An MP4 is a container: it holds a video track and, almost always, a sound track compressed with AAC. MP3 is a different, older audio codec, so the AAC sound has to be decoded and encoded again. Both formats are lossy, which means the conversion is lossy to lossy: the MP3 cannot be better than the AAC it came from, and a low bitrate adds a second layer of artefacts. At 192 kbps that second layer is inaudible to most listeners, which is why it is the default.
The encoder is LAME, the long-standing open-source MP3 encoder, compiled to WebAssembly and loaded only when you convert. In the rare MP4 whose sound is already MP3, filetidy copies it as is instead of encoding it twice. The video track is simply left out, so the result is a pure audio file.
MP3 size is easy to predict: bitrate in kilobits per second, times the length in seconds, divided by eight, gives kilobytes. A four-minute song at 192 kbps is 192 × 240 ÷ 8 = 5,760 KB, about 5.8 MB; at 320 kbps the same song is about 9.6 MB.
The MP3 keeps the source's sample rate and is stereo at most. Title, artist and cover art are not written: MP4 metadata is dropped rather than translated into ID3 tags, so add them afterwards in your music app or tag editor if you need them.
- Typical MP4 audio
- AAC, re-encoded to MP3 (lossy to lossy)
- Bitrates offered
- 128, 192 (default) or 320 kbps, constant
- Size formula
- kbps × seconds ÷ 8 = kilobytes (192 kbps ≈ 1.44 MB per minute)
- Encoder
- LAME MP3 encoder in WebAssembly, in a background worker
- Channels
- Stereo at most; surround sound is downmixed to two channels
- Tags
- Not copied; no ID3 title, artist or cover in the output