mmorise/World

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

1338 stars · 263 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

WORLD is a C++ library for high-quality speech analysis, manipulation, and synthesis based on a vocoder design. It estimates F0 (via DIO/Harvest), spectral envelope (CheapTrick), and band aperiodicity (D4C), and can resynthesize speech from those parameters.

Use cases

• estimate fundamental frequency from speech audio
• extract spectral envelope and aperiodicity from a waveform
• resynthesize speech from vocoder parameters
• modify pitch or timbre of speech recordings
• use a vocoder as an acoustic feature extractor for TTS or voice conversion
• analyze speech in real-time applications

When to choose

• you need a patent-free, BSD-licensed vocoder with C/C++ integration
• you want high-quality F0, spectral envelope, and aperiodicity estimation
• you are building TTS, voice conversion, or singing synthesis systems
• you need a well-established vocoder with wrappers in Python, C#, Swift, Java, and JS

When to avoid

• you need a differentiable vocoder for end-to-end deep learning training (consider diffsptk or neural vocoders)
• you want a ready-made application or GUI rather than a library
• you need official support for languages beyond C/C++ (wrappers are third-party)
• you require extensive tutorials or documentation

Facets

library · maturity stable

audio-processing speech-recognition machine-learning speech-processing machine-learning cpp cross-platform vocoder speech-synthesis f0-estimation spectral-envelope aperiodicity signal-processing audio

Member repositories

For agents

markdown · JSON · MCP: product_card(name="mmorise/World")