MP3 vs WAV

A WAV is the audio itself: a header saying how many samples per second and how many bits each one has, then the samples, unaltered. Nothing is encoded and nothing is interpreted — it is as close to a direct recording of the waveform as a file gets.

An MP3 is a description of that waveform, compressed by removing what a psychoacoustic model expects nobody to notice. It is roughly a tenth of the size and it is not the same audio. Both are useful and they are useful at different ends of the process.

CD-quality stereo is 44,100 samples a second, 16 bits each, two channels — about 10 megabytes a minute, so an hour of recording is around 600 megabytes. Studio settings of 48 kHz at 24 bits push that to roughly 900 megabytes an hour, and 96 kHz doubles it again.

The same hour as a 192 kbps MP3 is about 85 megabytes. That ratio is why nobody distributes WAV and why a podcast producer with a folder of raw interviews thinks about disk space constantly. It is also why WAV files are so awkward to email — the format is not big by accident, it is big because it is uncompressed.

Editing. Cutting, fading, normalising, applying a filter and saving again are all operations on the samples, and with a WAV each one starts from exactly what was there before. An MP3 has to be decoded, altered and re-encoded on every save, and each round discards a little more — after a few passes of ordinary editing the damage is audible.

The second reason is precision. An MP3 cannot be cut at an arbitrary point because it is organised in frames, so an edit lands on the nearest boundary and a little silence is added at the start of every encoded file. Neither matters when listening; both matter when several tracks have to line up.

WAV uses 32-bit fields for its chunk sizes, so the format cannot describe a file larger than about four gigabytes. At 48 kHz and 24 bits in stereo, that is roughly six and a half hours — comfortable for a song and a real risk for a conference, a concert or a multitrack session.

Recorders handle it by splitting into several files, which is why a long recording arrives in numbered parts. RF64 and BWF extend the format past the limit and are supported by professional tools and not by everything. If a long session matters, check what the recorder does before the event rather than after.

MP3 carries ID3 tags: artist, title, album, track number, cover art, and whatever else. Every player reads them, which is why a music library sorts itself.

WAV has an old and barely used tagging chunk that many programs ignore or discard on save. In practice a folder of WAVs is a folder of filenames, and the naming convention is the metadata. For finished audio going into a library, that alone is a reason to convert.

A WAV is described entirely by its sample rate and its bit depth. Compact disc audio is 44,100 samples a second at 16 bits; video production works at 48,000; studios often record at 24 bits to leave room for adjusting levels afterwards without the noise floor rising into audibility.

Neither number has an equivalent on the other side of this comparison, because MP3 replaced them with a bitrate that describes the compressed result rather than the samples. That is the clearest way to see what the two formats are: one states how the sound was measured, the other states how much room the description was given.

WAV to MP3 is the normal last step of any audio project: choose a high bitrate, 192 for speech and 256 or 320 for music, and the result is a tenth of the size and hard to distinguish. This is where the compression should happen — once, at the end, from the master.

MP3 to WAV makes the file ten times larger and does not restore anything; the audio inside is the same degraded signal in an uncompressed wrapper. There is exactly one good reason to do it: software that will only accept WAV — some editing tools, some broadcast and telephony systems, some game engines and CD burners. As an upgrade it is an illusion.

Record, edit and master in WAV. Keep the WAV master. Deliver MP3, or AAC or Opus if the destination supports them — both are better than MP3 at the same bitrate, and MP3’s advantage is only that everything ever made will play it.

For long-term storage of finished audio, FLAC is the better answer than either: it is bit-for-bit identical to the WAV, about half the size, and it holds proper tags. Keep WAV for the working session and FLAC for the archive, and let MP3 be the copy that travels.