Add third video amplification sampling
The following link demonstrates the high consistency between the image generated video and the reference image when using triple sampling
Can you create a workflow with three samplings
oh interesting concept ;-) will take a look at the video and try make a variant of that ;-)
I did try it, after recreating it. I didnt notice any immediate huge benefits, and the 3rd sampler can be quite slow.
But i'll clean it up a little, so its less "spaghetti mess", and upload so you can try it too ;-)
The main difference i think might be the image guider node (that is a bit different than the usual image in-place node) . But I'll try some more, if any improvements comes from that node, or the 3 sampler steps.
Edit:
That being said, I think it might give much better texture/details with 3 steps. It can give some quite nice results indeed ;-)
Did a few more test runs