WAV vs MP3: what every music producer needs to know
The technical difference between these two formats is small to describe and impossible to undo. This is where it matters in a real workflow, and where it genuinely does not.
Contents
If you produce music you have made the WAV-or-MP3 decision hundreds of times, usually without stopping. The decision is worth understanding once, because one direction of it cannot be reversed.
What actually makes WAV and MP3 different
WAV (Waveform Audio File Format) is a container for raw, uncompressed PCM audio. Every sample captured by the converter is stored exactly as it was recorded. Nothing is modelled, estimated or thrown away.
MP3 (MPEG-1 Audio Layer III) is lossy. A psychoacoustic model decides which parts of the signal a listener is least likely to notice, and those parts are removed from the file permanently. What comes back on decode is a reconstruction, not the original.
The size difference follows directly from that, and you can check it yourself rather than taking a number on trust:
- Uncompressed stereo at 44.1 kHz and 16 bits: 44,100 × 16 × 2 × 60 ÷ 8 ÷ 1,000,000 ≈ 10 MB per minute.
- MP3 at 128 kbps: 128,000 × 60 ÷ 8 ÷ 1,000,000 ≈ 0.96 MB per minute.
- MP3 at 320 kbps: 320,000 × 60 ÷ 8 ÷ 1,000,000 ≈ 2.4 MB per minute.
The bitrate is the whole story for an MP3’s size, which is why the same track at the same bitrate is the same size regardless of what is in it.
When WAV is the practical choice
Use WAV
- Recording — encoding at capture throws away data before you have made a single decision.
- Mixing and editing — processing amplifies what compression has already altered.
- Mastering delivery — uncompressed masters are the near-universal expectation, and many mastering engineers publish it as a requirement.
- Stems and collaboration — the receiving engineer should inherit your source, not your export settings.
- Broadcast and sync — deliverable specs for TV, film and advertising are written around uncompressed audio.
MP3 is fine
- Sending a rough mix for feedback.
- Personal listening, where the difference at high bitrates is genuinely hard to hear.
- Streaming submission, where the platform re-encodes from your master anyway.
- Short social clips, which are re-compressed by the platform regardless.
What happens when you import an MP3 into a DAW
Digital audio workstations do not process compressed audio directly. An MP3 has to be decoded back to PCM before any gain stage, EQ or effect can touch it — that is a property of how digital signal processing works, not a quirk of any one product. Some hosts decode to a cache file on disk, others decode into memory; the detail varies by product and version, and it does not change the outcome.
The outcome is this: your session is working with PCM either way, but the PCM it is working with was reconstructed from a file that had already lost information. Exporting that session as WAV produces an uncompressed copy of already-lossy audio. The container is lossless. The content is not.
Generation loss
Each MP3 encode discards more. Bounce to MP3, reimport, process, export to MP3 again, and the losses stack. This is generation loss, and it is the practical reason to keep a lossless working master even when everything you ship is compressed.
The bottom line
Use WAV for anything that will be edited, processed or delivered professionally. Use MP3 for sharing, previewing and listening. When you are unsure, keep the lossless file — storage is cheap and compression is always available later, while the reverse is not.
If you already have MP3 files that a DAW, mastering chain or distributor requires in WAV form, converting them takes seconds. Be clear about what that buys you: compatibility with a professional workflow, not restored quality. On that point the site is consistent everywhere, and the archival argument explains why it matters even for files you have already compressed.
Sources & method
Format behaviour below is derived from the published definitions of PCM and MPEG-1 Audio Layer III rather than from measurements of our own. File sizes are shown as arithmetic you can repeat, not as benchmarks. Where a claim depends on a specific product’s current internal behaviour, we describe the general, documented principle instead.
- FLAC — Free Lossless Audio Codec, Xiph.Org Foundation. xiph.org Accessed 2026-08-29
- mp3 — licensing program terminated, Fraunhofer IIS. iis.fraunhofer.de Accessed 2026-08-29