what I did with <Emphasis> this </Emphasis> stuff - I <Emphasis> based </Emphasis> it on theirs. It's just they hadn't actually come up with an external format yet. |
| </Segment> |
| <Segment StartTime="2157.598" EndTime="2158.092" Participant="me011"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2158.092" EndTime="2159.952" Participant="me011"> |
| So now that they <Emphasis> have </Emphasis> come up with a format, |
| </Segment> |
| <Segment StartTime="2160.619" EndTime="2161.887" Participant="mn017"> |
| <VocalSound Description="inbreath"/> <Uncertain> Mmm. </Uncertain> <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2160.921" EndTime="2163.104" Participant="me011"> |
| it doesn't - it seems pretty reasonable to use it. |
| </Segment> |
| <Segment StartTime="2163.757" EndTime="2165.142" Participant="me011"> |
| But let me look at it again. |
| </Segment> |
| <Segment StartTime="2164.865" EndTime="2165.869" Participant="fe016"> |
| O_K, great. |
| </Segment> |
| <Segment StartTime="2165.557" EndTime="2166.920" Participant="me011"> |
| As I said, that - |
| </Segment> |
| <Segment StartTime="2166.227" EndTime="2167.752" Participant="fe016"> |
| Cuz we actually can start - |
| </Segment> |
| <Segment StartTime="2166.920" EndTime="2171.671" Participant="me011"> |
| There's one level - there's one more level of indirection and I'm just blanking on exactly how it works. |
| </Segment> |
| <Segment StartTime="2171.671" EndTime="2173.372" Participant="me011"> |
| <VocalSound Description="inbreath"/> I gotta look at it again. |
| </Segment> |
| <Segment StartTime="2174.824" EndTime="2183.243" Participant="fe016"> |
| I mean, we can <Emphasis> start </Emphasis> with, um, I guess, this input from Dave's, which you had printed out, the <Emphasis> channelized </Emphasis> input. Cuz he has all of the |
| </Segment> |
| <Segment StartTime="2183.737" EndTime="2188.583" Participant="fe016"> |
| channels, you know, with the channels in the tag and stuff like that. So that would be <Uncertain> i- </Uncertain> directly, |
| </Segment> |
| <Segment StartTime="2185.214" EndTime="2186.143" Participant="me011"> |
| Yeah, I've seen it. |
| </Segment> |
| <Segment StartTime="2186.167" EndTime="2187.552" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2188.990" EndTime="2189.253" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2189.290" EndTime="2189.869" Participant="fe016"> |
| um - |
| </Segment> |
| <Segment StartTime="2189.664" EndTime="2191.120" Participant="me011"> |
| Easy - easy to map. |
| </Segment> |
| <Segment StartTime="2190.300" EndTime="2192.478" Participant="fe016"> |
| Yeah. And so then it would just be a matter of getting - |
| </Segment> |
| <Segment StartTime="2192.795" EndTime="2198.353" Participant="fe016"> |
| making sure to handle the annotations that are, you know, not at the word-level and, um, |
| </Segment> |
| <Segment StartTime="2198.759" EndTime="2200.477" Participant="me001"> |
| Where are those annotations coming from? |
| </Segment> |
| <Segment StartTime="2198.996" EndTime="2200.036" Participant="fe016"> |
| t- to import the- <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2200.727" EndTime="2203.892" Participant="fe016"> |
| Well, right now, I g- Jane would - <VocalSound Description="laugh"/> would - |
| </Segment> |
| <Segment StartTime="2202.382" EndTime="2202.910" Participant="me011"> |
| Mm-hmm. <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="2203.647" EndTime="2207.611" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> <VocalSound Description="clears throat"/> |
| </Segment> |
| <Segment StartTime="2204.544" EndTime="2205.124" Participant="fe016"> |
| <Uncertain> Yeah. </Uncertain> |
| </Segment> |
| <Segment StartTime="2207.611" EndTime="2209.482" Participant="fe008"> |
| Are you talking about the overlap a- annotations? |
| </Segment> |
| <Segment StartTime="2209.190" EndTime="2214.871" Participant="fe016"> |
| Yeah, any kind of annotation <Pause/> that, like, isn't already there. Uh, you know, anything you can envision. |
| </Segment> |
| <Segment StartTime="2213.184" EndTime="2214.613" Participant="fe008"> |
| <VocalSound Description="inbreath"/> Yeah. |
| </Segment> |
| <Segment StartTime="2214.950" EndTime="2220.041" Participant="fe008"> |
| So what <Emphasis> I </Emphasis> was imagining was - um, so Dave says we can have unlimited numbers of green ribbons. |
| </Segment> |
| <Segment StartTime="2219.982" EndTime="2220.430" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2220.379" EndTime="2223.948" Participant="fe008"> |
| And so put, uh, a - a green ribbon on for an overlap code. |
| </Segment> |
| <Segment StartTime="2223.948" EndTime="2230.828" Participant="fe008"> |
| <VocalSound Description="inbreath"/> And since we w- we - I - I think it's important to remain flexible regarding the time bins for now. |
| </Segment> |
| <Segment StartTime="2231.255" EndTime="2233.215" Participant="fe008"> |
| And so it's nice to have - |
| </Segment> |
| <Segment StartTime="2233.597" EndTime="2241.247" Participant="fe008"> |
| However, you know, you want to have it, uh, time- time- uh, <Emphasis> located </Emphasis> in the discourse. So, um, |
| </Segment> |
| <Segment StartTime="2241.622" EndTime="2246.635" Participant="fe008"> |
| if we - if we tie the overlap code to the first word in the overlap, |
| </Segment> |
| <Segment StartTime="2246.976" EndTime="2250.819" Participant="fe008"> |
| then you'll have a time-marking. It won't - it'll be independent of the time bins, |
| </Segment> |
| <Segment StartTime="2250.819" EndTime="2257.124" Participant="fe008"> |
| however these e- evolve, shrink, or whatever, increase, or - Also, you could have different time bins for different purposes. |
| </Segment> |
| <Segment StartTime="2257.124" EndTime="2264.297" Participant="fe008"> |
| <VocalSound Description="inbreath"/> And having it tied to the first word in an overlap segment is unique, uh, you know, anchored, clear. |
| </Segment> |
| <Segment StartTime="2265.015" EndTime="2266.981" Participant="fe008"> |
| And it would just end up on a separate ribbon. |
| </Segment> |
| <Segment StartTime="2267.280" EndTime="2269.741" Participant="fe008"> |
| So the <Emphasis> overlap </Emphasis> coding is gonna be <Emphasis> easy </Emphasis> with respect to that. |
| </Segment> |
| <Segment StartTime="2267.471" EndTime="2268.019" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="2270.486" EndTime="2271.862" Participant="fe008"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2271.862" EndTime="2272.922" Participant="fe008"> |
| You look puzzled. <Comment Description="to [me018]"/> |
| </Segment> |
| <Segment StartTime="2272.891" EndTime="2277.126" Participant="me018"> |
| <VocalSound Description="inbreath"/> I - I just - I don't quite understand what these things are. <Uncertain> Uh. </Uncertain> |
| </Segment> |
| <Segment StartTime="2273.402" EndTime="2274.440" Participant="fe008"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2274.949" EndTime="2279.700" Participant="fe008"> |
| <VocalSound Description="breath"/> O_K. What, the codes themselves? Or the - ? <Comment Description="leading question"/> <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2278.113" EndTime="2281.164" Participant="me018"> |
| Well, th- overlap codes. I'm not sure what that <Uncertain NumSyllables="1"> @@ </Uncertain> - |
| </Segment> |
| <Segment StartTime="2280.143" EndTime="2281.150" Participant="me011"> |
| Well, I mean, is that - |
| </Segment> |
| <Segment StartTime="2281.599" EndTime="2285.350" Participant="fe008"> |
| Well, we don't have to go into the codes. We don't have to go into the codes. But let me just - |
| </Segment> |
| <Segment StartTime="2282.113" EndTime="2284.000" Participant="me018"> |
| It probably doesn't matter. No, I d- |
| </Segment> |
