Extract Audio from Video (WAV export)

Features

• Extract the full audio track as lossless WAV
• Take a time range: fill in start and end seconds to export just one segment
• Downmix to mono — halves the size with almost no audible difference for speech
• Downsample to 48000, 44100, 22050 or 16000 Hz (16 kHz is the usual speech-recognition input)
• Peak normalisation to −1 dBFS, so quiet phone recordings need no manual gain
• 16-bit or 32-bit float WAV
• A 20-minute cap, refused clearly rather than taking the tab down
• A specific reason when the track cannot be decoded (unsupported codec, or no audio track at all)
• Everything runs locally in the browser; the video is never uploaded

How to use

1. Drop the video file in, or click to choose one
2. Fill in start and end if you only want a segment
3. For speech, tick the mono downmix and drop the sample rate to 16000 Hz
4. Press Start and download the WAV when it finishes

FAQ

Why is the output WAV and not MP3?

Because browsers cannot encode MP3 or AAC — the encoders are a burden in both patents and size. A WAV is a 44-byte header on top of the PCM, a few lines to write correctly, and lossless. The cost is size: roughly ten times an equivalent MP3. So the point of this tool is not format conversion but getting the track out losslessly; for MP3, hand the WAV to any player or transcoder afterwards.

The exported file is huge — what can I do?

Three switches, in order of effect: downmix to mono (halves it), downsample (44.1 kHz to 16 kHz is another drop to about 36%, inaudible for speech), and trim the range (keep only what you need). Ten minutes of 44.1 kHz stereo is 105 MB; with all three applied it is about 19 MB. A smaller file does not mean less content — none of these lose anything, they only remove what human ears cannot resolve for that kind of material.

Why is there a 20-minute cap?

Because decoded samples fill memory: 32-bit float at 44.1 kHz stereo is about 21 MB per minute, so twenty minutes is over 400 MB. Beyond that it is not "slower", it is the tab dying — and it dies after you have waited for the file you just dropped in. So it is refused up front, with a suggestion to trim the segment you need in an editor first.

Some videos yield no audio at all?

Two possibilities. Either the video has no audio track (screen recordings, some drone footage), in which case `decodeAudioData` fails, or the audio codec is one the browser cannot decode — typically PCM from some camcorders, AC-3 in some AVCHD files, or an encrypted track. Both cases are reported specifically rather than as a bare "processing failed". For the second, convert the video to MP4 (H.264 + AAC) in a player first and try again.

Is my video uploaded?

No. Decoding and encoding happen entirely in browser memory; the page makes no requests and works offline. Video usually carries voices and ambient sound, and none of it leaves your device.