MP3, WAV, FLAC: Which Audio Format Should You Export To?

The two families of formats

Every audio format falls into one of two categories, and that distinction governs everything else.

Lossless formats (WAV, AIFF, FLAC, ALAC) preserve all audio information. A decompressed FLAC file returns exactly the same samples as the original WAV, bit for bit. FLAC compression works like a ZIP archive: it reduces size without discarding anything.

Lossy formats (MP3, AAC, OGG, Opus) permanently remove part of the signal. The algorithm relies on psychoacoustic models: it identifies what the human ear will not perceive — a sound masked by a louder one at a nearby frequency, frequencies beyond the hearing threshold — and eliminates it. The discarded data cannot be recovered.

Comparison

Understanding bitrate

Bitrate expresses the amount of data per second of audio. At 320 kbps, each second uses 320 kilobits. The higher the bitrate, the less information the algorithm must discard.

The progression is not linear in terms of perceived quality. The gap between 128 and 192 kbps is clearly audible on music: cymbals turn "metallic" and the highs lose finesse. The gap between 256 and 320 kbps, by contrast, is very hard to perceive even in good listening conditions. Beyond 320 kbps, MP3 offers nothing useful.

Two encoding modes exist. Constant bitrate (CBR) uses the same rate from start to finish, including silent passages where it is wasted. Variable bitrate (VBR) adapts the rate to signal complexity, giving a better quality-to-size ratio. For music, high-quality VBR is generally preferable to CBR 320.

The problem with repeated encoding

This is the most important point in this article, and the most frequently ignored.

Every lossy encode permanently removes information. If you start from an MP3, process it, then export to MP3, you apply a second compression to an already depleted signal. The losses accumulate. After three or four cycles, artefacts become clearly audible: whistling in the highs, a "swimming" effect on sustained sounds, loss of transient precision.

The problem worsens with audio processing. Slowing a track and adding reverb changes its spectral content. The second MP3 encoder then encounters a signal different from the one the first encode's decisions were optimised for, and it discards different information. The flaws compound rather than overlap.

Practical consequence: if you plan to rework a file — cut it into a video, apply a second treatment, remix it — export to WAV. You will convert to MP3 only once, at the very end of the chain.

What to choose by destination

Publishing to YouTube, TikTok, Instagram

These platforms systematically re-encode whatever audio you upload, regardless of source format. Their encoder will do better work starting from a high-quality file. Upload WAV if the platform and your connection allow, otherwise MP3 320 kbps. Uploading a 128 kbps MP3 guarantees a degraded final result, since the platform compresses on top of it.

Personal listening on a phone

MP3 320 kbps is the best compromise. The difference from a lossless file is inaudible in a mobile listening context — transit, Bluetooth earbuds, ambient noise — and the space saving is substantial (roughly four times smaller).

Video editing

WAV, without hesitation. Your editing software will re-encode audio at final export. Starting from an uncompressed file avoids a pointless intermediate encode. File size is not a concern since it only lives on your machine.

Archiving your own work

FLAC. You keep full quality at roughly 40 % smaller size than WAV. You can always derive an MP3 later with no additional loss relative to the original.

Sending by email or messaging

MP3 192 kbps is generally enough and stays under common attachment limits. If the exchange is for work approval rather than final listening, say so, to avoid the other party judging quality on that version.

Sample rate and bit depth

Two parameters that accompany lossless formats and often cause confusion.

Sample rate determines the maximum reproducible frequency: half the sampling value. At 44.1 kHz — the CD standard — you cover up to 22 kHz, beyond the limit of human hearing. Moving to 96 kHz does not improve what you hear on a file meant for listening; it is useful in production, for certain processing, but not for distribution.

Bit depth determines dynamic range. 16-bit offers about 96 dB of dynamic range, ample for distribution. 24-bit (about 144 dB) is mainly for recording and mixing, where you keep headroom before clipping. For a final export, 16-bit / 44.1 kHz remains the sensible standard.

Common misconceptions

"Converting MP3 to WAV improves quality." No. The conversion reproduces exactly what the MP3 contained, restoring nothing. The file gets bigger with no benefit. The only valid reason to do it is to avoid recompression during later processing.

"FLAC sounds better than WAV." No. Both contain exactly the same audio data. FLAC is simply losslessly compressed, like a ZIP archive. On playback, the signal is identical.

"You should always export at 320 kbps." Not necessarily. For solo voice or a podcast, 128 to 192 kbps is plenty and halves the file size. High bitrates are justified on music rich in highs and dynamics.

Default recommendation

If you remember only one rule: export to WAV if the file will be reworked, and to MP3 320 kbps if it is destined for direct listening or publication. That single distinction prevents the vast majority of avoidable quality loss.