What is a truly lossless file?

"Lossless" gets thrown around a lot these days. Streaming services slap it on their premium tiers, download stores use it as a selling point, and file extensions like FLAC and WAV have become shorthand for "the good stuff." But what does it actually mean for a file to be truly lossless?

The word hides a distinction that trips up even careful listeners: there's lossless the format, lossless the content, and lossless the provenance. All three have to hold before a file deserves the label, and most files wearing it only guarantee the first. Unpacking that is the rest of this article — and it's the foundation everything else on this site is built on.

The simple definition

A lossless file contains every single bit of information that was in the original recording. Nothing thrown away, nothing approximated, nothing smoothed over. When you decode a lossless file, you get back a bit-perfect copy of the source audio — the same way unzipping a document gives you back the exact document you zipped up.

There's a more precise way to say it that turns out to matter: losslessness is a property of a transformation, not of a file sitting on disk. A FLAC is lossless relative to whatever was fed into the encoder. Feed it a studio master, and it's a perfect copy of a studio master. Feed it a 128 kbps MP3, and it's a perfect copy of a 128 kbps MP3. The format kept its promise in both cases. The promise was just narrower than most people assume.

This is different from lossy formats like MP3 or AAC, which work by permanently deleting audio information they think you won't notice. Mostly the very high frequencies, quiet sounds masked by louder ones, and fine spatial detail. Clever psychoacoustic tricks — but once that data is gone, it's gone forever.

The formats that qualify

The usual suspects: FLAC, ALAC (Apple's version), WAV and AIFF (uncompressed), plus less common ones like WavPack and Monkey's Audio. FLAC and ALAC compress the audio to save space — typically 40–60% smaller than WAV — but the compression is reversible, like a ZIP file for music.

A quick map of the landscape, since the ecosystem is more crowded than the big four:

Missing from that table, deliberately: MP3, AAC, Ogg Vorbis, and Opus are always lossy, no matter the bitrate. And MQA deserves its own footnote — it was marketed as lossless-class, required licensed decoding to reach its claims, and its owner went into administration in 2023. Treat "MQA" badges as a historical curiosity.

Here's the catch

A lossless format doesn't guarantee lossless content. This is the part nobody tells you.

Nothing stops someone from taking a 128 kbps MP3 — which has already thrown away most of its frequency content — and converting it to FLAC. The resulting file is bigger, says "FLAC" on the label, and will fool most players and plenty of people. But it's a fake. The audio inside is still the hollowed-out MP3 version, just wrapped in a lossless container. It's like photocopying a low-res photo onto expensive paper.

This happens constantly. Files get shared, re-encoded, upscaled, and re-shared until nobody knows where they came from. Some download stores have even been caught selling upscaled files as "lossless."

The scale of the problem surprises people the first time they audit a library. Community members who run batch analyzers over large, long-accumulated collections routinely report double-digit percentages of "lossless" files failing the cutoff test — with the worst hit rates in exactly the places you'd predict: free blog downloads, forum shares, and anything described as a "web rip" of unclear origin. Files from reputable stores fail rarely, but not never, because stores sell what labels deliver and labels occasionally ship a master with a lossy step in its own history. The point isn't that the world is full of scammers. It's that audio passes through too many hands for trust to substitute for measurement.

There's also a purely accidental route to fake lossless worth knowing about: default settings. For years, ripping software and converter tools defaulted to lossy output, and plenty of people "converted their library to FLAC" by first decoding MP3s they'd forgotten they had. The resulting files are honest mistakes with dishonest labels, and they circulate exactly like the deliberate ones.

The provenance chain: what "truly" actually means

Since losslessness is relative to the input, the only honest definition of truly lossless involves the whole chain from the microphone to your disk. Something like: every step between the original master and your file was either bit-perfect or lossless. One lossy step anywhere in that chain — one MP3 conversion in 2007 that nobody remembers — and the file inherits the damage forever, no matter how pristine every later step was.

