IN HIATUS UNTIL FURTHER NOTICE

Jan 31, 2024 

Feb 14 (Canceled)

Purpose of this site:  Share links on topics we talk about

Social media: 

Music and AI Group on LinkedIn

Signal Group

LLMs to try to generating lyrics, melody, chords

https://labs.perplexity.ai/ 

Developers: Seth Forsgren, Hayk Martiros

https://huggingface.co/riffusion/riffusion-model-v1 

https://www.riffusion.com/

Open AI  

Code:  https://github.com/openai/jukebox.  

Paper: https://arxiv.org/pdf/2005.00341.pdf

https://openai.com/research/musenet

Meta

audiocraft

MusicGen - inference time slow for live demo/testing

https://audiocraft.metademolab.com/

Google

MusicLM - https://google-research.github.io/seanet/musiclm/examples/    

Dataset:  https://www.kaggle.com/datasets/googleai/musiccaps

https://arxiv.org/pdf/2301.11325.pdf

https://magenta.tensorflow.org/

Open UNMIX  https://colab.research.google.com/github/pytorch/pytorch.github.io/blob/master/assets/hub/sigsep_open-unmix-pytorch_umx.ipynb 

A whole slew of links!  https://wiki.python.org/moin/PythonInMusic 

Music Synthesis in Python

https://www.youtube.com/watch?v=ROlkhVs15AM 

Reconstructing sounds of different instruments 

https://mapio.github.io/sinuous-violin/   

Matthieu Amiguet - Python for realtime audio processing in a live music context

https://www.youtube.com/watch?v=IdsrZYZ0TI0 

Marc Evanstein / music․py

https://www.youtube.com/watch?v=9Lk4gBBWMqw 

Sound Design in Python - Isaac Roberts - ADC20  - Drop the DAW 

https://www.youtube.com/watch?v=Q40qEg8Yq5c

Librosa Audio and Music Signal Analysis in Python | SciPy 2015 | Brian McFee

https://www.youtube.com/watch?v=MhOdbtPhbLU

Python libraries:

Librosa

pip install librosa

Spleeter

pip install spleeter

https://pypi.org/project/spleeter/

https://colab.research.google.com/github/deezer/spleeter/blob/master/spleeter.ipynb

https://pypi.org/project/omnizart/

https://scicoding.com/pitchdetection/

https://github.com/rupakvignesh/Lyrics-to-Audio-Alignment

https://github.com/MoonInTheRiver/DiffSinger

https://neuralsvb.github.io/

https://github.com/chitralekha18/AutoLyrixAlign

Real time jamming / Decrease inference time / Decrease model size

Metrics to sufficiently distinguish AI-generated music as unique (not too close to the training data)

Transcription that's not character level (not at the note level) that accurately describes a melody

Transcription

Music Transcription with Transformers:

https://magenta.tensorflow.org/transcription-with-transformers   - has examples

https://colab.research.google.com/github/magenta/mt3/blob/main/mt3/colab/music_transcription_with_transformers.ipynb