| <Segment StartTime="2283.292" EndTime="2286.187" Participant="me011"> |
| I mean, it doesn't. I mean, that - not for the topic of <Emphasis> this </Emphasis> meeting. |
| </Segment> |
| <Segment StartTime="2286.033" EndTime="2288.406" Participant="fe008"> |
| No. W- the idea is just to have a separate green ribbon, |
| </Segment> |
| <Segment StartTime="2288.715" EndTime="2295.083" Participant="fe008"> |
| you know, and - and - and let's say that this is a time bin. There's a word here. This is the first word of an overlapping segment of <Emphasis> any </Emphasis> <Emphasis> length, </Emphasis> |
| </Segment> |
| <Segment StartTime="2295.376" EndTime="2300.612" Participant="fe008"> |
| overlapping with <Emphasis> any </Emphasis> other, uh, word - uh, i- segment of any length. And, um, |
| </Segment> |
| <Segment StartTime="2301.440" EndTime="2304.325" Participant="fe008"> |
| then you can indicate that this here was perhaps a ch- a backchannel, |
| </Segment> |
| <Segment StartTime="2304.325" EndTime="2311.131" Participant="fe008"> |
| or you can say that it was, um, a usurping of the turn, or you can - you know, any - any number of categories. But the fact is, you have it <Emphasis> time-tagged </Emphasis> |
| </Segment> |
| <Segment StartTime="2311.400" EndTime="2315.941" Participant="fe008"> |
| in a way that's independent of the, uh, sp- particular time bin that the word ends up in. |
| </Segment> |
| <Segment StartTime="2312.730" EndTime="2314.706" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2316.094" EndTime="2320.208" Participant="fe008"> |
| If it's a large unit or a small unit, or we sh- change the boundaries of the units, |
| </Segment> |
| <Segment StartTime="2316.861" EndTime="2317.341" Participant="mn017"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="2320.050" EndTime="2320.616" Participant="fe016"> |
| Right. |
| </Segment> |
| <Segment StartTime="2320.485" EndTime="2323.551" Participant="fe008"> |
| it's still unique and - and, uh, fits with the format, |
| </Segment> |
| <Segment StartTime="2324.006" EndTime="2325.430" Participant="fe008"> |
| flexible, all that. |
| </Segment> |
| <Segment StartTime="2324.244" EndTime="2324.968" Participant="mn017"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2325.683" EndTime="2326.440" Participant="mn017"> |
| Um, |
| </Segment> |
| <Segment StartTime="2328.279" EndTime="2329.710" Participant="mn017"> |
| it would be nice - um, |
| </Segment> |
| <Segment StartTime="2329.710" EndTime="2331.925" Participant="mn017"> |
| eh, gr- this is sort of r- regarding - |
| </Segment> |
| <Segment StartTime="2331.925" EndTime="2338.850" Participant="mn017"> |
| <Uncertain> uh, uh </Uncertain> it's related but not directly germane to the topic of discussion, but, when it comes to annotations, um, |
| </Segment> |
| <Segment StartTime="2339.335" EndTime="2341.671" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2339.613" EndTime="2349.622" Participant="mn017"> |
| you often find yourself in the situation where you have <Pause/> different annotations <Pause/> of the same, say, word sequence. |
| </Segment> |
| <Segment StartTime="2349.798" EndTime="2350.485" Participant="fe008"> |
| Yeah. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2350.245" EndTime="2357.139" Participant="mn017"> |
| O_K? And sometimes the word sequences even differ slightly because they were edited s- at one place but not the other. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2350.906" EndTime="2351.435" Participant="fe008"> |
| Yeah. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2357.139" EndTime="2361.774" Participant="mn017"> |
| So, once this data gets out there, some people might start annotating this for, |
| </Segment> |
| <Segment StartTime="2363.421" EndTime="2367.582" Participant="mn017"> |
| I don't know, dialogue acts or, um, |
| </Segment> |
| <Segment StartTime="2368.984" EndTime="2374.879" Participant="mn017"> |
| you know, topics or what the heck. You know, there's a zillion things that people might annotate this for. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2374.879" EndTime="2376.930" Participant="mn017"> |
| And the only thing that |
| </Segment> |
| <Segment StartTime="2377.230" EndTime="2383.525" Participant="mn017"> |
| is really sort of common among all the versi- the various versions of this data is the word sequence, |
| </Segment> |
| <Segment StartTime="2383.227" EndTime="2383.820" Participant="fe008"> |
| Yep. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2383.804" EndTime="2385.397" Participant="mn017"> |
| or <Emphasis> approximately. </Emphasis> |
| </Segment> |
| <Segment StartTime="2384.675" EndTime="2385.811" Participant="fe016"> |
| Or the time. |
| </Segment> |
| <Segment StartTime="2385.397" EndTime="2386.189" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2386.189" EndTime="2392.696" Participant="mn017"> |
| Or the times. But, see, if you'd annotate dialogue acts, you don't necessarily want to - or topics - you don't really want to be dealing with <Emphasis> time-marks. </Emphasis> |
| </Segment> |
| <Segment StartTime="2387.103" EndTime="2387.657" Participant="fe016"> |
| I guess. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2392.696" EndTime="2396.375" Participant="mn017"> |
| You'd - it's much more efficient for them to just see the <Emphasis> word </Emphasis> sequence, right? |
| </Segment> |
| <Segment StartTime="2394.641" EndTime="2395.099" Participant="fe016"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="2396.375" EndTime="2397.087" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2397.087" EndTime="2404.150" Participant="mn017"> |
| I mean, most people aren't as <Emphasis> sophisticated </Emphasis> as - as we are <Emphasis> here </Emphasis> with, you know, uh, time alignments and stuff. So - |
| </Segment> |
| <Segment StartTime="2401.056" EndTime="2401.827" Participant="me011"> |
| <VocalSound Description="laugh, mike interaction"/> |
| </Segment> |
| <Segment StartTime="2401.665" EndTime="2403.513" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2402.851" EndTime="2405.253" Participant="fe016"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2404.150" EndTime="2405.495" Participant="mn017"> |
| <VocalSound Description="mouth"/> <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2404.274" EndTime="2405.126" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2405.800" EndTime="2408.590" Participant="mn017"> |
| So the - the - the <Emphasis> point </Emphasis> is - |
| </Segment> |
| <Segment StartTime="2405.937" EndTime="2406.827" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2405.985" EndTime="2406.439" Participant="fe016"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2407.481" EndTime="2410.170" Participant="me011"> |
| Should - should we mention some names on the people who are n- ? <Comment Description="abandoned as clear"/> |
| </Segment> |
| <Segment StartTime="2408.912" EndTime="2411.600" Participant="fe016"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2409.132" EndTime="2410.833" Participant="me018"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2409.598" EndTime="2411.300" Participant="me001"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2409.616" EndTime="2410.467" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="2410.843" EndTime="2416.263" Participant="mn017"> |
| So, um, the p- my <Emphasis> point </Emphasis> is that <Pause/> you're gonna end up with, |
| </Segment> |
| <Segment StartTime="2412.079" EndTime="2413.068" Participant="me018"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2417.009" EndTime="2422.431" Participant="mn017"> |
| uh, word sequences that are differently annotated. And <Pause/> you want some tool, |
| </Segment> |
| <Segment StartTime="2417.881" EndTime="2418.259" Participant="fe016"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2423.230" EndTime="2428.282" Participant="mn017"> |
