A bigger file preserves more information. Whether you can hear that information is a separate question.

How to take the lossless audio test

Choose 128 or 320 kbps MP3, play samples A and B to the end, then pick the one you think is lossless—or choose “I can’t tell”. Reveal the formats and use “Shuffle & try again” for another round.

Use the speakers or headphones you normally enjoy, in a quiet space, at a comfortable volume. Turn off spatial effects, volume normalisation and other optional processing where possible. Keep the playback route and volume unchanged between samples.

Does Bluetooth affect a lossless audio test?

For a cleaner comparison, use wired headphones or a wired connection to your system. A Bluetooth connection may apply another lossy codec to both samples. This page cannot verify the format that ultimately reaches your headphones or speakers.

Can you hear 320 kbps MP3 vs lossless?

It depends on the music, your listening conditions and the encoding. At 320 kbps, a difference may be subtle or inaudible to you. This test lets you compare the same excerpt on your own setup without knowing which format is playing.

Listen to the texture of the soft percussion, the attack of the plucked notes and the sound as it fades away. Lossy encoding can change these details. Avoid deciding that a brighter, louder or simply different sound must be the original.

Start at 128 kbps, then try 320 kbps. The higher setting gives the MP3 encoder more data to work with. Keep everything else the same when you change that setting.

What is the difference between WAV, MP3 and FLAC?

The WAV in this experiment contains uncompressed, 24-bit stereo PCM audio at 44.1 kHz. The MP3 is encoded directly from that same WAV. There is no different performance, remaster or EQ to confuse the comparison.

MP3 uses perceptual coding to reduce the data rate, discarding or approximating information according to a model of what we are likely to hear. Lossless formats such as FLAC and ALAC can reduce file size while preserving the decoded audio exactly. WAV is a container; here, it holds the uncompressed reference.

What does your listening-test result mean?

A forced choice between A and B has a 50% chance of being right by guessing. Shuffle the labels and repeat before drawing a conclusion. “I can’t tell” is a useful answer too.

Recognising a difference here does not rate your system, and missing it does not mean your hearing or equipment is poor. Your result belongs to this excerpt, this encoding and these listening conditions. A repeatable test with more music would give you stronger evidence.

This is an informal, label-hidden A/B comparison. It is not a controlled laboratory test or an ABX test, where an additional unknown sample must be matched to a reference.

How the samples and comparison work

“Small Hours” is an original composition made for this page in GarageBand, with piano, guitar, bass, strings, vibraphone and drums. These are software instruments, not a live acoustic recording. Both MP3 options are made with the LAME encoder from the exact same 25-second PCM master.

Both files load and decode before playback is enabled. The player matches their average RMS levels, applies the same conservative peak ceiling and plays them through the same browser audio path. RMS matching reduces an overall level cue; it does not guarantee identical perceived loudness. Each play begins at the start, with a short fade at either end.

The assignment of WAV to A or B is randomized independently for every round. The format labels appear only after both complete samples have played and you have selected an answer. Results stay in memory for this visit and reset when you reload the page. File formats are hidden in the interface, not from someone inspecting browser developer tools.

The original WAV, encoded files and generation notes are available with the sample documentation.

Further reading

← All Room Tuner guides