Guide
Extracting the sound track from a video file
A video file is a container with at least two streams: pictures and sound. Often the sound is what you want. A recorded lecture to listen to on a walk, an interview to transcribe, a song performed on camera, a podcast that was filmed. Pulling it out normally means installing an editor or a command-line tool. This page does it with the decoder your browser already has, and saves the result as a small audio file.
How the extraction works
When you drop a file, the page hands its bytes to the browser’s audio decoder, the same code that plays the video. The decoder returns the raw samples of the sound track, which the page cuts to your times and encodes into WAV, AAC or Opus. The picture is never decoded, so the work is fast even for a large video. All of this happens inside your browser; the video does not leave your computer.
A worked example
Press Sample to load a generated four-second stereo arpeggio and press Extract audio with the defaults (start 0, end filled in as 4.0). The page writes sample-song-audio.wav with 176,400 frames at 44,100 Hz and two channels, 689 KB, which is 4.000 seconds exactly. Change the start to 1 and the end to 3 and the result is 88,200 frames, two seconds. With a real MP4 you would see the video’s own sample rate, usually 48,000 Hz, and the length of its sound track.
Choosing a format
WAV is the safe choice for editing, because every program reads it, and it is lossless with respect to the decoded sound; expect about 10.6 MB per minute for CD-quality stereo. AAC in M4A is a small, widely playable file for listening, around 1 MB a minute at 128 kbps. Opus in WebM is smaller still for speech, and is excellent at 48 to 64 kbps, but fewer players open WebM audio. The format buttons and bitrate box sit below the waveform.
Cutting to the part you need
Use the two time boxes to extract only a passage, such as the song in the middle of a concert recording. If the end is left at 0 or lower than the start, the whole file from the start time is used. The waveform shows the entire sound track so you can see where the speech or music begins. For frame-accurate cuts of the video itself, use the video cutter instead.
Edge cases
A video with no sound track shows an error. Files with several audio tracks give the first one the browser chooses. Variable frame rate or damaged files may fail to decode. Mono tracks stay mono, and a track recorded at an unusual sample rate is converted to the browser’s rate before it reaches you. Extremely long videos need a lot of memory because the whole sound is held uncompressed.
Limits
MP3 output is not provided, only codecs the browser can both decode and encode are used. Lyrics, chapters and tags are not copied. The tool saves one audio file per video and cannot separate dialogue from music. Nothing is uploaded.
Related tools
- Audio Converter changes the extracted file to another format or sample rate.
- Audio Cutter trims the sound further on a waveform.
- Video Cutter & Trimmer cuts the video itself.
- Mute Video does the opposite and keeps the picture without the sound.
Checked against the current version of the Extract Audio from Video on .