MP3 vs WAV: which one should you actually use?
One is a working format and one is a delivery format. Confusing the two is how people end up emailing 400 MB of audio, or handing a mastering engineer a file that has already been squeezed.
The short version. Use WAV while you are recording, editing, or archiving, because it keeps every sample and never degrades when you re-save it. Use MP3 when the file has to travel: it is roughly seven times smaller at 192 kbps and plays on everything. Converting between them cannot add quality back, only take it away.
The comparison, in one table
| WAV | MP3 | |
|---|---|---|
| Compression | None. Raw PCM samples | Lossy. Audio is permanently discarded |
| Quality after conversion | Unchanged, every time | Degrades a little with each re-encode |
| Size, 4 minute stereo track | About 40 MB | 3.7 MB at 128 kbps, 9.2 MB at 320 kbps |
| Introduced | 1991, by Microsoft and IBM | 1993, as MPEG-1 Audio Layer III |
| Editing | Ideal. No loss on repeated saves | Poor. Each save re-compresses |
| Tags and artwork | Limited and inconsistently supported | ID3 tags, supported everywhere |
| Streaming | Impractical at that size | What it was designed for |
| Typical use | DAWs, CD masters, samples, broadcast | Phones, cars, downloads, podcasts |
What a WAV file actually is
WAV is a container, defined by Microsoft and IBM in 1991, that almost always holds uncompressed PCM audio. Inside, the file is a header describing the format followed by a long list of numbers, one per sample per channel. Nothing clever happens. That simplicity is the reason a WAV recorded in 1995 still opens today in any editor you can name.
Because there is no compression, size is entirely predictable. Multiply sample rate by channels by bytes per sample and you have the bytes per second. A stereo 44.1 kHz 16-bit file is 10.1 MB per minute whether it contains a full orchestra or silence.
The other consequence: re-saving a WAV twenty times does nothing to it. Every save writes the same numbers back. That is why editing work stays in WAV until the final export.
What an MP3 file actually is
MP3 arrived in 1993 as the audio layer of the MPEG-1 standard, and it works by modelling what human hearing misses. A loud sound hides a quiet one just after it. Frequencies above roughly 16 kHz are inaudible to most adults. The encoder finds audio it predicts you will not notice, removes it, and stores what is left far more compactly.
It works remarkably well, which is why the format outlived every attempt to replace it. But the removal is permanent, and it happens again every time you re-encode. Editing an MP3, saving it, and editing it again stacks up damage that eventually becomes obvious even to casual listeners.
Can you hear the difference?
Honestly, it depends on the bitrate, the material, and the listener, and the answer is less dramatic than either camp claims.
At 128 kbps, differences are usually audible on the right material. Cymbals and hi-hats develop a thin, watery hiss. Applause turns to static. Reverb tails lose their edges. On a podcast or a pop mix played in a car, most people would not notice anything wrong.
At 320 kbps, reliably identifying the MP3 in a blind comparison is genuinely difficult, and plenty of people who are certain they can hear it discover otherwise when the labels are hidden. That is worth knowing before you spend hours re-ripping a library.
The part that is not a matter of opinion: the discarded data is gone. Whether you can hear it today matters less if the file will be re-encoded, pitch shifted, time stretched, or heavily processed later, because those operations expose damage that normal playback hides.
Which one to use, by situation
| What you are doing | Use | Why |
|---|---|---|
| Recording or mixing in a DAW | WAV | No generation loss, and some DAWs will not import MP3 at all |
| Burning an audio CD | WAV, 44.1 kHz 16-bit | Exactly what the Red Book CD standard requires |
| DJing with Serato, Traktor, or Rekordbox | WAV | Faster analysis and no decode overhead mid set |
| Emailing a mix for feedback | MP3, 192 to 320 kbps | 40 MB will bounce. 5 MB will not |
| Publishing a podcast | MP3, 96 to 128 kbps | Speech compresses well and listeners download faster |
| Filling a phone or car stereo | MP3 | Storage runs out long before quality does |
| Long term archiving | WAV or FLAC | Keep the version you can still make decisions from |
| Sound effects in a game engine | WAV | Predictable decode cost and sample accurate playback |
A workflow that avoids the usual mistakes
Keep one WAV master per project and treat every MP3 as disposable. When you need a new MP3, make it from the WAV again rather than from an older MP3. This costs nothing and avoids the single most common way people quietly ruin their own audio.
If storage is the problem, FLAC is the compromise worth knowing about. It is lossless like WAV, usually a little over half the size, and it handles tags properly. The catch is that not every piece of software reads it, which is exactly the gap WAV still fills.
Converting between them
Both directions work in your browser here, without uploading anything:
- MP3 to WAVFor getting audio into a DAW, an editor, or onto a CD.
- WAV to MP3For sending, uploading, or fitting more onto a device.
Common questions
Is WAV better than MP3?
WAV is more accurate, which is not the same as better for your situation. WAV keeps every sample the recording captured. MP3 throws some away to save space. If you are editing, mixing, mastering, or archiving, use WAV. If you are sending, streaming, or filling a phone, MP3 is the sensible choice and a 320 kbps file is hard to fault.
Can you actually hear the difference between MP3 and WAV?
At 128 kbps, often yes: cymbals get a papery hiss and reverb tails smear. At 320 kbps most listeners cannot reliably tell in a blind test on normal equipment. The difference is easiest to hear on cymbals, applause, and solo piano, and hardest on dense loud mixes. Note that "cannot hear it" and "the data is still there" are different claims. Only the second one is about the file.
Does converting MP3 to WAV restore the lost quality?
No. Conversion changes the container, not the content. You get a larger file that sounds the same. Here is the longer explanation.
Which format should I use for a podcast?
Record and edit in WAV, publish in MP3. Speech compresses extremely well, so 96 to 128 kbps mono is usually plenty for the published episode, and your listeners will thank you for the smaller download.
Do WAV files hold artist and album tags?
Badly. MP3 has ID3 tags that every player understands. WAV can carry metadata in optional INFO or iXML chunks, but support is patchy and many programs drop it when they re-save the file. If your library depends on tags, MP3 or FLAC will treat you better.
What about FLAC, AAC, and ALAC?
FLAC and ALAC are lossless like WAV but roughly half the size, and they handle tags properly. AAC is lossy like MP3 and generally sounds better at the same bitrate. WAV survives because it is simple and every piece of audio software on earth reads it.