WAV to MP3 Compressor
WAV stores audio as raw uncompressed samples, which is why a three-minute recording runs to about 30 MB. That is the right format while you are editing and the wrong one for distribution. Compression here runs in the browser.
How to use it
- Drop in a .wav file.
- FFmpeg loads as WebAssembly, then encodes the samples with LAME.
- The .mp3 downloads when encoding completes.
What you are actually giving up
A standard CD-quality WAV holds 16-bit samples at 44,100 samples per second across two channels, which works out to roughly 10 MB per minute. Nothing is compressed, so every sample is stored exactly as captured.
MP3 applies a psychoacoustic model: it identifies sounds that are masked by louder sounds nearby in time or frequency, and discards them on the basis that you cannot hear them anyway. At around 128 kbps that yields roughly a ten to one reduction, so a 30 MB WAV becomes about 3 MB.
For most listeners, on most equipment, with most material, the difference is not perceptible. It becomes audible on content with sharp transients such as cymbals and applause, and on sustained pure tones, where the encoder has less masking to exploit.
When to keep the WAV
Keep the uncompressed file whenever the audio is still going to be worked on. Every subsequent edit and re-export of an MP3 re-encodes it, and lossy generations compound in the same way they do for images. Master in WAV, export to MP3 once at the end.
Keep it also for archival, for anything heading to a mastering engineer or a broadcaster, and for short samples that will be layered and processed. The storage cost of a WAV is trivial compared with the cost of discovering later that the only surviving copy is a 128 kbps export.
Distribution is the case where MP3 wins outright: podcast feeds, email attachments, embedded web audio, and anything going to a device with limited storage.
Sample rate and channel notes
The output preserves the source sample rate and channel layout. A 48 kHz stereo WAV produces a 48 kHz stereo MP3, and a mono voice recording stays mono, which halves the size again relative to encoding it as stereo.
One thing to be aware of: 24-bit and 32-bit float WAVs, common in professional recording, are reduced to the 16-bit resolution the MP3 encoder works with. The extra headroom those formats provide matters during mixing and is irrelevant once the mix is finished, so this is only a concern if you are compressing a file that is not yet final.
At a glance
| Accepted input | .wav |
|---|---|
| Output | .mp3, LAME encoder |
| Typical reduction | Around ten to one |
| Engine | FFmpeg compiled to WebAssembly, single-threaded |
Frequently asked questions
How much smaller will the file be?
Roughly a tenth of the original. A 30 MB WAV typically lands near 3 MB as an MP3 at default quality.
Will I hear the difference?
Usually not on speech or typical music through typical equipment. Sharp transients such as cymbals and applause are where lossy encoding is most audible.
Should I compress before or after editing?
After. Edit in WAV and export to MP3 once, at the end. Editing an MP3 and re-exporting stacks a second generation of loss on top of the first.
Is my mono recording encoded as stereo?
No. The channel layout of the source is preserved, so a mono file stays mono and stays correspondingly smaller.
Read more
Audio and video: containers, codecs, and transcoding — An MP4 is not a format, it is a box. Knowing what is in the box explains most media conversion behaviour.