| uh, that is able to sort of merge these different annotations back into a single, |
| </Segment> |
| <Segment StartTime="2428.816" EndTime="2430.291" Participant="mn017"> |
| uh, version. |
| </Segment> |
| <Segment StartTime="2430.813" EndTime="2431.682" Participant="mn017"> |
| <VocalSound Description="mouth"/> O_K? |
| </Segment> |
| <Segment StartTime="2432.040" EndTime="2438.544" Participant="mn017"> |
| Um, and we had this problem very massively, uh, at S_R_I when we worked, uh, a while back on, |
| </Segment> |
| <Segment StartTime="2438.544" EndTime="2439.016" Participant="mn017"> |
| <VocalSound Description="mouth"/> |
| </Segment> |
| <Segment StartTime="2439.016" EndTime="2440.366" Participant="mn017"> |
| uh - |
| </Segment> |
| <Segment StartTime="2440.366" EndTime="2445.347" Participant="mn017"> |
| well, on dialogue acts as well as, uh, you know, um, what <Emphasis> was </Emphasis> it? <Comment Description="embedded"/> uh, |
| </Segment> |
| <Segment StartTime="2445.307" EndTime="2447.400" Participant="fe016"> |
| Well, all the Switchboard in it. Yeah. |
| </Segment> |
| <Segment StartTime="2445.864" EndTime="2451.028" Participant="mn017"> |
| utterance types. There's, uh, automatic, uh, punctuation and stuff like that. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2450.392" EndTime="2451.644" Participant="me018"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2451.028" EndTime="2454.984" Participant="mn017"> |
| Because we had one set of <Pause/> annotations that were based on, |
| </Segment> |
| <Segment StartTime="2455.538" EndTime="2462.981" Participant="mn017"> |
| uh, one version of the transcripts with a particular segmentation, and then we had <Emphasis> another </Emphasis> version that was based on, |
| </Segment> |
| <Segment StartTime="2463.562" EndTime="2468.511" Participant="mn017"> |
| uh, a different s- slightly <Emphasis> edited </Emphasis> version of the transcripts with a <Emphasis> different </Emphasis> segmentation. So, |
| </Segment> |
| <Segment StartTime="2468.511" EndTime="2468.855" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2468.855" EndTime="2470.790" Participant="mn017"> |
| we had these two different |
| </Segment> |
| <Segment StartTime="2471.272" EndTime="2477.474" Participant="mn017"> |
| versions which were - you know, you could tell they were from the same source but they weren't identical. So it was extremely hard |
| </Segment> |
| <Segment StartTime="2477.474" EndTime="2477.989" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2477.989" EndTime="2483.640" Participant="mn017"> |
| to reliably merge these two back together to correlate the information from the different annotations. |
| </Segment> |
| <Segment StartTime="2483.027" EndTime="2488.210" Participant="me011"> |
| Yep. I - I don't see any way that <Emphasis> file </Emphasis> formats are gonna <Emphasis> help </Emphasis> us with that. It's - it's all a question of semantic. |
| </Segment> |
| <Segment StartTime="2486.001" EndTime="2486.569" Participant="mn017"> |
| No. |
| </Segment> |
| <Segment StartTime="2486.128" EndTime="2487.131" Participant="fe016"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2487.353" EndTime="2491.331" Participant="mn017"> |
| <VocalSound Description="mouth"/> No. But once you have a file format, I can imagine writing - |
| </Segment> |
| <Segment StartTime="2488.556" EndTime="2489.048" Participant="fe016"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="2491.845" EndTime="2496.986" Participant="mn017"> |
| not personally, but someone writing a tool that is essentially an alignment tool, |
| </Segment> |
| <Segment StartTime="2493.246" EndTime="2494.070" Participant="fe016"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2493.462" EndTime="2494.234" Participant="me001"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2496.839" EndTime="2497.425" Participant="me011"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2497.390" EndTime="2501.269" Participant="mn017"> |
| um, that mediates between various versions, and - |
| </Segment> |
| <Segment StartTime="2501.269" EndTime="2505.556" Participant="mn017"> |
| uh, sort of like th- uh, you know, you have this thing in UNIX where you have, uh, |
| </Segment> |
| <Segment StartTime="2505.369" EndTime="2505.965" Participant="me011"> |
| Diff. <Comment Description="completing [mn017]'s utterance"/> |
| </Segment> |
| <Segment StartTime="2505.846" EndTime="2507.524" Participant="fe016"> |
| <Emphasis> W_diff </Emphasis> or diff. Yeah. |
| </Segment> |
| <Segment StartTime="2505.979" EndTime="2506.267" Participant="mn017"> |
| diff. |
| </Segment> |
| <Segment StartTime="2506.267" EndTime="2506.683" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2506.683" EndTime="2512.341" Participant="mn017"> |
| There's the, uh, diff that actually tries to reconcile different - |
| </Segment> |
| <Segment StartTime="2511.639" EndTime="2512.817" Participant="fe008"> |
| Is it <Emphasis> S_diff </Emphasis>? <Comment Description="to [mn017]"/> |
| </Segment> |
| <Segment StartTime="2512.341" EndTime="2514.701" Participant="mn017"> |
| two diffs f- <Comment Description="increased volume"/> based on the same original. <Comment Description="faster tempo on "original""/> |
| </Segment> |
| <Segment StartTime="2514.108" EndTime="2514.583" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2514.653" EndTime="2515.563" Participant="fe008"> |
| Mmm. <Comment Description="quietly, with some mike noise"/> |
| </Segment> |
| <Segment StartTime="2514.701" EndTime="2515.332" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2515.332" EndTime="2516.515" Participant="mn017"> |
| Something like that, |
| </Segment> |
| <Segment StartTime="2516.840" EndTime="2523.630" Participant="mn017"> |
| um, but operating on these lattices that are really what's behind this - uh, this annotation format. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2522.259" EndTime="2522.708" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2523.630" EndTime="2524.358" Participant="mn017"> |
| So - |
| </Segment> |
| <Segment StartTime="2523.677" EndTime="2525.369" Participant="fe016"> |
| You could definitely do that with the - |
| </Segment> |
| <Segment StartTime="2523.796" EndTime="2527.980" Participant="me011"> |
| There's actually a diff library you can use <Pause/> to do things like that that - |
| </Segment> |
| <Segment StartTime="2528.654" EndTime="2534.957" Participant="mn017"> |
| So somewhere in the A_P_I you would like to have like a merge or some - some function that merges two - |
| </Segment> |
| <Segment StartTime="2529.068" EndTime="2530.514" Participant="me011"> |
| so you have different formats. |
| </Segment> |
| <Segment StartTime="2535.370" EndTime="2538.127" Participant="me011"> |
| Yeah, I think it's gonna be very hard. Any sort of structured |
| </Segment> |
| <Segment StartTime="2535.639" EndTime="2536.350" Participant="mn017"> |
| two versions. |
| </Segment> |
| <Segment StartTime="2537.300" EndTime="2546.082" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2538.642" EndTime="2539.286" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="2538.800" EndTime="2542.543" Participant="me011"> |
| <Emphasis> anything </Emphasis> when you try to merge is really, really hard because you ha- i- |
| </Segment> |
| <Segment StartTime="2542.543" EndTime="2547.737" Participant="me011"> |
| The hard part isn't the file format. The hard part is specifying what you <Emphasis> mean </Emphasis> by "merge". <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2544.942" EndTime="2545.739" Participant="mn017"> |
| Is - |
| </Segment> |
| <Segment StartTime="2546.677" EndTime="2548.110" Participant="mn017"> |
| Exactly. <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="2547.186" EndTime="2552.744" Participant="fe016"> |
