Audio File Formats
WAV, AIFF, FLAC, MP3, AAC — what each format does, and when to use which.
The file format is the container; the audio inside is the same regardless of container, but the wrapper changes how the data is stored, how big the file is, and what platforms can read it.
The Three Categories
Uncompressed. Audio stored as raw samples. Largest files; bit-perfect.
- WAV (Microsoft / IBM)
- AIFF (Apple)
- BWF (Broadcast Wave Format — WAV with metadata)
Lossless compressed. Audio compressed with no quality loss. Smaller than uncompressed; identical when decoded.
- FLAC (Free Lossless Audio Codec)
- ALAC (Apple Lossless Audio Codec)
Lossy compressed. Audio compressed by discarding "less important" information. Smallest files; quality is reduced and cannot be recovered.
- MP3 (MPEG-1 Layer III)
- AAC (Advanced Audio Coding)
- Opus (modern, high-efficiency)
- Ogg Vorbis (open-source MP3 alternative)
When to Use Which
| Use case | Format | |----------|--------| | Recording session | WAV or AIFF, 24-bit, project sample rate | | Mastering source | WAV, 24-bit, 48 kHz minimum | | Archive | FLAC (saves space) or WAV (universal) | | Streaming delivery | WAV (services convert to AAC/MP3 themselves) | | Personal listening | FLAC or ALAC for lossless; AAC 256+ kbps for portable | | Posting on the web | MP3 320 kbps or AAC 256 kbps | | Voice / podcast | AAC or Opus at 96–128 kbps mono |
WAV vs AIFF
Functionally identical. Same audio data, different file headers. WAV is more universal across Windows software; AIFF is more universal across older Mac software. Modern tools handle both transparently.
If you're sharing files across platforms, WAV is the safer choice.
FLAC vs ALAC
Both lossless compressed. Both will decode to the same bit-perfect data as the source.
- FLAC is open-source, plays on more platforms, and has wider tool support.
- ALAC is Apple's equivalent, preferred in iTunes and Apple Music ecosystems.
If you're not specifically in Apple's ecosystem, default to FLAC.
MP3 vs AAC
Both lossy. AAC is technically superior — at the same bitrate, AAC sounds better.
- MP3 320 kbps — the highest-quality MP3. Fine for distribution; some audible artifacts on critical listening.
- AAC 256 kbps — sounds noticeably better than MP3 at the same data size. Default for streaming services.
- Opus — modern codec with excellent low-bitrate performance. Great for voice and streaming.
For new content creation, AAC is the better default. MP3 remains the most universal.
Bitrate Guidance for Lossy
| Bitrate | Quality | Use case | |---------|---------|----------| | 64 kbps | Voice only | Podcast, audiobook, AM radio quality | | 128 kbps | OK for voice; mediocre for music | Background music, low-bandwidth streams | | 192 kbps | Good for most music | YouTube audio, casual streaming | | 256 kbps | Very good | Apple Music, premium streaming | | 320 kbps | Near-transparent | Spotify Premium, audiophile listening |
Variable bitrate (VBR) is generally preferable to constant bitrate (CBR) at the same average — it allocates more data to complex passages and less to silence.
Sample Rate and Bit Depth Are Independent
A WAV file can be 44.1 kHz / 16-bit or 96 kHz / 32-bit float. A FLAC can be either. The container and the codec don't dictate the sample rate or bit depth — they hold whatever you put in.
The exception: lossy formats often have tight rate ranges. MP3 supports 16, 22.05, 24, 32, 44.1, 48 kHz. AAC supports up to 96 kHz but most decoders cap at 48 kHz. Higher rates than these aren't usually needed for lossy delivery.
What About 32-bit Float WAVs?
Some DAWs export 32-bit float WAV files. These are useful as working files (no clipping in the file even if the master bus clipped) but are larger than standard 24-bit WAVs and not universally accepted by mastering services and distributors.
If you're delivering a 32-bit float file, confirm the recipient can read it. Otherwise convert to 24-bit before sending.
Common Mistakes
Sending MP3s to a mastering engineer. They need full-resolution data to work with. WAV / 24-bit minimum.
Re-encoding MP3s. Decoding an MP3 and re-encoding to MP3 (or any lossy format) compounds the data loss. Always start from the original WAV / FLAC source.
Streaming a low-bitrate MP3 thinking high-bitrate streaming will fix it. The damage is in the encode. Re-encoding from a 128 kbps MP3 to 320 kbps gives you a bigger file, not a better-sounding one.
Inconsistent sample rate within a project. A session at 48 kHz with a 44.1 kHz audio file imported will need resampling — most DAWs do this automatically but the result isn't always great. Match all assets to the project rate.