# magenta/ddsp

DDSP: Differentiable Digital Signal Processing

Repository: https://github.com/magenta/ddsp
Canonical: https://ross.abutalabs.com/products/ddsp
Homepage: https://magenta.tensorflow.org/ddsp
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-07-09T03:02:15+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 91, release rhythm 8, longevity 100
- inputs: {"age_days": 2423, "days_push": 55, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3344, forks 381 (observed 2026-08-28T04:07:57.008656+00:00)

## What it is
DDSP is a Python library of differentiable digital signal processing components (synthesizers, filters, waveshapers) that can be embedded in deep learning models, particularly as output layers for audio generation. It is developed by Google's Magenta team and powers audio synthesis tasks like timbre transfer and self-supervised pitch detection.

## Use cases
- generate audio from neural network outputs
- convert my voice into a violin with timbre transfer
- train an autoencoder model on audio files
- detect pitch in audio in a self-supervised way
- use interpretable DSP elements as differentiable output layers
- synthesize realistic instrument sounds with neural control

## When to choose
- you need differentiable DSP components inside a TensorFlow deep learning audio model
- you want interpretable audio generation rather than a black-box neural vocoder
- you are doing music or audio research like timbre transfer or pitch detection

## When to avoid
- you need a general-purpose non-differentiable DSP toolkit for production audio processing
- you work in PyTorch or another framework without adaptation effort
- you need real-time low-latency audio synthesis in a DAW or plugin

## Facets
- artifact type: library
- maturity: active
- function: audio-processing, machine-learning, deep-learning, llm-training
- domain: machine-learning, deep-learning, artificial-intelligence
- platform: python, cross-platform
- tags: dsp, audio-synthesis, music, timbre-transfer, tensorflow, neural-audio, synthesizers, pitch-detection, audio

## Member repositories
- magenta/ddsp (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:57.008656+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:41:02.062654+00:00, confidence not recorded.
  - readme: https://github.com/magenta/ddsp (fetched 2026-08-28T04:07:57.008656+00:00, sha 9bf2cb992a17)
  - homepage: https://magenta.tensorflow.org/ddsp (fetched 2026-08-29T09:34:21.719274+00:00, sha b10d2215abcc)
  - registry_pypi: https://pypi.org/pypi/ddsp/json (fetched 2026-08-29T09:34:21.728507+00:00, sha 9832615e6eca)
- Data as of 2026-08-30T08:39:29.467469+00:00.