What Is Lossless Audio? Can You Hear the Difference From MP3?

2026-11-21 ·

Lossless means compression without loss: after decoding, the file matches the original sound bit for bit. MP3 and AAC work differently. They discard information that the algorithm judges hard to hear in order to save space, and that information cannot be recovered. In blind tests at high bitrates, listeners often cannot reliably tell a lossy file from the original, but at 128 kbps the difference is easier to hear. Below: what lossless is, how FLAC, ALAC and WAV differ, what hi-res adds and what it does not, why a "lossless" file made from an MP3 is a fake, and how to test the difference yourself.

What lossless is, and how it differs from lossy

Any digital sound is a sequence of numbers: how many times a second the air pressure is measured, and how precisely each value is stored. The uncompressed form of that sequence is called PCM, and it is usually stored in WAV or AIFF. Its size is predictable. For stereo at CD quality (16-bit, 44.1 kHz) it is about 1,411 kbps, or roughly 10 MB per minute.

Lossless compression works a little like an archiver. The algorithm finds repetition and predictable patterns, stores them more compactly, and on decoding restores everything to the last bit. FLAC and ALAC (Apple Lossless) work this way. A FLAC file is usually around half the size of the same WAV, and the sound is identical.

Lossy codecs such as MP3, AAC and Opus rely on psychoacoustics. They split the signal into bands and discard what a model of hearing says should be inaudible, for example quiet components next to loud ones. The saving is large: MP3 at 320 kbps is about 23 % of the size of the WAV, and at 128 kbps about 9 %. The price is irreversible loss. You cannot turn a lossy file back into lossless, because what was discarded is gone.

Format Type What it stores Size compared with WAV Typical use
WAV uncompressed PCM 100 % studios, exchanging files between programs, mastering
FLAC lossless PCM, compressed without loss usually about 50–60 % archives, catalogues, releases
ALAC lossless PCM, compressed without loss, in an M4A container usually about 50–60 % Apple libraries
AAC lossy perceptual coding depends on bitrate streaming, phones
MP3 lossy perceptual coding about 23 % at 320 kbps, about 9 % at 128 kbps almost any device

FLAC, ALAC and WAV: what differs in practice

These three formats sound the same: if the compression is lossless, everything matches the source after decoding. What differs is convenience.

WAV is the simplest option. It is uncompressed, and almost any program opens it, but it takes the most space, and its metadata (title, artist, cover art) is handled less well than in FLAC.

FLAC is an open lossless format. It supports tags and cover art, most players and programs read it, and that is why it is a common choice for archives and catalogues.

ALAC is lossless inside an M4A container. It fits well in the Apple ecosystem, while support on other devices is usually weaker.

For choosing a format for a particular job, see the audio formats for mastering article, which compares WAV, FLAC, MP3 and AAC and says which to choose for streaming, CD, vinyl and podcasts.

Hi-res: what it adds and what it does not

Hi-res usually means 24-bit depth and a sampling rate above 44.1 kHz, for example 24-bit / 96 kHz. The two parameters work differently.

Bit depth sets the dynamic range. In theory it is about 96 dB at 16 bits and about 144 dB at 24 bits. For listening to a finished track this headroom is rarely needed: the room's background noise and the equipment's own noise sit well above the noise floor of a 16-bit file. Where 24 bits does help is in recording and processing, because quiet passages and intermediate calculations accumulate fewer errors.

Sampling rate sets the upper frequency limit. At 44.1 kHz the signal reaches 22.05 kHz, and the threshold of hearing for most people is around 20 kHz. At 96 kHz the limit rises to 48 kHz, and all of the added range lies beyond what anyone can hear. For studio processing, a higher rate can give extra margin against distortion from non-linear effects, but that is a property of the processing, not something the listener hears.

Blind-test results are mixed. A study by Meyer and Moran (JAES, 2007) found no audible difference between a CD-standard version and high-resolution audio at normal and loud listening levels. A meta-analysis by Joshua Reiss (JAES, 2016) of 18 published experiments reached a different conclusion: listeners have a small but statistically significant ability to tell the formats apart, and trained listeners do better. Some critics questioned how the studies were selected. None of this supports a flat claim that hi-res sounds better, and presenting it as a proven fact would be dishonest.

Why a "lossless" file made from an MP3 is a fake

Sometimes a file labelled FLAC or WAV was in fact an MP3. It happens in download catalogues, on sites that offer free downloads, and from sellers of beats. Converting an MP3 to FLAC makes the file bigger but restores nothing: the lost information does not come back. The sound stays the sound of an MP3, even if the name says lossless.

You can check this with a spectrum. Almost any audio editor or free analyser can show a spectrum or spectrogram. In a file that was once an MP3, the top frequencies usually end in a sharp cut-off. At 128 kbps the cut-off usually sits around 16 to 17 kHz; at higher bitrates it is higher, but the exact values depend on the encoder and its version. A cut-off on its own is a reason to be suspicious, not proof, because a natural roll-off also occurs in honest recordings.

