https://github.com/johnvinyard/music-synthesis

Models, training and evaluation code for a two-stage music synthesis pipeline
https://github.com/johnvinyard/music-synthesis

audio deep-learning generative-adversarial-network machine-learning music-synthesis

Last synced: about 1 month ago
JSON representation

Models, training and evaluation code for a two-stage music synthesis pipeline

• Host: GitHub
• URL: https://github.com/johnvinyard/music-synthesis
• Owner: JohnVinyard
• Created: 2019-12-09T15:03:10.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2020-04-23T17:46:33.000Z (almost 5 years ago)
• Last Synced: 2024-11-10T04:31:47.672Z (3 months ago)
• Topics: audio, deep-learning, generative-adversarial-network, machine-learning, music-synthesis
• Language: Jupyter Notebook
• Homepage: https://johnvinyard.github.io/synthesis/2020/04/01/gan-vocoder.html
• Size: 171 MB
• Stars: 2
• Watchers: 4
• Forks: 0
• Open Issues: 11
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Music Synthesis

In this repo, I'm developing models for a two-stage music synthsis pipeline.

1. A generative model that can produce sequences of low-frequency audio features, 

   such as a mel spectrogram, or a sequence of chroma and MFCC features.

2. A conditional generative model that can produce raw audio from the 

   low-frequency features.

![spectrogram](https://generation-report-filterbankmultiscaleexperiment.s3.amazonaws.com/35f29c_feature.png)

![audio](https://music-vocoder-blog-post.s3.amazonaws.com/audio_full.png)

The second stage is inspired by papers developing spectrogram-to-speech vocoders 

such as:

 

- [MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis](https://arxiv.org/abs/1910.06711)

- [High Fidelity Speech Synthesis with Adversarial Networks](https://arxiv.org/abs/1909.11646)

You can read more about early experiments developing models for the second 

stage in [this blog post](https://johnvinyard.github.io/synthesis/2020/04/01/gan-vocoder.html).