Video to Audio
Grab the sound out of a video and leave the picture behind.
only the soundtrack is read, the picture is ignored
- Length
- ·
- Sample rate
- ·
- Channels
- ·
- Size
- ·
- Length
- ·
- Sample rate
- ·
- Channels
- ·
- Size
- ·
Only the soundtrack is read, the picture is discarded. The output length matches the video exactly.
Pick MP3 for something you just want to listen to, or WAV if the audio is going into an editor next. Both write without any download.
How to use it
- Drop a video file onto the page. MP4, MOV, M4V or WEBM go straight through, because your browser can already read the audio track inside them.
- The soundtrack is decoded and drawn as a waveform. The picture is ignored entirely; only the audio track is touched.
- Play it back to confirm you have the right file and the right track before you spend time encoding.
- Choose an output format and click Download. MP3 and WAV write immediately; M4A, OGG and FLAC load a converter once.
A video file is a box with two things in it
The extension on a video tells you about the container, not the audio. An .mp4 is an ISO base media file: a structure of nested boxes holding one or more tracks, each track compressed with its own codec, plus a table saying which chunk belongs to which moment. Extracting the sound means reading that table, walking to the audio track, and ignoring everything else. The video frames are never decoded, which is why pulling audio out of a twenty-minute recording takes a second or two rather than the minutes a full transcode would.
Browsers already contain a demultiplexer and a set of audio decoders, because playing video on the web requires both. Handing an MP4 to the audio decoding API takes advantage of exactly that machinery: it finds the AAC track, decodes it, and hands back raw samples. No extra code is needed and nothing large has to be downloaded. The same is true of WEBM, where the audio is Vorbis or Opus.
Where the browser stops helping
Support is narrower than most people expect, and it is not about the file being exotic. Matroska is the clearest example: WEBM is a deliberately restricted subset of Matroska that browsers implement, but a full .mkv (the format most desktop recorders and rips default to) is not something any browser will open, even when the audio inside is ordinary AAC. AVI, WMV and FLV are legacy containers no browser has ever supported.
The other failure mode is a supported container with an unsupported codec inside. An MP4 ripped from a Blu-ray often carries AC-3 or DTS surround audio, which browsers do not license. A ProRes master from an editing suite usually carries uncompressed PCM in a MOV, which some browsers handle and others do not. In both cases the container opens fine and the audio track cannot be read, and the honest answer is to use the converter path, which loads a full ffmpeg build on demand.
Choosing what comes out
Whatever you pick, the soundtrack is being re-encoded, because there is no way to hand a decoded signal back to a new container without compressing it again. If the video's audio was already lossy. And in a phone recording or a download it certainly was. You are on your second generation. Keep the output bitrate at or above whatever the source was using so the second pass has room to work; 192 kbps MP3 is a safe floor for a video that had 128 or 160 kbps AAC in it.
If the audio is going anywhere near an editor, a transcription tool or a podcast workflow, export WAV instead and compress once at the end. And if you only need a fragment (a quote, a piece of music, a sound effect) extract the whole thing here and trim it afterwards, rather than trying to guess boundaries from the video timeline.