Chains you can trust tend to be short and documented: a CD you ripped yourself with a secure ripper (one generation, zero conversions); a purchase from a store that takes label masters directly (Qobuz, Bandcamp, HDtracks, Bleep); a download straight from the artist. Chains you can't trust are long and anonymous: forum re-shares, blog aggregates, "complete discography" torrents, marketplace sellers with thousands of titles and no quality control. The audio doesn't know where it's been, but it carries the scars.

A useful mental model: provenance is like a chain of custody in evidence law. Every handoff is a chance for tampering, and "found it on the internet" is not a chain of custody.

How to spot a fake

Your ears might not catch it, but the spectrum doesn't lie. Lossy encoders leave a fingerprint: a hard cutoff in the high frequencies. A 128 kbps MP3 typically chops everything above ~16 kHz; a 320 kbps MP3 usually cuts around 20 kHz. A genuine lossless file has content all the way up to the Nyquist limit (half the sample rate — so 22.05 kHz for CD-quality 44.1 kHz audio).

That's exactly what WhatsMyBitrate checks. It looks at where the frequency content actually stops, compares it to where it should stop for a genuine file, and tells you whether your "lossless" FLAC is the real deal or a re-encode in disguise. Try it on a few files in your library — the results can be eye-opening.

For the full forensic method — the cutoff table, the compression-ratio trick, the metadata fingerprints, and what a clean result looks like — the FLAC verification guide walks through it step by step, and the WAV guide covers the hi-res variant of the same scam.

What about streaming "lossless"?

Apple Music, Tidal, Deezer, Amazon Music, and Spotify's long-promised tier all advertise lossless streaming. Are those files truly lossless? Mostly, with caveats worth knowing. The streams themselves are generally genuine FLAC or ALAC at 16-bit/44.1 kHz or better — the services have too much scrutiny to fake the container. The caveats live one step up the chain: the service plays whatever the label delivered, and labels have been caught delivering upsampled or loudness-squashed masters. "Lossless delivery of a compromised master" is still lossless delivery. The format is honest; the provenance is only as good as the label's.

Hi-res tiers (24-bit/96 kHz and up) add a second caveat: the ultrasonic content those formats can store is inaudible, so their practical advantage is usually that they're simply different masters — often better ones, sometimes worse ones. The badge guarantees the format, never the mastering.

Frequently asked questions

Is CD quality "truly lossless"? Yes, with one honest boundary: a CD rip is a bit-perfect copy of the CD, full stop. Whether the CD captured everything the studio had is a different question — CDs are limited to 16-bit/44.1 kHz by design, and a master recorded at higher resolution was already reduced to fit. Truly lossless means lossless from the CD onward, which for most of music history is as far back as the chain can practically go.

Is a vinyl rip lossless? It can be losslessly stored, but vinyl is an analog medium — the rip is a recording of a playback event, complete with the cartridge's noise, the turntable's rumble, and the wear on that particular copy. That makes vinyl rips genuinely unique objects rather than perfect copies, which is either their charm or their limitation depending on what you wanted.

Does converting FLAC to WAV (or back) lose anything? No. Lossless-to-lossless conversions are mathematically exact in both directions — decode one and encode the other and the PCM samples are bit-identical. You can bounce between FLAC, ALAC, WAV, and AIFF forever without a single sample changing. The losses only enter when a lossy codec touches the chain. The one practical footnote: WAV's metadata support is historically messy, so tags and album art sometimes fail to survive the trip even though the audio does. If your tags matter, FLAC or ALAC are the safer long-term containers.

Why does truly lossless matter if I can't hear the difference? Archiving and processing headroom — covered in depth in the blind-testing article. The short version: you archive for the devices, formats, and processing you haven't thought of yet. Lossy files make permanent decisions about what you're allowed to keep; lossless files don't.

The takeaway

A truly lossless file is one where the audio data traces back to the original master without any lossy step in between. The container matters less than the provenance. Trust the spectrum, not the extension.

If you take one habit from this article, make it this: any time a lossless file enters your library from a source you don't fully control, spend five seconds checking it before it gets filed away. Verification is cheap, the tool is free, and a library built on checked files is the only kind whose labels you can actually believe.