| But the one thing that would work here actually for i- that is more reliable than the utterances is the - |
| </Segment> |
| <Segment StartTime="2547.737" EndTime="2549.473" Participant="me011"> |
| And that's very difficult. <VocalSound Description="outbreath"/> |
| </Segment> |
| <Segment StartTime="2548.990" EndTime="2550.098" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2553.773" EndTime="2558.324" Participant="fe016"> |
| the speaker ons and offs. So if you have a good, um - |
| </Segment> |
| <Segment StartTime="2558.252" EndTime="2561.778" Participant="me011"> |
| But this is exactly what I <Emphasis> mean, </Emphasis> is that - that the problem i- |
| </Segment> |
| <Segment StartTime="2559.098" EndTime="2559.652" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2560.531" EndTime="2563.814" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2560.621" EndTime="2562.836" Participant="fe016"> |
| You just have to know wha- what to tie it to. |
| </Segment> |
| <Segment StartTime="2562.397" EndTime="2568.640" Participant="me011"> |
| Yeah, exactly. The problem is saying "what are the semantics, what do you <Emphasis> mean </Emphasis> by "merge"?" So. |
| </Segment> |
| <Segment StartTime="2563.302" EndTime="2564.102" Participant="fe016"> |
| And - |
| </Segment> |
| <Segment StartTime="2566.164" EndTime="2567.289" Participant="fe016"> |
| Right, right. |
| </Segment> |
| <Segment StartTime="2566.567" EndTime="2567.147" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="2567.775" EndTime="2572.939" Participant="mn017"> |
| So - so just to let you know what we - where we <Emphasis> kluged </Emphasis> it by, uh, doing - |
| </Segment> |
| <Segment StartTime="2573.720" EndTime="2575.259" Participant="mn017"> |
| uh, by doing - Hhh. <Comment Description="vocal gesture, difficulty"/> |
| </Segment> |
| <Segment StartTime="2575.259" EndTime="2582.817" Participant="mn017"> |
| Both were based on <Emphasis> words, </Emphasis> so, bo- we have two versions of the same words intersp- you know, sprinkled with - with different <Emphasis> tags </Emphasis> for annotations. |
| </Segment> |
| <Segment StartTime="2581.980" EndTime="2585.443" Participant="me011"> |
| And then you did diff. <VocalSound Description="laugh"/> Yeah, that's just what I thought. <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="2582.817" EndTime="2584.669" Participant="mn017"> |
| <VocalSound Description="breath"/> And we did <Emphasis> diff. </Emphasis> Exactly! |
| </Segment> |
| <Segment StartTime="2583.196" EndTime="2584.320" Participant="fe016"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2585.030" EndTime="2587.502" Participant="mn017"> |
| And that's how - Yeah. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2585.443" EndTime="2590.286" Participant="me011"> |
| That's just wh- how <Emphasis> I </Emphasis> would have done it. <Comment Description="while laughing"/> <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2587.502" EndTime="2592.785" Participant="mn017"> |
| But, you know, it had lots of errors and things would end up in the wrong order, and so forth. Uh, so, |
| </Segment> |
| <Segment StartTime="2592.870" EndTime="2593.373" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2593.150" EndTime="2595.653" Participant="mn017"> |
| um, if you had a more - |
| </Segment> |
| <Segment StartTime="2595.653" EndTime="2602.593" Participant="mn017"> |
| Uh, it - it was a kluge because it was basically reducing everything to - uh, to - uh, uh, to <Emphasis> textual </Emphasis> alignment. |
| </Segment> |
| <Segment StartTime="2600.421" EndTime="2601.296" Participant="me011"> |
| A textual - <Comment Description="completing [mn017]'s utterance"/> |
| </Segment> |
| <Segment StartTime="2602.743" EndTime="2603.831" Participant="fe016"> |
| <VocalSound Description="inbreath"/> <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2602.932" EndTime="2604.287" Participant="mn017"> |
| Um, so - |
| </Segment> |
| <Segment StartTime="2603.831" EndTime="2608.412" Participant="fe016"> |
| But, d- isn't that something where whoever - if <VocalSound Description="clears throat"/> - if the people |
| </Segment> |
| <Segment StartTime="2607.642" EndTime="2608.838" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2608.720" EndTime="2613.892" Participant="fe016"> |
| who are making changes, say in the transcripts, cuz this all happened when the transcripts were different - |
| </Segment> |
| <Segment StartTime="2614.181" EndTime="2619.510" Participant="fe016"> |
| ye- um, if they <Emphasis> tie </Emphasis> it to <Emphasis> something, </Emphasis> like if they tied it to the acoustic segment - if they - |
| </Segment> |
| <Segment StartTime="2619.888" EndTime="2627.577" Participant="fe016"> |
| You know what I mean? Then - Or if they tied it to an acoustic segment and we had the <Emphasis> time-marks, </Emphasis> that would help. But the problem is exactly as Adam said, that you get, |
| </Segment> |
| <Segment StartTime="2623.596" EndTime="2624.985" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2625.105" EndTime="2625.495" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2628.516" EndTime="2632.860" Participant="fe016"> |
| you know, y- you don't <Emphasis> have </Emphasis> that information or it's lost in the merge somehow, so - |
| </Segment> |
| <Segment StartTime="2632.012" EndTime="2638.556" Participant="fe008"> |
| <VocalSound Description="inbreath"/> Well, can I ask one question? It - it seems to me that, um, we will have o- an official version of the corpus, which will be only one - |
| </Segment> |
| <Segment StartTime="2635.100" EndTime="2638.561" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2638.902" EndTime="2641.248" Participant="fe008"> |
| one version in terms of the words - where the words are concerned. |
| </Segment> |
| <Segment StartTime="2641.248" EndTime="2645.045" Participant="fe008"> |
| We'd still have the - the merging issue maybe if coding were done independently of the - |
| </Segment> |
| <Segment StartTime="2645.192" EndTime="2650.574" Participant="mn017"> |
| <VocalSound Description="mouth"/> And you're gonna <Emphasis> get </Emphasis> that because if the data gets out, people will do all kinds of things to it. |
| </Segment> |
| <Segment StartTime="2645.449" EndTime="2646.700" Participant="fe008"> |
| But - but - |
| </Segment> |
| <Segment StartTime="2650.644" EndTime="2650.941" Participant="fe008"> |
| <VocalSound Description="inbreath"/> <Comment Description="to begin speaking"/> |
| </Segment> |
| <Segment StartTime="2651.272" EndTime="2657.130" Participant="mn017"> |
| And, uh, s- you know, several years from now you might want to look into, um, |
| </Segment> |
| <Segment StartTime="2657.625" EndTime="2659.330" Participant="mn017"> |
| the prosody of |
| </Segment> |
| <Segment StartTime="2660.967" EndTime="2668.313" Participant="mn017"> |
| referring expressions. And someone at the university of who knows where has annotated the referring expressions. |
| </Segment> |
| <Segment StartTime="2668.313" EndTime="2672.607" Participant="mn017"> |
| So you want to get that annotation and bring it back in line with your data. |
| </Segment> |
| <Segment StartTime="2668.335" EndTime="2668.824" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="2672.357" EndTime="2673.852" Participant="fe008"> |
| <VocalSound Description="inbreath"/> O_K, then - |
| </Segment> |
| <Segment StartTime="2672.602" EndTime="2675.860" Participant="me011"> |
| But unfortunately they've also hand-edited it. |
| </Segment> |
| <Segment StartTime="2672.960" EndTime="2673.812" Participant="mn017"> |
| O_K? |
| </Segment> |
| <Segment StartTime="2674.029" EndTime="2681.060" Participant="fe016"> |
| But they've also - Exactly. And so that's exactly what we should - somehow when you distribute the data, say that - you know, |
