WAV files are big because they aren't compressed at all: every sample is stored as a plain number. A minute of a typical recorder file (24-bit, stereo, 48 kHz) takes over 17 MB, so a few minutes of an interview won't go through email, and a long recording fills a phone. The good news is that some of the biggest savings cost no quality at all.
We tested every method on 8 October 2026 on Windows 11 with Chrome, using the same 1 minute 42 second recording of speech, saved the way field recorders and lavalier kits often save it: 24-bit, two channels, 48 kHz, with a couple of seconds of room noise at each end. It started at 29.35 MB.
The short answer
| Method | Quality | Our file | Minutes in a 25 MB email |
|---|---|---|---|
| Nothing (the original WAV) | — | 29.35 MB | 1.4 |
| ZIP it | Unchanged | 19.37 MB (66%) | 2.2 |
| Mono, when both channels are the same | Unchanged | 14.68 MB (50%) | 2.9 |
| … plus 16-bit instead of 24-bit | No audible change | 9.78 MB (33%) | 4.3 |
| … plus trimming the silence at each end | Unchanged | 9.40 MB (32%) | 4.5 |
| FLAC instead of WAV | Unchanged (lossless) | 2.24 MB (8%) | about 10 |
| MP3, 128 kbps mono | Lossy, very good for speech | 1.63 MB (6%) | 26 |
| MP3, 64 kbps mono | Lossy, fine for speech | 816 KB (3%) | 52 |
If the file just has to be smaller and nobody will edit it again, convert it to MP3: that's the biggest cut by far. If quality must stay exactly the same, go to mono (when it's safe) and FLAC. The sections below show how, and when each step is a bad idea.
1. Check whether your "stereo" file is really mono
Many recordings have two channels with the same sound in both: phone recorders, video-call exports, interview kits with a single microphone. Half of that WAV is a copy, and throwing the copy away loses nothing. In our test it halved the file, from 29.35 MB to 14.68 MB.
The tricky part is knowing whether the channels really are identical. WAV to MP3 checks them sample by sample before converting and tells you ("Both channels are identical"), and if one side is silent, as happens when a microphone is plugged into one input, it takes the side with the voice instead of mixing it with silence, which would make it 6.5 dB quieter. If the two sides differ, as in music or a stereo field recording, keep both.
2. Cut the silence and the parts you don't need
Every second of a WAV costs the same, silent or not. Trimming two seconds from each end of our recording saved only 4%, but on real recordings the savings are usually in the long pauses before someone starts talking, the minutes after the interview ended, or a false start. Cutting a WAV loses no quality: Trim Audio cuts WAV files at the exact sample you choose, without re-encoding, and shows you the waveform so you can see the silence.
3. Use 16-bit instead of 24-bit for the final copy
Recorders use 24-bit so there's room to fix levels afterwards. For listening, 16-bit (CD quality) is more than enough, and it takes a third less space: our mono file went from 14.68 MB to 9.78 MB. Keep the 24-bit original if you'll edit it again. In Audacity, File > Export Audio lets you choose WAV and, under encoding, 16-bit. Lowering the sample rate from 48 kHz to 44.1 kHz saves only 8% and does change the audio, so it's rarely worth it.
4. Save it as FLAC if quality must not change
FLAC is to WAV what ZIP is to a document, except it's built for sound, so it works far better. It's lossless: decode a FLAC and you get back exactly the WAV you started with. Our recording went from 9.40 MB as a trimmed 16-bit mono WAV to 2.24 MB as FLAC. Speech with pauses compresses especially well; music usually shrinks less. Audacity exports FLAC from the same Export Audio window.
The catch is compatibility: many upload forms and some older players don't take FLAC. If the other side asked for a WAV or an MP3, FLAC won't help.
5. Convert it to MP3 when it just needs to be small
MP3 throws away detail you're unlikely to hear, which is why it gets so much smaller: our recording became 1.63 MB at 128 kbps mono, 18 times smaller than the original, and 816 KB at 64 kbps, which is still perfectly clear for speech. For music, use 256 kbps stereo (about 1.9 MB per minute).
WAV to MP3 does this in the browser, with nothing uploaded. It also catches the two things that go wrong most often: 32-bit float recordings whose peaks go above full scale, which an MP3 would clip with an audible crackle (it lowers the volume just enough instead), and stereo files that should have been mono. Keep the WAV as your master if you might edit the recording later, because every round of MP3 encoding loses a little more.
What doesn't work
- Zipping the WAV. ZIP looks for repeated patterns, and sound has few of them. Our file only went to 66% of its size, and a plain 16-bit mono WAV only to 79%. FLAC got to 8% from the same recording.
- Renaming it to .mp3. The file stays a WAV with the wrong name. Some players refuse it, and the size doesn't change. It has to be converted.
- Converting to MP3 at 320 kbps "to be safe". For a voice recording it's five times the size of 64 kbps mono with nothing extra to hear.
Which one should you use?
- To email or upload a recording of speech: MP3, 64 or 128 kbps mono.
- To send music or a mix for listening: MP3 at 256 kbps stereo.
- To archive or hand over for editing, with nothing lost: mono if the channels are identical, trimmed, then FLAC, or 24-bit WAV if the recipient needs WAV.
- When a form says "WAV only" and the file is too big: mono (if identical), trim, and 16-bit. Together they took our file to 32% of its size while staying a WAV.