audio-alignment

A fast python library for aligning similar audio snippets passed in as NumPy arrays

Verify that a subtitle actually matches and syncs with a video using audio-based timing analysis. Detect wrong subtitles, find the correct timing offset, and automatically sync them.

Estimation of audio alignment relative to a given reference.

Align various Sanskrit texts and audio

Multi-track audio timebase estimation, alignment and clock-drift correction for recordings from independent devices. TimeMap abstraction, GCC-PHAT fine alignment, drift/changepoint estimation, multi-track graph solver, validation, Adobe Audition SESX export, and a dedicated MP4 wireless-microphone channel-sync pipeline.

Python CLI to facilitate making karaoke videos

Local-first toolkit for fixed-camera band rehearsal videos: YAMNet song/highlight/fun-moment detection, multicam-style auto editing on virtual cameras, lossless raw-cut export, and a reusable audio-to-video alignment library for swapping in a clean recorder take of the same performance.

MMS Forced alignment package for JavaScript, TypeScript and Node.js: word-level timestamps from audio + transcript for captions, subtitles and karaoke. Meta MMS (wav2vec2) via ONNX Runtime with a CTC Viterbi decoder. Hindi, Hinglish and 20 Indic languages. No Python.

歌词字幕对齐命令行工具：demucs 起唱点提取 + 带约束单调动态规划，中位偏差 0.00s（人工校验真值对照）

LAE (Layered Audio Extension): hi-res audio on top of a bit-exact 16-bit CD layer that every player plays. Lossy tier LAEY; lossless tier LAES in design. Reference encoder and decoder, alignment to an existing CD (the CD stays unchanged), CD-Extra packaging (lae_cdx), C decoder library and foobar2000 component. Windows and Linux (x86, ARM64).

Danmaku Studio 弹幕一站式解决方案

🎧 Align audio to text effortlessly with precise word-level timing using OpenAI Whisper, outputting clean JSON for seamless integration.

Open toolkit for human-directed AI music-video production: storyboards, animatics, prompts, media review, audio alignment, lip-sync research and GPU workflows.

Non-destructive WAV/MP3 and TextGrid timeline alignment for Praat 7 on macOS, with gapless decoding and multi-pair opening.

🕺 Compare your dance take against a reference video and get plain-language coaching tips. On-device pose estimation, audio alignment, and a phrase-bank engine in Swift.

Open word-by-word timestamps for the whole Qur'an — Muhammad Al-Luhaidan (محمد اللحيدان), 114 surahs, quran.com/QUL format · توقيت كلمات القرآن كلمة بكلمة · CC0

Align and graft dubbed audio across media releases using distributed audio anchors.

Audio + transcript → sentence-level timestamps, JSON, and a browser QA page. Built for non-verbatim transcripts (BBC Learning English style).

Local-first guitar practice with MusicXML import, PDF tablature recognition, automatic audio alignment, synchronized score following, and slow-speed looping.