Hugging Face's logo

Datasets:
humyn-labs
/
Indic-High-Fidelity-MultiSpeaker-ASR

Tasks:
Automatic Speech Recognition
Modalities:
Audio
Text
Formats:
parquet
Languages:
Hindi
Malayalam
Marathi
Size:
< 1K
Tags:
ASR
Conversational-speech
multi-speaker
indic-languages
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
language
stringlengths
5
12
file_name
stringlengths
11
18
audio
audioduration (s)
285
1.2k
transcript_json
stringlengths
4.68k
33.3k
Assamese
Assamese_1.wav
[{"start": "00:00:02.260", "end": "00:00:23.100", "speaker": "Speaker 1", "text": "নমস্কাৰ! আজি আমি ৱৰ্ক লাইফ বেলেঞ্চ, কৰ্ম জীৱনৰ ভাৰসাম্যতা অৰ্থাৎ কাম আৰু ব্যক্তিগত জীৱনৰ মাজত সমতাৰ বিষয়ে আলোচনা কৰিম। আজিৰ দ্ৰুতগতিৰ আগবাঢ়ি যোৱা জীৱনৰ এই বিষয়টো বহুত গুৰুত্বপূৰ্ণ হৈ পৰিছে।"}, {"start": "00:00:23.180", "end": "00:00:3...
Assamese
Assamese_2.wav
[{"start": "00:00:02.560", "end": "00:00:14.440", "speaker": "Speaker 1", "text": "নমস্কাৰ! আজি আমাৰ আলোচনাৰ বিষয়বস্তু হ'ল ৱৰ্ক লাইফ বেলেঞ্চ, অৰ্থাৎ কৰ্ম জীৱনৰ ভাৰসাম্যতা। আ, প্ৰথমে মই আপোনাক নমস্কাৰ জনাইছোঁ।"}, {"start": "00:00:14.480", "end": "00:00:15.980", "speaker": "Speaker 2", "text": "নমস্কাৰ।"}, {"start": "00...
Bengali
Bengali_1.wav
[{"start": "00:00:00.040", "end": "00:00:24.600", "speaker": "Speaker 1", "text": "শ্রমীকল্যাণ হতে হবে। আর মানুষের ভালোবাসা অর্জন করতে হবে। এই তিনটে কোয়ালিটি আমার মানে তাদের প্রতি সাজেশন থাকলো। আমি তো আর উপদেশ দিতে পারি না, সাজেশন। সততা একটা রাজনীতিবিদকে তার আইডেন্টিফিকেশন।"}, {"start": "00:00:24.640", "end": "00:01:0...
Bengali
Bengali_2.wav
[{"start": "00:00:03.120", "end": "00:03:23.260", "speaker": "Speaker 1", "text": "ঠিক আছে। আমার স্ত্রীকে বললাম, দেখো, এরকম বলছে। বলে, তো একদিন নাই বা বেরোলো তুমি, আমি তো মর্নিং ওয়াক তখন। তো একদিন মর্নিং ওয়াক নাই বা করলে? তাই আমি পরের দিন সকালবেলা পৌনে আটটা নাগাদ এলো। বলে, স্যার, আমি আজ আপনার কাছে এসেছি, আমার স্ত্রী ...
Bengali
Bengali_3.wav
[{"start": "00:00:00.240", "end": "00:02:02.440", "speaker": "Speaker 1", "text": "লড়াই কার? সিপিএমের বিরুদ্ধে চৌত্রিশ বছরের সিপিএমকে ধুলি, মাটির ধুলির সঙ্গে মিশিয়ে দেয়। মমতা ব্যানার্জি এমন একজন ব্যক্তি যিনি দেখিয়ে দিয়েছেন একা। ভারতবর্ষে কেউ দেখাতে পারে নি। কংগ্রেস থেকে বেরিয়ে এসে সবাইকে বলেছিল প্রণব মুখার্জির মত...
Bhojpuri
Bhojpuri_1.wav
[{"start": "00:00:00.320", "end": "00:00:02.920", "speaker": "Speaker 1", "text": "अरे! का हो भाई, कहाँ रहता रह।"}, {"start": "00:00:03.000", "end": "00:00:06.900", "speaker": "Speaker 2", "text": "अरे कहाँ रह जी हो, घरे दुआरे रहतनी, का करी।"}, {"start": "00:00:06.960", "end": "00:00:08.500", "speaker": "Speaker 1", "t...
Bhojpuri
Bhojpuri_2.wav
"[{\"start\": \"00:00:00.240\", \"end\": \"00:00:04.080\", \"speaker\": \"Speaker 1\", \"text\": \"(...TRUNCATED)
Bhojpuri
Bhojpuri_3.wav
"[{\"start\": \"00:00:02.660\", \"end\": \"00:00:06.960\", \"speaker\": \"Speaker 1\", \"text\": \"(...TRUNCATED)
Chattisgarhi
Chattisgarhi_1.wav
"[{\"start\": \"00:00:00.440\", \"end\": \"00:01:03.790\", \"speaker\": \"Speaker 1\", \"text\": \"(...TRUNCATED)
Chattisgarhi
Chattisgarhi_2.wav
"[{\"start\": \"00:00:00.370\", \"end\": \"00:00:04.940\", \"speaker\": \"Speaker 1\", \"text\": \"c(...TRUNCATED)
End of preview. Expand in Data Studio

