16 March 2026 | 6 min read
How to extract audio from MP4 without losing quality
The audio inside your MP4 is already compressed. The only question is whether you make it worse on the way out.
Extracting audio from a video file is one of the most searched audio tasks there is, and most of the tools that answer the search do the same thing: upload your footage to a stranger's server and hand back an MP3. For anyone working with client material, that is not acceptable, and the MP3 is not necessary either.
What is inside an MP4
An MP4 is a container. Inside it, video and audio sit as separate streams. The audio is usually AAC, sometimes PCM on files from professional cameras. Extraction means taking that stream out, and what happens next determines the quality outcome.
Three outcomes, only two of them good
- Stream copy: the audio stream is lifted out untouched into a new container. Perfectly lossless, but the result stays AAC.
- Decode to PCM: the AAC is decoded once into uncompressed WAV or AIFF. No further generation loss, and the result is editable everywhere.
- Transcode to MP3: the audio is decoded and then compressed again with a different lossy codec. This is the one that costs you quality for no reason.
For editing, mixing or transcription work, decoding to WAV is the right answer. You accept the compression that already happened when the camera recorded, and you add none of your own.
Choose the sample rate and bit depth deliberately
Cameras commonly record 48 kHz AAC. Decoding to 48 kHz 24 bit WAV matches the source and matches film delivery specs. Upsampling to 96 kHz adds nothing but file size, and downsampling to 44.1 kHz for a video project just creates a mismatch your NLE will have to fix later.
Do it locally
Uploading rushes to an anonymous converter is a confidentiality problem, a bandwidth problem and a waiting problem. Modern browsers can decode audio streams from MP4, MOV and WEBM with the Web Audio API and WebAssembly, on your own machine.
The audio extractor here does exactly that. Drop the video in, choose WAV or AIFF, choose the sample rate and bit depth, download. Nothing is uploaded, there is no queue, and a two hour interview finishes as fast as your laptop can decode it.
Common cases
- Interview shot on a mirrorless camera: extract to 48 kHz 24 bit WAV, then sync with the recorder audio.
- Screen recording for a tutorial: extract to WAV, clean it, place it back on the timeline.
- Reference audio for transcription: WAV is more reliable input for speech to text engines than a re-encoded MP3.