| </Segment> |
| <Segment StartTime="2675.920" EndTime="2676.670" Participant="mn017"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2676.637" EndTime="2677.004" Participant="me011"> |
| Yep. |
| </Segment> |
| <Segment StartTime="2677.201" EndTime="2677.841" Participant="mn017"> |
| Right. |
| </Segment> |
| <Segment StartTime="2678.187" EndTime="2678.925" Participant="fe008"> |
| Well, then the - |
| </Segment> |
| <Segment StartTime="2680.423" EndTime="2681.284" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2681.350" EndTime="2686.094" Participant="fe016"> |
| that - have some way of knowing how to merge it back in and asking people to try to do that. |
| </Segment> |
| <Segment StartTime="2684.991" EndTime="2687.782" Participant="me018"> |
| What's - what's wrong with <Pause/> doing times? I - |
| </Segment> |
| <Segment StartTime="2687.132" EndTime="2688.885" Participant="fe008"> |
| I agree. That was what I was wondering. |
| </Segment> |
| <Segment StartTime="2687.665" EndTime="2693.974" Participant="fe016"> |
| Uh, yeah, time is the - <VocalSound Description="breath"/> Time is passing! <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2689.474" EndTime="2693.718" Participant="fe008"> |
| Time is unique. You were saying that you didn't think we should - Andreas was saying - Yeah. |
| </Segment> |
| <Segment StartTime="2689.989" EndTime="2690.543" Participant="me011"> |
| Well, |
| </Segment> |
| <Segment StartTime="2691.882" EndTime="2694.654" Participant="mn017"> |
| Time - time - times are |
| </Segment> |
| <Segment StartTime="2692.052" EndTime="2693.360" Participant="me001"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2694.840" EndTime="2697.146" Participant="me011"> |
| what if they haven't notated with them, times? |
| </Segment> |
| <Segment StartTime="2695.010" EndTime="2696.129" Participant="mn017"> |
| ephemeral. |
| </Segment> |
| <Segment StartTime="2695.379" EndTime="2695.834" Participant="fe016"> |
| <Uncertain> Yeah. </Uncertain> <Comment Description="whispered"/> |
| </Segment> |
| <Segment StartTime="2696.071" EndTime="2698.830" Participant="fe016"> |
| He - he's a language modeling person, though. <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="2696.129" EndTime="2697.791" Participant="mn017"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2696.449" EndTime="2697.853" Participant="me001"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2697.791" EndTime="2698.468" Participant="mn017"> |
| Um - |
| </Segment> |
| <Segment StartTime="2697.962" EndTime="2706.053" Participant="me011"> |
| So - so imagine - I think his - his example is a good one. Imagine that this person who developed the corpus of the referring expressions didn't include time. |
| </Segment> |
| <Segment StartTime="2704.556" EndTime="2705.208" Participant="mn017"> |
| Mm-hmm. |
| </Segment> |
| <Segment StartTime="2706.170" EndTime="2706.577" Participant="mn017"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2706.525" EndTime="2707.240" Participant="fe008"> |
| Ach! <Comment Description="while laughing"/> |
| </Segment> |
| <Segment StartTime="2706.680" EndTime="2712.740" Participant="me011"> |
| He included references to <Emphasis> words. </Emphasis> He said that at this <Emphasis> word </Emphasis> is when - when it happened. Or she. |
| </Segment> |
| <Segment StartTime="2708.064" EndTime="2708.973" Participant="fe008"> |
| Well, then - |
| </Segment> |
| <Segment StartTime="2708.605" EndTime="2709.118" Participant="mn017"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2710.879" EndTime="2711.571" Participant="mn017"> |
| Or she. |
| </Segment> |
| <Segment StartTime="2711.255" EndTime="2715.437" Participant="fe008"> |
| But then couldn't you just indirectly figure out the time <Pause/> tied to the <Emphasis> word? </Emphasis> |
| </Segment> |
| <Segment StartTime="2711.571" EndTime="2713.250" Participant="mn017"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2712.007" EndTime="2713.273" Participant="fe016"> |
| But still they - |
| </Segment> |
| <Segment StartTime="2713.720" EndTime="2714.974" Participant="fe016"> |
| Exactly. Yeah. |
| </Segment> |
| <Segment StartTime="2714.352" EndTime="2716.843" Participant="me011"> |
| Sure. But what if - what if they change the <Emphasis> words? </Emphasis> |
| </Segment> |
| <Segment StartTime="2716.632" EndTime="2720.910" Participant="fe008"> |
| Not - Well, but you'd have <Emphasis> some </Emphasis> anchoring point. He couldn't have changed <Emphasis> all </Emphasis> the words. |
| </Segment> |
| <Segment StartTime="2717.715" EndTime="2719.554" Participant="me011"> |
| <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2720.412" EndTime="2723.142" Participant="me018"> |
| But can they change the words without changing the <Emphasis> time </Emphasis> of the word? |
| </Segment> |
| <Segment StartTime="2721.018" EndTime="2722.695" Participant="me011"> |
| <VocalSound Description="mouth"/> Sure. But they could have changed it a little. |
| </Segment> |
| <Segment StartTime="2723.792" EndTime="2732.742" Participant="me011"> |
| The - the <Emphasis> point </Emphasis> is, that - that they may have annotated it off a <Emphasis> word </Emphasis> transcript that isn't the same as <Emphasis> our </Emphasis> word transcript, so how do you merge it back in? I understand what you're saying. <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2731.803" EndTime="2732.746" Participant="mn017"> |
| Mmm. Mm-hmm. |
| </Segment> |
| <Segment StartTime="2732.742" EndTime="2734.766" Participant="me011"> |
| And I - I guess the <Emphasis> answer </Emphasis> is, |
| </Segment> |
| <Segment StartTime="2733.082" EndTime="2736.920" Participant="mn017"> |
| <NonVocalSound Description="mike noise"/> <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2733.453" EndTime="2733.790" Participant="fe016"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2735.221" EndTime="2738.195" Participant="me011"> |
| um, it's gonna be <Emphasis> different </Emphasis> every <Emphasis> time. </Emphasis> |
| </Segment> |
| <Segment StartTime="2736.678" EndTime="2737.271" Participant="fe016"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2738.195" EndTime="2739.890" Participant="me011"> |
| It's j- it's just gonna be - <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2738.260" EndTime="2738.814" Participant="fe016"> |
| Yeah. |
| </Segment> |
| <Segment StartTime="2738.638" EndTime="2739.331" Participant="fe008"> |
| Yeah. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2739.760" EndTime="2741.840" Participant="fe016"> |
| You only know the boundaries of the - |
| </Segment> |
| <Segment StartTime="2740.133" EndTime="2744.376" Participant="me011"> |
| I- it's <Emphasis> exactly </Emphasis> what I said before, which is that "what do you mean by "merge"?" |
| </Segment> |
| <Segment StartTime="2744.376" EndTime="2748.983" Participant="me011"> |
| So in this case where you have the words and you don't have the times, well, what do you <Emphasis> mean </Emphasis> by "merge"? |
| </Segment> |
| <Segment StartTime="2748.983" EndTime="2752.110" Participant="me011"> |
| If you tell me what you <Emphasis> mean, </Emphasis> I can write a program to <Emphasis> do </Emphasis> it. |
| </Segment> |
| <Segment StartTime="2749.071" EndTime="2749.392" Participant="fe016"> |
| Right. |
| </Segment> |
| <Segment StartTime="2751.162" EndTime="2755.979" Participant="fe016"> |
| <VocalSound Description="inbreath"/> Right. You can merge at the level of the representation that the other person preserved and that's it. |
| </Segment> |
| <Segment StartTime="2755.336" EndTime="2757.537" Participant="me011"> |
| Right. And that's about all you can do. |
| </Segment> |
| <Segment StartTime="2755.979" EndTime="2760.530" Participant="fe016"> |
