What is a fake 320 MP3?

A file encoded at 320 kbps from a source that was already low quality, usually a 128 kbps rip. The header says 320 and the file is large, but the audio that was thrown away by the first encode never comes back.

Can you hear a fake 320?

Sometimes, on a good system: cymbals and air sound dull or splashy. On club speakers at volume it is much harder, which is why it is worth checking the files rather than trusting your ears in the booth.

How can you prove it?

Look at the frequency cutoff. Every lossy encoder discards everything above a ceiling that depends on the bitrate it used. A real 320 carries content up to roughly 20 kHz. A 128 rip stops near 16 kHz and re-encoding it at 320 does not move that wall.

TrackGemz · Guides

How to tell if an MP3 is a fake 320

The bitrate in a file's header is a claim, not a measurement. Anyone can take a 128 kbps rip, re-encode it at 320 and hand you a file three times the size with none of the detail back. Pools and packs are full of them, and they sound fine until the room is loud.

Why upscaled files exist

Mostly not malice. A track gets ripped once at low quality, then passes through a chain of re-encodes as it is shared, converted and repacked. Every step writes a new header. By the time it reaches you it says 320, and nothing in the file records what it used to be.

The audio itself does remember, though. Lossy encoding throws away the top of the frequency range, and a later encode at a higher bitrate cannot invent it back. That missing ceiling is the fingerprint.

What the cutoff tells you

Open a track in a spectrogram view and look at the top edge. Roughly:

A real 320 carries content up to about 20 kHz, with a soft top edge.
A 192 tends to stop around 19 kHz.
A 128 stops near 16 kHz, and the edge is a hard flat line.

A flat wall well below 20 kHz on a file that claims 320 is the giveaway. The wall is the ceiling of the first encode, preserved perfectly by every encode after it.

A low cutoff is not always fraud. Old recordings, vinyl rips, live bootlegs and some mastering choices genuinely have nothing up there. That is why a cutoff should be reported for you to judge rather than acted on automatically.

Checking one file, and checking ten thousand

For a single track, any spectrogram tool will do. That approach stops working the moment you want to know about a whole library, because you cannot eyeball forty thousand pictures.

TrackGemz measures the cutoff of every file and compares it with what the file claims, then sorts the results into tiers so you can start with the worst. It reports rather than deletes: you see the claimed bitrate, the effective one and the cutoff, and you decide what to replace.

Scan your library or a folder.
Open the upscales list. Files are grouped by how far the measurement falls short of the claim.
Work down from the worst tier, replacing what matters to you and leaving the vinyl rips alone.

On one real library the bulk of the suspect files were exactly 128 kbps under a 320 label, which is what a single generation of upscaling looks like.

Download for Mac See pricing

Free to try. macOS 12 or later, Apple Silicon and Intel.