How to Extract the Audio from a Video (Without Re-encoding It Twice)
Pulling audio out of a video can be a lossless copy or a re-encode. Which you get depends on the format you ask for, and it decides the quality you end up with.
Updated August 11, 20265 min read
A video file is a container holding separate streams: usually one video, one audio, sometimes subtitles. Extracting the audio means taking one stream out and writing it to its own file. Done well that is a copy — instant, and bit-for-bit identical to what was in the video. Done carelessly it is a re-encode, which takes longer and loses quality for no benefit. The difference is entirely down to which output format you ask for.
Copy the stream, do not convert it
Almost every MP4 you encounter contains AAC audio. If you extract to M4A — which is the same AAC data in an audio-only container — nothing has to be decoded or re-encoded. The bytes are copied across. It takes a second regardless of length and the result is exactly as good as the original.
Ask for MP3 instead and the AAC has to be decoded to raw audio and then re-encoded as MP3. That is a generational loss: you are compressing something that was already compressed, with a different algorithm making different assumptions. It is usually not dramatic, but it is entirely avoidable.
| Video contains | Ask for this | Result |
|---|---|---|
| AAC (most MP4, MOV) | M4A | Copy — instant, lossless |
| AAC | MP3 | Re-encode — slower, slight loss |
| Opus (many WebM) | OGG | Copy — instant, lossless |
| Opus | MP3 | Re-encode — slower, slight loss |
| Anything | WAV | Decode to uncompressed — huge file, no further loss |
When MP3 is still the right answer
MP3 is thirty years old and less efficient than AAC or Opus, but it remains the format that everything plays without argument: car stereos, gym equipment, DJ software, old MP3 players, embedded devices, podcast tooling. If the destination is any of those, take the small re-encode and be done.
One rule if you do: encode at a decent bitrate. Going from AAC to a 96kbps MP3 stacks two lossy passes at low quality and it will be audible. 192kbps or higher keeps the second pass essentially transparent (192kbps is the free maximum in our extractor; Pro goes up to 320kbps). There is more on the trade-offs in audio formats explained.
The workflow
Trim the video first if you only need part of it
Extracting from a 40-minute recording to keep 90 seconds wastes time and disk. Cut it with the video cutter first — trimming video is faster than trimming audio afterwards.
Extract to the matching format
In extract audio, choose M4A for MP4 and MOV sources. You will notice how fast it is when no re-encoding is involved.
Convert only if something demands it
Use the audio converter as a separate, deliberate step rather than folding a format change into the extraction.
Fix the level if it needs it
Video audio is frequently quiet or wildly inconsistent. The normalizer evens it out — useful for lecture recordings and interviews in particular.
Runs on your own device, not our servers. Free takes one video up to 300 MB and extracts up to 60 minutes of audio from it; Pro lifts both caps and takes up to 20 videos at once. AVI, FLV, WMV and 3GP files are capped at 150 MB on every plan.
Extract audio from a videoCommon uses and what each one needs
- Lecture or meeting recordings. Extract to M4A, normalise, and you have something a fraction of the video's size that you can listen to while walking. Video adds nothing to a talking head.
- Podcast from a video interview. Extract losslessly, then edit as audio. Podcast platforms want MP3 or M4A, so convert at the final export rather than at extraction.
- Music from your own footage. Extract, then trim to the section you want.
- Transcription. Extract first — audio-only files upload and process far faster in transcription tools, and the video was never being used.
- Removing audio instead. If the goal is a silent video rather than a separate audio file, that is the mute tool, not this one.
Things that go wrong
Silent output. Either the video genuinely has no audio track, or it has several — a film with multiple language tracks, or a screen recording with separate system and microphone streams — and the extraction took one you were not expecting.
Audio drifts out of sync. This shows up when the source has a variable frame rate, which many screen recordings and phone videos do. The audio itself is fine; it is the video timing that was elastic. Extracting the audio on its own is unaffected, but re-joining it to a re-encoded video later can drift.
The file is much bigger than expected. You extracted to WAV, which is uncompressed — roughly 10MB per minute in stereo. That is the right choice if you are about to edit and master, and the wrong one if you just wanted something to listen to.
A closing note on copyright, since it comes up: extracting audio from your own recordings is unremarkable, and extracting from material you do not own is governed by the same rules as any other copy. Whether a personal copy is allowed varies by country — the UK, for one, has no private-copying exception — and redistribution without the rights holder's permission is not allowed anywhere.
Frequently asked questions
How do I convert an MP4 to MP3?
Extract the audio stream and choose MP3 as the output. Be aware that MP4 audio is almost always AAC, so asking for MP3 forces a re-encode. Extracting to M4A instead is an instant, lossless copy of the same audio — convert to MP3 only if the destination requires it.
Does extracting audio from a video lose quality?
Not if the output format matches what is already inside the container — AAC to M4A, or Opus to OGG, is a straight copy. Asking for a different format means decoding and re-encoding, which costs a little quality and a lot more time.
What format should I extract audio to?
M4A from MP4 and MOV files, because it is a lossless copy and plays almost everywhere. Use MP3 when the destination is older hardware or software that insists on it, and encode at 192kbps or above so the second lossy pass stays inaudible (192kbps is the free maximum in our extractor; Pro goes up to 320kbps).
Why is my extracted audio silent?
Either the video has no audio track, or it has more than one and the extraction picked a different one — common with multi-language films and with screen recordings that keep system audio and microphone as separate streams.