| And beyond that, all you know is - is relative ordering and sometimes even that is wrong. |
| </Segment> |
| <Segment StartTime="2759.442" EndTime="2760.838" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2760.148" EndTime="2765.563" Participant="me011"> |
| <VocalSound Description="inbreath"/> So - so in - so in this one you would have to do a best match between the word sequences, |
| </Segment> |
| <Segment StartTime="2761.036" EndTime="2761.414" Participant="fe016"> |
| So. |
| </Segment> |
| <Segment StartTime="2761.441" EndTime="2762.509" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2765.370" EndTime="2765.851" Participant="mn017"> |
| Mm-hmm. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2765.563" EndTime="2770.490" Participant="me011"> |
| extract the times f- from the best match of theirs to yours, and use that. |
| </Segment> |
| <Segment StartTime="2769.278" EndTime="2772.937" Participant="fe016"> |
| And then infer that their time-marks are somewhere in between. Yeah, exactly. |
| </Segment> |
| <Segment StartTime="2771.160" EndTime="2771.576" Participant="me011"> |
| Right. |
| </Segment> |
| <Segment StartTime="2772.182" EndTime="2779.146" Participant="fe008"> |
| <VocalSound Description="inbreath"/> But it could be that they just - uh, I mean, it could be that they chunked - they - they lost certain utterances and all that stuff, or - |
| </Segment> |
| <Segment StartTime="2777.549" EndTime="2781.410" Participant="me011"> |
| Right, exactly. So it could get very, very ugly. <VocalSound Description="laugh"/> |
| </Segment> |
| <Segment StartTime="2777.637" EndTime="2778.942" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2778.978" EndTime="2780.503" Participant="fe016"> |
| Definitely. Definitely. |
| </Segment> |
| <Segment StartTime="2780.281" EndTime="2780.567" Participant="fe008"> |
| Yeah. <Comment Description="quietly"/> |
| </Segment> |
| <Segment StartTime="2780.503" EndTime="2781.452" Participant="fe016"> |
| <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2781.452" EndTime="2789.127" Participant="fe016"> |
| Alright. Well, I guess, w- I - I didn't want to keep people <Emphasis> too </Emphasis> long and Adam wanted t- people - <Emphasis> I'll </Emphasis> read the digits. If anyone else offers to, that'd be great. And |
| </Segment> |
| <Segment StartTime="2781.726" EndTime="2783.009" Participant="fe008"> |
| That's interesting. <Comment Description="whispered"/> |
| </Segment> |
| <Segment StartTime="2787.404" EndTime="2789.344" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2787.831" EndTime="2788.306" Participant="me011"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2789.344" EndTime="2795.437" Participant="mn017"> |
| Ah, well. <Comment Description="sighing"/> For th- for the - <NonVocalSound Description="noise"/> for the benefit of science we'll read the digits. |
| </Segment> |
| <Segment StartTime="2789.552" EndTime="2789.987" Participant="me011"> |
| <Uncertain> Yeah. </Uncertain> |
| </Segment> |
| <Segment StartTime="2789.621" EndTime="2790.868" Participant="fe016"> |
| if not, I guess - |
| </Segment> |
| <Segment StartTime="2790.521" EndTime="2792.903" Participant="me011"> |
| More digits, the better. O_K, this is |
| </Segment> |
| <Segment StartTime="2791.792" EndTime="2798.206" Participant="fe016"> |
| Thanks - thanks a lot. <NonVocalSound Description="spike on "lot""/> It's really helpful. I mean, Adam and Don <NonVocalSound Description="spike on "Don""/> will sort of meet and I think that's great. Very useful. |
| </Segment> |
| <Segment StartTime="2795.437" EndTime="2796.465" Participant="mn017"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2796.001" EndTime="2796.911" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2796.465" EndTime="2797.412" Participant="mn017"> |
| <VocalSound Description="breath-laugh"/> |
| </Segment> |
| <Segment StartTime="2797.890" EndTime="2800.270" Participant="me011" DigitTask="true"> |
| transcript two seven three one, two seven five O_. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2801.110" EndTime="2802.603" Participant="me011" DigitTask="true"> |
| eight five O_ five one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2803.023" EndTime="2804.678" Participant="me011" DigitTask="true"> |
| nine five zero six one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2805.038" EndTime="2806.008" Participant="me011" DigitTask="true"> |
| O_ seven eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2806.008" EndTime="2806.631" Participant="me011" DigitTask="true"> |
| zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2806.454" EndTime="2807.344" Participant="fe016"> |
| <Uncertain> Go next. </Uncertain> <Comment Description="whispered"/> |
| </Segment> |
| <Segment StartTime="2806.631" EndTime="2807.598" Participant="me011" DigitTask="true"> |
| one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2806.744" EndTime="2808.524" Participant="me001"> |
| <VocalSound Description="breath"/> |
| </Segment> |
| <Segment StartTime="2807.598" EndTime="2808.565" Participant="me011" DigitTask="true"> |
| two zero two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2809.005" EndTime="2811.212" Participant="me011" DigitTask="true"> |
| four four three four four two seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2811.515" EndTime="2812.628" Participant="me011" DigitTask="true"> |
| five five six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2812.628" EndTime="2814.866" Participant="me011" DigitTask="true"> |
| six six O_ O_ nine eight five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2815.200" EndTime="2816.421" Participant="me011" DigitTask="true"> |
| seven six five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2816.789" EndTime="2817.840" Participant="me011" DigitTask="true"> |
| eight eight four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2817.840" EndTime="2818.420" Participant="me011" DigitTask="true"> |
| nine <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2818.619" EndTime="2820.268" Participant="me011" DigitTask="true"> |
| O_ O_ one nine one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2820.591" EndTime="2822.094" Participant="me011" DigitTask="true"> |
| one five six one <Emphasis> four </Emphasis> <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2822.403" EndTime="2824.909" Participant="me011" DigitTask="true"> |
| two seven four eight five six zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2825.308" EndTime="2826.220" Participant="me011" DigitTask="true"> |
| three four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2826.220" EndTime="2827.315" Participant="me011" DigitTask="true"> |
| four five O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2827.315" EndTime="2828.116" Participant="me011" DigitTask="true"> |
| five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2828.116" EndTime="2828.755" Participant="me011" DigitTask="true"> |
| six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2828.755" EndTime="2829.730" Participant="me011" DigitTask="true"> |
| seven. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2829.860" EndTime="2833.356" Participant="me018" DigitTask="true"> |
| Transcript one five one one dash one five three zero. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2830.056" EndTime="2830.986" Participant="fe016" CloseMic="false"> |
| <Uncertain> Let me - </Uncertain> |
| </Segment> |
| <Segment StartTime="2834.072" EndTime="2834.736" Participant="me018" DigitTask="true"> |
| zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2835.416" EndTime="2837.713" Participant="me018" DigitTask="true"> |
| one O_ six six seven three nine <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2838.344" EndTime="2840.409" Participant="me018" DigitTask="true"> |