The reverse is not reliable either. A spectrum that is completely empty above 22 kHz proves nothing, because many 96 kHz recordings contain no ultrasonic content. The more dependable approach is to ask the seller or the artist where the file came from, and then check with a blind test.

A blind test you can run yourself

The reliable method is ABX. The steps are these:

  1. Prepare two versions of the same passage: A is the original lossless file, and B is the same track as an MP3 at the bitrate you want to test.
  2. Someone, or a program, randomly chooses which version will be X. You listen to A, B and X as many times as you need, then say whether X matches A or B.
  3. Repeat this 16 times. Guessing at random gives 8 correct answers out of 16 on average.
  4. Twelve correct answers out of sixteen gives a probability of about 4 % that the result is chance, which is already statistically noticeable. Thirteen out of sixteen gives about 1 %.

To keep the test fair, match the loudness: the louder version almost always sounds better. Measure the integrated loudness (LUFS) of both versions and adjust the levels so that the difference is a fraction of a decibel. Use headphones you trust, and do not turn it into a marathon, since fatigue spoils the result.

What to listen for:

  • cymbals and hi-hats: the decay can sound smeared or metallic at low bitrates;
  • sibilants in vocals: "s" and "sh" sounds can turn distorted;
  • reverb tails: the quiet decay, where coding artefacts are easier to hear;
  • drum transients: a faint smear just before a hit (pre-echo) is a typical sign of lossy coding.

When lossless is needed, and when it is not

Lossless is needed in three situations.

Working on a track. Mixing and mastering are best started from a source without loss. Every lossy re-encode adds artefacts, and no tool can fully remove them at the end.

Archiving. Keep your source files in lossless. You can make an MP3 at any bitrate from them, but you cannot get lossless back from an MP3.

Sending a track for mastering. Use WAV or FLAC, so the service does not have to work with sound that has already lost information.

For everything else, lossless is usually not necessary. Listening on a phone, background music, streaming and podcasts are all well served by a good lossy file. Lossless files take noticeably more space: a FLAC is usually two to three times the size of an MP3 at 320 kbps, and several times the size of an MP3 at 128 kbps. To choose the download format for a finished master, WAV, MP3 or FLAC, read download your master as WAV, MP3 or FLAC.

How to do it in Magic Master

The audio converter converts files between WAV, MP3 and FLAC. The outputs available are MP3 at 320 kbps, WAV at 16 or 24 bits, and FLAC. The converter shows LUFS and True Peak, so it is also handy for measuring loudness before a blind test. Files can be up to 100 MB, with five conversions per day from one address, and conversions do not use tokens.

Mastering accepts WAV, FLAC, MP3, M4A, OGG and other common formats. Its outputs include WAV at 16 and 24 bits, FLAC at 24 bits, and MP3 from 128 to 320 kbps. The 16-bit WAV uses TPDF dither by default, while the 24-bit WAV has no dither. Base mastering in the app is free within the quota: guests get three runs per calendar month, and signed-in accounts without tokens get one per day.

Common mistakes

  • Converting an MP3 to FLAC and expecting the sound to improve.
  • Comparing versions at different loudness: the louder one almost always seems better.
  • Judging a file by its format name or extension instead of checking the spectrum and running a blind test.
  • Chasing hi-res for the number on the label: for listening, it does not guarantee any audible difference.
  • Keeping only compressed versions, and then looking for a source file when you need to master.

Summary

Lossless is an honest format for working and archiving, because it keeps everything that was recorded. For listening, a good MP3 or AAC is usually enough, and arguing about the difference makes little sense until it has been tested blind. If you need a master, use WAV or FLAC and do not re-encode. If you are unsure, run an ABX test on your own track.

Questions and answers

What does lossless mean?

Lossless means compression without loss: after decoding, the file matches the original sound bit for bit. MP3 and AAC throw away part of the information to save space, and that information cannot be brought back.

Which music format sounds best?

Uncompressed WAV and the lossless formats FLAC and ALAC carry the original sound without loss, so they are equal in quality. Choose by convenience: FLAC takes less space, while almost every studio program reads WAV.

What is FLAC?

FLAC is an open lossless compression format. A FLAC file is usually about half the size of the same WAV, and it decodes to the exact original, which is why it suits archives and releases.

Is AAC better than MP3?

Both are lossy. At the same bitrate AAC usually sounds cleaner, which is why it is common for streaming and phones. The difference is most noticeable at low bitrates.

Can you hear the difference between lossless and MP3?

It depends on the bitrate and the listener: at a high bitrate, listeners often cannot pick out the MP3 in a blind test, while at 128 kbps the difference is easier to hear. An ABX test on your own track is the honest way to find out.

Try Magic Master

Upload a track — a finished master in about a minute.

Open mastering →AI mastering onlineLUFS analyzer