MP4 or MP3: Choosing Between Video and Audio
The practical difference between a video container and an audio file, what you lose by extracting audio, and how to pick a bitrate that is worth the storage.
The choice between MP4 and MP3 looks like a choice between two file formats. It is really a choice about what you intend to do with the file, and it is worth a moment's thought because one direction is irreversible.
MP4 is a container, MP3 is not
This distinction explains most of the confusion around media formats. An MP4 file is a container: a wrapper holding one or more streams, typically a video stream, an audio stream, and often subtitles and chapter markers. The container specifies how those streams are stored alongside each other, not how each one is compressed.
MP3, by contrast, describes both a container and a specific audio compression method. An MP3 file holds exactly one thing: a single compressed audio stream. There is nowhere in the format to put a picture track.
This is why converting MP4 to MP3 is a one-way door. You are not changing a wrapper; you are discarding the video stream and keeping the audio. The file gets much smaller because you have thrown away the largest part of it, and nothing can bring it back.
Audio bitrates in practice
What each bitrate is genuinely suited to:
| Bitrate | Approximate size per hour | Suited to |
|---|---|---|
| 64 kbps | ~29 MB | Speech only; noticeably degraded on music |
| 128 kbps | ~58 MB | The long-standing default; fine for casual listening |
| 192 kbps | ~86 MB | Music on decent headphones; a sensible middle ground |
| 256 kbps | ~115 MB | Close to transparent for most listeners |
| 320 kbps | ~144 MB | The MP3 maximum; the practical ceiling of the format |
When each choice makes sense
- Keep the video. Anything where the picture carries meaning: tutorials, demonstrations, anything with text on screen, sport, film. Also keep it whenever you are unsure, since you can always extract the audio later.
- Take the audio only. Interviews, podcasts, lectures, music and language practice, where the picture is either static or incidental. The storage saving is large and nothing you care about is lost.
A note on re-encoding
Every lossy re-encode degrades quality, because each pass discards information based on what the previous pass left behind. Converting a 128 kbps source into a 320 kbps file does not recover detail; it produces a larger file containing the same damaged audio, plus whatever the second encoder introduced.
The practical rule: choose the bitrate you actually want at the point of extraction, and avoid repeatedly converting the same file between formats.