| three one zero five eight eight one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2841.282" EndTime="2842.655" Participant="me018" DigitTask="true"> |
| four five three eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2843.287" EndTime="2844.981" Participant="me018" DigitTask="true"> |
| five nine four three four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2844.867" EndTime="2845.386" Participant="me011"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2845.371" EndTime="2847.635" Participant="me018" DigitTask="true"> |
| six six O_ one three four eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2848.216" EndTime="2848.834" Participant="me018" DigitTask="true"> |
| seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2849.363" EndTime="2851.573" Participant="me018" DigitTask="true"> |
| eight zero one one two four three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2852.280" EndTime="2853.675" Participant="me018" DigitTask="true"> |
| eight O_ seven two nine. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2854.916" EndTime="2855.757" Participant="me018"> |
| Scratch that. <VocalSound Description="inbreath"/> |
| </Segment> |
| <Segment StartTime="2855.757" EndTime="2857.238" Participant="me018" DigitTask="true"> |
| nine O_ seven two nine <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2858.062" EndTime="2859.127" Participant="me018" DigitTask="true"> |
| zero one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2859.913" EndTime="2861.378" Participant="me018" DigitTask="true"> |
| one three nine four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2861.762" EndTime="2862.880" Participant="me018" DigitTask="true"> |
| two two five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2863.367" EndTime="2864.721" Participant="me018" DigitTask="true"> |
| three eight seven O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2864.307" EndTime="2867.670" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2865.037" EndTime="2865.784" Participant="me018" DigitTask="true"> |
| four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2866.154" EndTime="2866.794" Participant="me018" DigitTask="true"> |
| five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2867.391" EndTime="2868.518" Participant="me018" DigitTask="true"> |
| seven two two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2868.731" EndTime="2869.760" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="2869.017" EndTime="2870.120" Participant="me018" DigitTask="true"> |
| eight one zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2870.526" EndTime="2871.843" Participant="me018" DigitTask="true"> |
| nine five one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2872.340" EndTime="2873.371" Participant="me018" DigitTask="true"> |
| O_ nine eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2873.584" EndTime="2874.450" Participant="me018" DigitTask="true"> |
| zero. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2876.380" EndTime="2879.237" Participant="fe008" DigitTask="true"> |
| Transcript one four five one, one four seven O_. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2879.649" EndTime="2881.298" Participant="fe008" DigitTask="true"> |
| nine five two O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2881.521" EndTime="2883.213" Participant="fe008" DigitTask="true"> |
| O_ three O_ four seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2883.638" EndTime="2884.057" Participant="fe008"> |
| s- |
| </Segment> |
| <Segment StartTime="2884.057" EndTime="2886.512" Participant="fe008" DigitTask="true"> |
| zero seven one three six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2887.285" EndTime="2888.640" Participant="fe008" DigitTask="true"> |
| one nine O_ two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2889.089" EndTime="2889.911" Participant="fe008" DigitTask="true"> |
| two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2889.911" EndTime="2892.269" Participant="fe008" DigitTask="true"> |
| three O_ six O_ seven three two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2890.703" EndTime="2891.692" Participant="me011"> |
| Oh, right. <Comment Description="whispered"/> |
| </Segment> |
| <Segment StartTime="2892.652" EndTime="2894.514" Participant="fe008" DigitTask="true"> |
| four zero one zero four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2895.115" EndTime="2896.937" Participant="fe008" DigitTask="true"> |
| six two O_ one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2897.344" EndTime="2898.547" Participant="fe008" DigitTask="true"> |
| seven five one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2899.007" EndTime="2901.347" Participant="fe008" DigitTask="true"> |
| eight five one two four five five. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2901.896" EndTime="2903.490" Participant="fe008" DigitTask="true"> |
| nine five seven one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2903.756" EndTime="2905.263" Participant="fe008" DigitTask="true"> |
| O_ O_ four nine <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2905.573" EndTime="2906.334" Participant="fe008" DigitTask="true"> |
| zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2906.914" EndTime="2909.100" Participant="fe008" DigitTask="true"> |
| one zero eight one seven nine zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2909.807" EndTime="2911.332" Participant="fe008" DigitTask="true"> |
| three three six one three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2911.649" EndTime="2913.323" Participant="fe008" DigitTask="true"> |
| four four five eight zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2913.772" EndTime="2915.263" Participant="fe008" DigitTask="true"> |
| five two five one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2915.427" EndTime="2916.668" Participant="fe008" DigitTask="true"> |
| six six zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2916.862" EndTime="2918.565" Participant="fe008" DigitTask="true"> |
| seven seven four three seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2918.691" EndTime="2919.188" Participant="fe008" DigitTask="true"> |
| eight. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2919.885" EndTime="2920.241" Participant="fe016"> |
| <VocalSound Description="mouth"/> |
| </Segment> |
| <Segment StartTime="2920.609" EndTime="2924.453" Participant="fe016" DigitTask="true"> |
| Transcript number three three three one, three three five zero. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2925.040" EndTime="2927.372" Participant="fe016" DigitTask="true"> |
| one three seven seven eight one zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2927.917" EndTime="2929.674" Participant="fe016" DigitTask="true"> |
| two five nine five five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2929.991" EndTime="2930.853" Participant="fe016" DigitTask="true"> |
| three eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2931.125" EndTime="2932.586" Participant="fe016" DigitTask="true"> |
| four eight nine two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2932.923" EndTime="2933.565" Participant="fe016" DigitTask="true"> |
| five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2933.975" EndTime="2934.892" Participant="fe016" DigitTask="true"> |
| six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2934.892" EndTime="2936.588" Participant="fe016" DigitTask="true"> |
| seven zero five one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2936.902" EndTime="2937.699" Participant="fe016" DigitTask="true"> |
| nine one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2938.099" EndTime="2939.205" Participant="fe016" DigitTask="true"> |
| O_ three three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2939.603" EndTime="2941.175" Participant="fe016" DigitTask="true"> |