Dataset Overview

This dataset contains high-quality multi-speaker conversational audio recordings curated for Automatic Speech Recognition (ASR) research across multiple Indic languages.

The dataset includes:

  • Paired audio + timestamped transcripts
  • Natural, non-scripted conversational speech
  • Dual-speaker interactions
  • Segment-level speaker annotations
  • Regionally diverse accents

Audio Specifications

  • Format: WAV (PCM 16-bit)
  • Sampling Rate: 16 kHz
  • Channel: Mono
  • Speech Type: Natural conversational dialogue
  • Recording Style: Dual-speaker spontaneous interaction
  • Typical Duration: 10–30 minutes per recording

All audio files are normalized to ensure consistent duration reporting and playback compatibility.

Supported Languages

This dataset includes conversational speech recordings in:

  • Assamese
  • Odia
  • Bengali
  • Bhojpuri
  • Chhattisgarhi
  • Gujarati
  • Haryanvi
  • Hindi
  • Punjabi
  • Marathi
  • Tamil
  • Kannada
  • Malayalam
  • Telugu

The dataset preserves natural accent variation and conversational speech characteristics.

Speaker Representation

  • Dual-speaker conversational recordings
  • Natural, spontaneous dialogue
  • Regionally representative speakers
  • Conversational turn-taking preserved

Dataset Creation Methodology

Data Collection

Speech data was collected from native speakers across multiple Indian regions to ensure:

  • Accent diversity
  • Natural conversational flow
  • Real-world dialogue patterns
  • Informal and semi-formal speech contexts

Topics include:

  • Everyday life discussions
  • Social interactions
  • Business and finance
  • Public affairs
  • General conversational topics

Transcription Process

  • Manual transcription by native speakers
  • Reviewed for linguistic accuracy
  • Timestamp-level segmentation
  • Speaker-labeled segments
  • Preserves conversational fillers and natural pauses

Each transcript entry contains:

  • start timestamp
  • end timestamp
  • speaker label
  • text content

Intended Use

Designed for:

  • Training and fine-tuning ASR models
  • Conversational ASR benchmarking
  • Speaker diarization research
  • Speaker turn detection
  • Multi-speaker modeling
  • Academic and open research

Out-of-Scope Uses

This dataset is not intended for:

  • Safety-critical or real-time production systems without additional validation
  • Commercial deployment without attribution (CC BY 4.0 required)
  • Medical, clinical, legal, or diagnostic applications

License

Creative Commons Attribution 4.0 International (CC BY 4.0)

📬 Contact For dataset-related queries, please contact:-

[support@humynlabs.ai]
Downloads last month
550
Number of rows:
37
Total file size:
1.06 GB