faroit / awesome-python-scientific-audio Goto Github PK

https://github.com/AppliedAcousticsChalmers/ReTiSAR

Would be happy to add it, but I am not sure if it too narrow.

• pystoi
• speex (seems not maintained anymore)
• python-pesq

https://github.com/snakers4/silero-vad

https://github.com/midas-research/audino

https://github.com/fgnt/nara_wpe

https://github.com/persephone-tools/persephone

https://github.com/DeepSpectrum/DeepSpectrum
https://github.com/marl/openl3

https://github.com/NVIDIA/OpenSeq2Seq

If a software packages was reviewed separately for software quality, a tagged should be added to indicate that a package is of potential high quality

https://github.com/audioanalytic/psds_eval

in room acoustics library is possible to compute the RIR for a spherical microphone (e.g. eigenmike) and is there a way to compute spherical harmonic coefficients given some signal inputs?

Seems like great implementation of ITU-R BS.1770-4 to measure loudness by @csteinmetz1

website

I implemented a collection of multi-pitch/chord detection algorithms: https://github.com/sevagh/chord-detection

My repo is relatively popular and maybe it's a good addition.

https://github.com/amsehili/auditok

Open-Source Practices for Music Signal Processing Research: Recommendations for Transparent, Sustainable, and Reproducible Audio Research

https://ieeexplore.ieee.org/abstract/document/8588406/

Hi, great idea collecting these resources for Python, thank you.

I have a similar repo, dedicated to open source forced aligners (not just Python):

https://github.com/pettarin/forced-alignment-tools

and some of them --- including my own aeneas --- are written (at least partially) in Python. So my question is: do you take pull requests?

1. If yes, would you mind creating a "forced aligners" subsection, or do you prefer filing them under the current "speech processing"?
2. If not, would you mind adding those you deem interesting (feel free to copy from the list linked above)?

(BTW, you might want to state your contribution policy somewhere in the README, so that you will not get similar queries later.)

Some of the tools have publications so we might hot-link them directly from this list

https://github.com/MTG/DCASE-models

https://github.com/ynop/audiomate

https://github.com/SuperKogito/spafe

https://pypi.org/project/audiomath/

See here: https://github.com/TUT-ARG/sed_eval

https://github.com/alexanderlerch/pyACA

https://github.com/jonashaag/RealRIRs

add: https://github.com/audeering/audtorch

https://github.com/furkanyesiler/acoss

https://github.com/emilio-molina/audio_degrader

https://github.com/r9y9/pysptk

https://github.com/pzelasko/lhotse

https://github.com/SuperKogito/pydiogment

https://github.com/chipmuenk/pyfda

https://github.com/iver56/audiomentations

https://github.com/CPJKU/partitura

https://github.com/lowerquality/gentle

This is a suggestion to add the missing link to PyPi for Essentia:

• Main package: https://pypi.org/project/essentia/
• Including additional TensorFlow wrapper algorithms for inference with pre-trained models): https://pypi.org/project/essentia-tensorflow/

Both packages are Manylinux wheels, only available for Linux x86_64.

https://github.com/spotify/klio

possible add to ML section

• more focus on deep learning
• reorganize feature extraction as proposed [by @yomguy](* #21 (comment))
• add more ressources and intros
• indicate the top_n packages one would need to get started

https://github.com/cokelaer/spectrum

Awesome Criteria (proposal)

• The package is actively maintained
• Ideally it runs on python 2.x and python 3
• The package/algorithm is generic and therefore useful for more than reproducing a single publication
• package is listed on pypi

https://github.com/resemble-ai/Resemblyzer

possibly add https://github.com/mozilla/DeepSpeech

https://www.springer.com/gp/book/9783030512484

https://github.com/DeepSpectrum/DeepSpectrum
https://github.com/marl/openl3`

https://github.com/salu133445/muspy

@salu133445

https://github.com/KinWaiCheuk/nnAudio

nnAudio is a pytorch-based toolkit for working with audio spectrograms using 1D Convolutions that supports GPU/TPU-acceleration.

https://github.com/ArjaanAuinger/pyaudiodsptools

https://github.com/novoic/surfboard