Lossy vs Lossless Audio: A Practical Guide
Every digital audio format falls into one of two families: lossless, which preserves every bit of the original recording, and lossy, which throws some information away to make files much smaller. Knowing which is which — and when each makes sense — will save you storage, avoid quality mistakes and make conversions painless.
Uncompressed, lossless and lossy
Uncompressed PCM is the raw form of digital audio: a long list of sample values, typically 44,100 or 48,000 per second per channel. WAV and AIFF files almost always store audio this way. Nothing is lost, but nothing is saved either — CD-quality stereo needs about 10 MB per minute.
Lossless compression shrinks that data without discarding anything. It works like a ZIP file: patterns and predictability in the waveform are encoded more compactly, and decoding reconstructs the original exactly. FLAC, ALAC (Apple Lossless, usually in .m4a files), WavPack and Monkey’s Audio are lossless. Typical savings are 40–60% versus WAV.
Lossy compression goes much further — often 80–95% smaller than WAV — by permanently removing information that a psychoacoustic model predicts you won’t hear. MP3, AAC, Vorbis, Opus and WMA are lossy.
How lossy codecs decide what to remove
Human hearing has limits that codecs exploit:
- Frequency masking: a loud sound hides quieter sounds at nearby frequencies.
- Temporal masking: a loud sound briefly hides quieter sounds just before and after it.
- Absolute threshold: very quiet sounds and very high frequencies (above ~16–18 kHz for many adults) are hard to hear at all.
The encoder splits audio into frequency bands, estimates what is masked, and spends its limited bit budget on what remains audible. At generous bitrates the result is perceptually transparent. At low bitrates you start to hear artefacts: a “swirly” or underwater quality, smeared cymbals, or pre-echo before sharp drum hits.
Can you hear the difference?
Honest answer: at high bitrates, usually not. Many controlled blind tests have found that most listeners cannot reliably distinguish a well-encoded 256–320 kbps MP3 or AAC from the lossless original on typical music. Differences become audible with low bitrates, certain “killer” samples (harpsichords, castanets, applause), and repeated lossy-to-lossy conversions.
So why does lossless matter? Because it is future-proof. A lossless master can be converted to any present or future format at any quality, with no accumulated damage. A lossy file only ever gets worse with each conversion.
Generation loss: the trap of converting lossy to lossy
Every time a lossy file is decoded and re-encoded, the new encoder makes its own decisions about what to discard, on top of what was already removed. One generation at a sensible bitrate is usually inaudible; several generations add up. That is why you should:
- Convert from the best source available — lossless if you have it.
- When you must transcode lossy audio (for example, M4A to MP3), use an output bitrate equal to or higher than the source.
- Never “upgrade” by converting lossy to lossless: an MP3 converted to FLAC is just a larger file with the same quality.
Comparing the formats
| Format | Type | Typical size per minute | Strengths |
|---|---|---|---|
| WAV / AIFF | Uncompressed | ~10 MB | Editing, universal in studios |
| FLAC | Lossless | ~5–7 MB | Open, great tagging, archives |
| ALAC (.m4a) | Lossless | ~5–7 MB | Apple ecosystem |
| MP3 | Lossy | ~1–2.4 MB | Universal compatibility |
| AAC (.m4a) | Lossy | ~1–2 MB | Efficient, Apple/streaming standard |
| Opus / Vorbis (.ogg) | Lossy | ~0.5–2 MB | Very efficient, open |
For a deeper look at each format, see our common audio formats compared guide.
A sensible library strategy
Most people are happiest with a two-tier approach:
- Archive tier (lossless): rip CDs to FLAC or ALAC, keep studio masters as WAV, and buy lossless downloads when offered. Store this on a computer or backup drive.
- Portable tier (lossy): generate MP3 or AAC copies for phones, cars and sharing. Use FLAC to MP3 or WAV to MP3 at 192–320 kbps.
If you lose or replace a device, you simply regenerate the portable copies from your archive.
When lossy is simply the right choice
Lossless isn’t always better in practice. For speech — podcasts, audiobooks, lectures, voice memos — lossy at 64–128 kbps is indistinguishable from the original and makes files tiny. For streaming and messaging, small files load faster and use less data. And for storage-limited devices, a 192 kbps library holds five to seven times more music than FLAC.
Summary
Lossless formats (WAV, FLAC, ALAC) keep everything and are ideal for editing and archiving. Lossy formats (MP3, AAC, Opus, Vorbis) remove inaudible detail to save space and are ideal for listening and sharing. Keep lossless masters when you can, create lossy copies for everyday use, and avoid stacking lossy conversions.