Getting the sound out of a video
A video file is a container holding separate video and audio tracks. Extracting the audio means pulling that second track out and saving it on its own. That is useful more often than you would expect: turning a recorded lecture into something you can listen to while walking, keeping the audio from an interview for transcription, lifting a backing track out of a performance recording, or simply stripping a 200 MB video down to a 4 MB file when only the words matter.
Which format should you pick?
| Format | Size | Best for |
|---|---|---|
| MP3 | Small | Maximum compatibility — car stereos, old phones, anything |
| M4A (AAC) | Small | Better quality than MP3 at the same size; ideal for modern devices |
| WAV | Very large | Importing into editing or transcription software with no quality loss |
If you are unsure, MP3 at 192 kbps is the safe answer. It plays on literally everything, and at that bitrate the difference from the original is inaudible to almost everyone. Choose M4A if you know the file is staying on modern hardware — it sounds better for the same number of bytes. Reach for WAV only when something downstream demands uncompressed audio, because an hour of it runs to roughly 600 MB.
What the bitrate actually changes
Bitrate is how many bits per second are spent describing the sound. More bits means more detail preserved and a bigger file, in a near-linear relationship: one minute of 192 kbps audio is about 1.4 MB, so an hour is roughly 86 MB. Doubling to 320 kbps doubles that. For speech — lectures, podcasts, interviews — 128 kbps is genuinely enough, and halves your file size against the default. For music, 192 kbps or higher is worth it.
Trimming while you extract
If you only need ninety seconds out of a forty-minute recording, extracting the whole soundtrack and cutting it afterwards wastes both time and disk space. Turn on Trim, set a start and end, and only that section is processed. Times are entered as minutes and seconds, so 2:30 means two minutes thirty. The estimated output size updates as you change the range.
Nothing is uploaded
Extraction runs on your own machine, using FFmpeg compiled to WebAssembly inside the browser tab. Your video is never transmitted anywhere, which matters if the recording is a private meeting, a medical consultation, a legal deposition, or anything else you would not hand to a stranger's server.
Audio extraction is also much faster than video compression, because the video track is discarded rather than re-encoded. Most files finish in a few seconds once the encoder has loaded. The first run downloads that encoder — about 30 MB — and your browser caches it for every visit afterwards.