OpenVINO™ AI Plugins for Audacity* 🤘

A set of AI-enabled effects, generators, and analyzers for Audacity®. These AI features run 100% locally on your PC 💻 -- no internet connection necessary! OpenVINO™ is used to run AI models on supported accelerators found on the user's system such as CPU, GPU, and NPU.

• Music Separation🎵 -- Separate a mono or stereo track into individual stems -- Drums, Bass, Vocals, & Other Instruments.
• Noise Suppression🧹 -- Removes background noise from an audio sample.
• Music Generation & Continuation🎶 -- Uses MusicGen LLM to generate snippets of music, or to generate a continuation of an existing snippet of music.
• Whisper Transcription🎤 -- Uses whisper.cpp to generate a label track containing the transcription or translation for a given selection of spoken audio or vocals.
• Super Resolution✨ -- Upscales and enriches audio for improved clarity and detail.

Installation 💾

Go here to find installation packages & instructions for the latest Windows release.

Build Instructions 🔨

• Windows Build Instructions
• Linux Build Instructions

Help, Feedback, & Bug Reports 🙋‍♂️

We welcome you to submit an issue here for

• Questions
• Bug Reports
• Feature Requests
• Feedback of any kind -- how can we improve this project?

Contribution 🤝

Your contributions are welcome and valued, no matter how big or small. Feel free to submit a pull-request!

Acknowledgements & Citations 🙏

• Audacity® development team & Muse Group-- Thank you for your support!
• Audacity® GitHub -- https://github.com/audacity/audacity
• Whisper transcription & translation analyzer uses whisper.cpp (with OpenVINO™ backend): https://github.com/ggerganov/whisper.cpp
• Music Generation & Continuation use MusicGen model, from Meta.
  • We currently have support for MusicGen-Small, and MusicGen-Small-Stereo
  • The txt-to-music pipelines were ported from python to C++, referencing logic from the Hugging Face transformers project: https://github.com/huggingface/transformers
• Music Separation effect uses Meta's Demucs v4 model (https://github.com/facebookresearch/demucs), which has been ported to work with OpenVINO™
• Noise Suppression:
  • noise-suppression-denseunet-ll: from OpenVINO™'s Open Model Zoo: https://github.com/openvinotoolkit/open_model_zoo
  • DeepFilterNet2 & DeepFilterNet3:
    • Ported the models & pipeline from here: https://github.com/Rikorose/DeepFilterNet
    • We also made use of @grazder's fork / branch (https://github.com/grazder/DeepFilterNet/tree/torchDF-changes) to better understand the Rust implementation, and so we also based some of our C++ implementation on torch_df_offline.py found here.
    • Citations:

      @inproceedings{schroeter2022deepfilternet2,
      title = {{DeepFilterNet2}: Towards Real-Time Speech Enhancement on Embedded Devices for Full-Band Audio},
      author = {Schröter, Hendrik and Escalante-B., Alberto N. and Rosenkranz, Tobias and Maier, Andreas},
      booktitle={17th International Workshop on Acoustic Signal Enhancement (IWAENC 2022)},
      year = {2022},
      }
        
      @inproceedings{schroeter2023deepfilternet3,
      title = {{DeepFilterNet}: Perceptually Motivated Real-Time Speech Enhancement},
      author = {Schröter, Hendrik and Rosenkranz, Tobias and Escalante-B., Alberto N. and Maier, Andreas},
      booktitle={INTERSPEECH},
      year = {2023},
      }
• Super Resolution feature is a port (python -> C++, pytorch -> OpenVINO IR) of this project: https://github.com/haoheliu/versatile_audio_super_resolution
  • Citations:

    @article{liu2023audiosr,
    title={{AudioSR}: Versatile Audio Super-resolution at Scale},
    author={Liu, Haohe and Chen, Ke and Tian, Qiao and Wang, Wenwu and Plumbley, Mark D},
    journal={arXiv preprint arXiv:2309.07314},
    year={2023}
    }
• OpenVINO™ Notebooks -- We have learned a lot from this awesome set of python notebooks, and are still using it to learn latest / best practices for implementing AI pipelines using OpenVINO™!