WAV to MP3

How much smaller

Stereo WAV at CD quality runs roughly 10 MB per minute. MP3 at 192 kbps is about 1.4 MB per minute, around a seventh of the size, and for most people on most equipment it is indistinguishable from the original.

For voice recordings, 96 kbps mono halves it again with no meaningful loss. Set channels to mono on the audio converter if the source is a single microphone.

This is the one conversion you cannot undo

WAV holds every sample the recorder captured. MP3 decides which of them you were never going to hear and discards those permanently. The result is excellent and the decision is final: converting the MP3 back to WAV later gives you a large file containing exactly what the MP3 contained, not the recording you started with.

So the rule is simple. If the WAV is a master, something you recorded, something you will edit again, or the only copy, keep it. Convert a copy. Disk is cheaper than a re-record, and the moment you discover otherwise is always after the original is gone.

Where the size goes

Uncompressed stereo audio at CD quality is 44,100 samples a second, two channels, 16 bits each, which works out to about 1.4 megabits per second, or roughly 10 MB per minute. That number does not care what the recording is: an hour of silence and an hour of an orchestra are the same size, because every sample is stored regardless.

MP3 at 192 kbps is about 1.4 MB per minute, a seventh of the size. At 128 kbps it is under a megabyte a minute. At 96 kbps mono, which is the right setting for a single-microphone voice recording, an hour of speech is around 40 MB rather than 600.

Choosing the bitrate honestly

• 192 kbps for music. This is where most people stop hearing a difference on ordinary equipment, and it is the sensible default.
• 256 or 320 kbps if the source is genuinely lossless and the destination is good headphones in a quiet room. This is the one case where the extra bits are doing real work, because there are no earlier artefacts to preserve.
• 128 kbps for speech in stereo, or anything going onto a device where space is tight.
• 96 kbps mono for interviews, lectures, dictation and voice memos. Mono halves the data before the encoder even starts, and a single microphone recorded two identical channels anyway.

If the WAV is going into an editor rather than onto a player, consider FLAC instead, on the audio converter. It is lossless, so nothing is discarded, and it is still around half the size of the WAV.

Common questions

Further reading

Related tools