WAV vs MP3: Quality, Size and When Each Wins
WAV and MP3 sit at opposite ends of a single trade-off: fidelity against size. WAV stores sound exactly. MP3 stores an approximation your ears are unlikely to catch, in a tenth of the space. Both are answers to different questions, and once you know which question you are asking, the choice makes itself.
How each one stores sound
A WAV file is essentially raw PCM. The microphone signal is measured thousands of times per second (44,100 for CD quality) and every measurement is written down verbatim. Nothing gets thrown away, so playback reconstructs the waveform exactly: lossless in the strong, bit-for-bit sense of the word.
MP3 is perceptual coding. Rather than storing the waveform, it stores a description of what you would hear, exploiting quirks of human hearing: a loud sound masks quieter sounds near it in pitch and time, and we are insensitive to the highest frequencies. The encoder discards whatever its model says you cannot perceive and spends its limited bits on what you can. If it helps, picture WAV as a photograph of the waveform and MP3 as a skilled artist’s rendering: remarkably faithful, but the discarded detail is gone for good.
The size difference, concretely
| Format | One minute, stereo 44.1 kHz | A 4-minute song |
|---|---|---|
| WAV (16-bit PCM) | ≈ 10 MB | ≈ 40 MB |
| MP3 at 320 kbps | ≈ 2.4 MB | ≈ 9.6 MB |
| MP3 at 128 kbps | ≈ 1 MB | ≈ 3.8 MB |
WAV’s size is fixed arithmetic, sample rate times bit depth times channels, regardless of content, so a minute of silence costs the same 10 MB as a minute of orchestra. MP3’s size is whatever bitrate you choose, which is exactly the dial that trades quality for space.
Can you actually hear the difference?
Decades of blind listening tests point the same way: at 320 kbps almost nobody can reliably tell MP3 from the lossless original. Not casual listeners, and mostly not audio engineers on studio monitors either. At 192 to 256 kbps the same holds for nearly all material.
Low bitrates are a different story. At 128 kbps and below, trained ears pick out swishy, underwater-sounding cymbals and smeared transients, and at 96 or 64 anyone can hear it. A few killer samples such as harpsichords, castanets and dense cymbal work expose encoders sooner than ordinary music does. If someone claims they can hear 320 kbps MP3 a mile away, the polite response is a blind ABX test.
When WAV wins
- Editing and production: every edit, effect or re-export of a lossy file re-encodes it, stacking losses generation on generation. Work in WAV and go lossy once, at the very end.
- Archival masters: a recording you can never make again, like an interview or a live take, deserves the format that discards nothing. At today’s storage prices there is no reason not to.
- Compatibility with pro tools: DAWs, video editors, broadcast systems and sampler hardware all treat WAV as the native currency.
When MP3 (or AAC) wins
Distribution, which is almost everything else. Podcasts, music libraries, voice notes, email attachments, anything streamed or downloaded: a file one-tenth the size that listeners cannot distinguish is simply the correct engineering choice. AAC (.m4a) is MP3’s successor, a little better at every bitrate and just as universally supported, so for new material it is the stronger pick. MP3 remains the format that literally everything built since 1998 can play.
The browser quirk: decode anything, encode WAV
A practical note if you use in-browser audio tools. Browsers ship licensed decoders for MP3, AAC and OGG, so the Web Audio API can read nearly any file, but they expose no built-in lossy encoders, which means writing a file back out produces uncompressed PCM. That is why our Audio to WAV Converter turns MP3, M4A or OGG into WAV on your own device instead of on a processing server, and why the Voice Recorder saves takes in lossless form. Handy precisely when the destination is a DAW, a sampler or a transcription tool that wants clean PCM input.
One conversion that can’t help you
Converting an MP3 to WAV does not restore quality. The detail the encoder discarded is permanently gone, and the WAV is a ten-times-larger container around the same approximation. The conversion is still often worthwhile for compatibility with equipment or software that requires WAV, but it is a container change rather than an upgrade.
Which leaves a simple working rule: record and edit lossless, distribute lossy, and treat the trip from lossy back to lossless as something that never actually happens.