| zero eight five eight three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2941.686" EndTime="2943.060" Participant="fe016" DigitTask="true"> |
| one four O_ one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2943.506" EndTime="2944.055" Participant="fe016" DigitTask="true"> |
| two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2944.466" EndTime="2945.906" Participant="fe016" DigitTask="true"> |
| three O_ three six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2946.163" EndTime="2946.991" Participant="fe016" DigitTask="true"> |
| four zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2947.290" EndTime="2948.546" Participant="fe016" DigitTask="true"> |
| six one three zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2948.846" EndTime="2950.095" Participant="fe016" DigitTask="true"> |
| seven six five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2950.528" EndTime="2951.852" Participant="fe016" DigitTask="true"> |
| eight three seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2951.852" EndTime="2953.179" Participant="fe016" DigitTask="true"> |
| nine nine three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2953.450" EndTime="2954.020" Participant="fe016" DigitTask="true"> |
| O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2954.318" EndTime="2955.200" Participant="fe016" DigitTask="true"> |
| zero. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2955.836" EndTime="2959.798" Participant="mn017" DigitTask="true"> |
| <VocalSound Description="mouth"/> Transcript s- thirty-six eleven, thirty-six thirty. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2960.326" EndTime="2962.682" Participant="mn017" DigitTask="true"> |
| two two five two five nine O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2963.180" EndTime="2964.295" Participant="mn017" DigitTask="true"> |
| three four two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2964.574" EndTime="2965.071" Participant="mn017" DigitTask="true"> |
| four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2965.399" EndTime="2966.006" Participant="mn017" DigitTask="true"> |
| five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2966.006" EndTime="2966.752" Participant="mn017" DigitTask="true"> |
| six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2966.752" EndTime="2968.347" Participant="mn017" DigitTask="true"> |
| seven O_ eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2968.347" EndTime="2969.983" Participant="mn017" DigitTask="true"> |
| nine four eight two six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2969.983" EndTime="2970.892" Participant="mn017" DigitTask="true"> |
| O_ three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2970.892" EndTime="2972.399" Participant="mn017" DigitTask="true"> |
| zero eight eight zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2972.399" EndTime="2973.040" Participant="mn017" DigitTask="true"> |
| one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2973.040" EndTime="2973.661" Participant="mn017" DigitTask="true"> |
| two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2973.661" EndTime="2974.183" Participant="mn017" DigitTask="true"> |
| three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2974.183" EndTime="2975.028" Participant="mn017" DigitTask="true"> |
| four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2975.028" EndTime="2975.973" Participant="mn017" DigitTask="true"> |
| six one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2975.973" EndTime="2977.289" Participant="mn017" DigitTask="true"> |
| seven five seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2977.289" EndTime="2979.706" Participant="mn017" DigitTask="true"> |
| eight four five three seven nine one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2979.706" EndTime="2981.081" Participant="mn017" DigitTask="true"> |
| nine five seven zero. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2981.377" EndTime="2982.587" Participant="mn017" DigitTask="true"> |
| O_ O_ two five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2982.587" EndTime="2983.209" Participant="mn017" DigitTask="true"> |
| zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2983.209" EndTime="2984.210" Participant="mn017" DigitTask="true"> |
| one O_. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2984.490" EndTime="2988.071" Participant="me001" DigitTask="true"> |
| Transcript two nine seven one dash two nine nine zero. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2988.780" EndTime="2990.181" Participant="me001" DigitTask="true"> |
| seven seven nine <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2990.181" EndTime="2991.075" Participant="me001" DigitTask="true"> |
| eight nine <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2991.075" EndTime="2992.348" Participant="me001" DigitTask="true"> |
| nine O_ two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2992.348" EndTime="2992.984" Participant="me001" DigitTask="true"> |
| O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2993.280" EndTime="2994.818" Participant="me001" DigitTask="true"> |
| one one zero one <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2995.291" EndTime="2996.937" Participant="me001" DigitTask="true"> |
| two four five two <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2996.937" EndTime="2998.047" Participant="me001" DigitTask="true"> |
| three seven <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2998.047" EndTime="2999.175" Participant="me001" DigitTask="true"> |
| four four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2999.175" EndTime="2999.938" Participant="me001" DigitTask="true"> |
| five <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="2999.938" EndTime="3000.777" Participant="me001" DigitTask="true"> |
| six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3000.777" EndTime="3001.885" Participant="me001" DigitTask="true"> |
| seven O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3002.116" EndTime="3007.496" Participant="fe016"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
| <Segment StartTime="3002.511" EndTime="3004.387" Participant="me001" DigitTask="true"> |
| nine one four six zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3004.387" EndTime="3006.057" Participant="me001" DigitTask="true"> |
| O_ three nine four <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3006.057" EndTime="3006.881" Participant="me001" DigitTask="true"> |
| zero <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3007.306" EndTime="3009.574" Participant="me001" DigitTask="true"> |
| one eight one nine nine five eight <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3010.241" EndTime="3012.264" Participant="me001" DigitTask="true"> |
| two eight three seven five nine O_ <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3012.264" EndTime="3013.077" Participant="me001" DigitTask="true"> |
| three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3013.077" EndTime="3014.669" Participant="me001" DigitTask="true"> |
| four O_ seven three <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3014.981" EndTime="3016.774" Participant="me001" DigitTask="true"> |
| six two six two six <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3015.158" EndTime="3015.447" Participant="mn017"> |
| <Comment Description="microphone is turned off"/> |
| </Segment> |
| <Segment StartTime="3016.774" EndTime="3018.760" Participant="me001" DigitTask="true"> |
| seven three, zero one. <Comment Description="Digits"/> |
| </Segment> |
| <Segment StartTime="3020.243" EndTime="3021.371" Participant="me001"> |
| <NonVocalSound Description="mike noise"/> <Comment Description="removing mike"/> |
| </Segment> |
| <Segment StartTime="3020.591" EndTime="3021.968" Participant="fe008"> |
| <NonVocalSound Description="mike noise"/> |
| </Segment> |
|
|
| </Transcript> |
|
|
| </Meeting> |
|
|
|
|
