Take the soundtrack out of a video. Copy the existing track sample for sample, or re-encode to MP3, M4A, AAC, OGG, WAV or FLAC with the cost of that choice stated before you run it.
Initializing in your browser…
Take frames out of a video by index rather than by guesswork. The frame rate is measured, every image is labelled with the frame it came from, and they all come down as one ZIP archive.
Cut a clip out of a video, choosing whether the cut lands on the exact frame (re-encodes to MP4) or leaves the streams untouched (instant and lossless, but only at a keyframe). The result says which ran and reads the real length back off the file.
Convert a video between MP4, WebM, OGG, MOV, AVI and MKV, each naming the codecs it really writes and each producing a file that actually plays. One quality dial normalised into four different encoder scales.
You only need the audio from an interview video to send to a transcription service.
Input
interview.mp4, AAC inside · keep the original track
Output
interview.m4a, 32,371 bytes, the same samples the video carried, bit for bit
Copying the track moves the samples into an audio container without decoding them, so nothing at all is lost and it is instant. Re-encoding the same two seconds as a 320 kbps MP3 gives 83,739 bytes carrying exactly the same information plus a second generation of loss, and the page says so when the bitrate you picked is higher than the source can supply.
Pull the soundtrack out of a video. The default copies the existing track sample for sample, which is exact, instant and the smallest thing this page can produce; MP3, M4A, AAC, OGG, WAV and FLAC are there when you need one of them, with the cost of the choice stated before you run it.
Save the audio from music videos, concert recordings, or video-based music releases, copying the track untouched where the format allows it.
Extract the audio track from a video podcast or webinar recording for audio-only distribution.
Pull audio from interview or meeting recordings to feed into transcription tools that only accept audio input.
Take a WAV or FLAC into an editor when you need an uncompressed decode to work on, with the page saying plainly that this cannot undo the loss already in the video.
The soundtrack inside a video is already compressed, almost always as AAC or Opus. That one fact decides everything else on this page, and it is why the default here is to copy the track rather than re-encode it. Copying runs -vn -c:a copy: the samples are lifted out of the video container and written into an audio one without being decoded, so the result is exact, instant and smaller than any of the alternatives. Measured on a two second test clip, the copied AAC came to 32,371 bytes where the same audio re-encoded as a 320 kbps MP3 came to 83,739.
When you do want another format, the page says what it will cost before it runs. Choosing a bitrate higher than the source has does not recover anything: extracting 320 kbps MP3 from a 96 kbps Opus track produced 82,605 bytes carrying exactly the information that was in 24,000 bytes of Opus, plus a second generation of loss. The tool reads the source bitrate and says so when the setting you have chosen is doing that.
The uncompressed formats are sized from the file rather than from a rule of thumb. WAV is written as 16 bit PCM and its rate is the sample rate times the channel count times 16, which for the mono 44.1 kHz test clip is 705.6 kbps and not the 1411 kbps everyone quotes: 1411 is 16 bit STEREO at 44.1 kHz and applies to that alone. The prediction for the test clip was 176,444 bytes against a real 178,254, which is one percent. FLAC keeps the same samples losslessly in roughly 60 percent of the space, and that figure is given as an estimate rather than a promise because it depends on the material.
Seven outputs are offered: a straight copy, MP3, M4A, AAC, OGG Vorbis, WAV and FLAC. The bitrate control appears only for the four formats where it means anything, and where it does not the page says why: a copy has no bitrate because nothing is being encoded, and a lossless format's size follows the sample rate and the channel count rather than a setting. A video with no audio stream is refused with that reason rather than run and failed inside ffmpeg. Afterwards the result panel prints the real size against the prediction, the difference between them as a percentage, and the exact command that ran.
Keeping the original track, every time. It is the samples that are already in the video, moved into an audio container without being decoded, so nothing at all is lost. Every other option decodes and re-encodes.
Only if the source is at least that high. A video's soundtrack is typically 96 to 192 kbps, and encoding it at 320 makes a larger file carrying the same information plus a second generation of loss. The page reads the source bitrate and warns when your setting is doing that.
No. 1411 kbps is 16 bit stereo at 44.1 kHz. The real figure is the sample rate times the channel count times 16, so a mono 44.1 kHz source is 705.6 kbps, exactly half. This page uses your file's own numbers.
This tool extracts the full audio track. To get a specific section, trim the video first using the Video Trimmer, then extract audio from the trimmed clip.
Video is decoded and re-encoded on your own device. Large files never leave your machine.