WhisperX

What is it • Setup • Example usage

Made by Max Bain • 🌐 https://www.maxbain.com

Whisper-Based Automatic Speech Recognition (ASR) with improved timestamp accuracy using forced alignment.

What is it 🔎

This repository refines the timestamps of openAI's Whisper model via forced aligment with phoneme-based ASR models (e.g. wav2vec2.0), multilingual use-case.

Whisper is an ASR model developed by OpenAI, trained on a large dataset of diverse audio. Whilst it does produces highly accurate transcriptions, the corresponding timestamps are at the utterance-level, not per word, and can be inaccurate by several seconds.

Phoneme-Based ASR A suite of models finetuned to recognise the smallest unit of speech distinguishing one word from another, e.g. the element p in "tap". A popular example model is wav2vec2.0.

Forced Alignment refers to the process by which orthographic transcriptions are aligned to audio recordings to automatically generate phone level segmentation.

Setup ⚙️

Install this package using

pip install git+https://github.com/m-bain/whisperx.git

You may also need to install ffmpeg, rust etc. Follow openAI instructions here https://github.com/openai/whisper#setup.

Example usage💬

English

Run whisper on example segment (using default params)

whisperx examples/sample01.wav

For increased timestamp accuracy, at the cost of higher gpu mem, use a bigger alignment model e.g.

whisperx examples/sample01.wav --model medium.en --align_model WAV2VEC2_ASR_LARGE_LV60K_960H --output_dir examples/whisperx

Result using WhisperX with forced alignment to wav2vec2.0 large:

sample01.mp4

Compare this to original whisper out the box, where many transcriptions are out of sync:

sample_whisper_og.mov

Other Languages

For non-english ASR, it is best to use the large whisper model.

French

whisperx examples/sample_fr_01.wav --model large --language fr --align_model VOXPOPULI_ASR_BASE_10K_FR --output_dir examples/whisperx

sample_fr_01_vis.mov

German

whisperx examples/sample_de_01.wav --model large --language de --align_model VOXPOPULI_ASR_BASE_10K_DE --output_dir examples/whisperx

sample_de_01_vis.mov

Italian

whisperx examples/sample_it_01.wav --model large --language it --align_model VOXPOPULI_ASR_BASE_10K_IT --output_dir examples/whisperx

sample_it_01_vis.mov

Japanese

whisperx --model large --language ja examples/sample_ja_01.wav  --align_model jonatasgrosman/wav2vec2-large-xlsr-53-japanese --output_dir examples/whisperx --align_extend 2

sample_ja_01_vis.mov

Limitations ⚠️

• Not thoroughly tested, especially for non-english, results may vary -- please post issue to let me know the results on your data
• Whisper normalises spoken numbers e.g. "fifty seven" to arabic numerals "57". Need to perform this normalization after alignment, so the phonemes can be aligned. Currently just ignores numbers.
• Assumes the initial whisper timestamps are accurate to some degree (within margin of 2 seconds, adjust if needed -- bigger margins more prone to alignment errors)
• Hacked this up quite quickly, there might be some errors, please raise an issue if you encounter any.

Coming Soon 🗓

[x] Multilingual init

[x] Subtitle .ass output

[ ] Automatic align model selection based on language detection

[ ] Incorporating word-level speaker diarization

[ ] Inference speedup with batch processing

Contact

Contact maxbain[at]robots[dot]ox[dot]ac[dot]uk if using this commerically.

Acknowledgements 🙏

Of course, this is mostly just a modification to openAI's whisper. As well as accreditation to this PyTorch tutorial on forced alignment

Citation

If you use this in your research, just cite the repo,

@misc{bain2022whisperx,
  author = {Bain, Max},
  title = {WhisperX},
  year = {2022},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/m-bain/whisperX}},
}

as well as the whisper paper,

@article{radford2022robust,
  title={Robust speech recognition via large-scale weak supervision},
  author={Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya},
  journal={arXiv preprint arXiv:2212.04356},
  year={2022}
}

and any alignment model used, e.g. wav2vec2.0.

@article{baevski2020wav2vec,
  title={wav2vec 2.0: A framework for self-supervised learning of speech representations},
  author={Baevski, Alexei and Zhou, Yuhao and Mohamed, Abdelrahman and Auli, Michael},
  journal={Advances in Neural Information Processing Systems},
  volume={33},
  pages={12449--12460},
  year={2020}
}