TuneLab
Lossless Audio Checker
Verify if your FLAC, WAV, or AIFF files are genuine lossless or transcoded. Runs on your device — files never uploaded.
Drop audio file here or click to browse
FLAC, WAV, AIFF, MP3, M4A · processed on your device
How to Check for Lossless Audio
- 1 Upload a FLAC, WAV, AIFF, MP3, or M4A file
- 2 Read the spectrogram — a sharp horizontal cutoff means audio was lossy-encoded
- 3 Use the overlay buttons to highlight evidence: cutoff line, spectral holes, SBR, pre-echo
- 4 Check the verdict and use the forensic guide below the result to understand what was detected
How to Read Your Spectrogram
A spectrogram shows every frequency in your audio over time. The horizontal axis is time (left to right), the vertical axis is frequency (low at the bottom, high at the top), and brightness indicates energy. Here's what to look for.
What does genuine lossless look like?
A true lossless recording fills the entire spectrogram from bottom to top. You'll see color and texture all the way up to the Nyquist frequency (22.05 kHz for 44.1 kHz files). Energy naturally fades toward the top — most music has less high-frequency content — but there's no abrupt cutoff. Just a gradual, uneven fade with visible noise texture above the music content.
What to check: Toggle the Cutoff overlay — it should sit at or near the Nyquist limit. The Holes overlay should show very few dots (natural audio doesn't have psychoacoustic gaps).
What does a transcoded file look like?
A file converted from MP3/AAC to FLAC/WAV shows a sharp horizontal line where all content abruptly disappears. Above the line: solid black. Below it: normal audio. This "shelf" is the signature of lossy compression — the encoder permanently discarded frequencies above that point.
What to check: Toggle Cutoff — the red dashed line sits at the encoder's frequency ceiling. Toggle Holes — orange dots will cluster in the upper bands where the psychoacoustic model removed masked frequencies. If the file was HE-AAC, the SBR overlay may show a blue band where frequencies were synthetically reconstructed.
What do the overlay layers show?
| Overlay | Color | What It Shows | Good Sign | Bad Sign |
|---|---|---|---|---|
| Cutoff | Red line | Detected frequency ceiling | Near Nyquist (>20.5 kHz) | Sharp shelf at 16-19 kHz |
| Holes | Orange dots | Silent bins next to loud neighbors | Few or no dots | Dense clusters = codec masking |
| SBR | Blue band | Synthetically replicated frequencies | No band visible | HE-AAC/mp3PRO signature |
| Pre-echo | Yellow lines | Noise before loud transients | No lines | MDCT block artifact (MP3/AAC) |
How to use the zoom controls
The forensic workstation lets you inspect specific frequency ranges in detail:
- Frequency slider — hover on the left edge of the spectrogram to reveal two vertical sliders. Drag them to zoom into a frequency band (e.g. 15-22 kHz to inspect the cutoff region).
- Ctrl+Scroll — zoom the frequency axis centered on your cursor position. Great for quick inspection without leaving the spectrogram.
- Drag to select — click and drag a rectangle to zoom into a specific time + frequency region. Useful for inspecting individual transients for pre-echo.
- Double-click or press the Reset button to return to the full view.
Frequency cutoff reference table
Different bitrates cut at characteristic frequencies. If the detected cutoff matches one of these, the file was almost certainly transcoded:
| Cutoff | Likely Source | What It Means |
|---|---|---|
| < 16.5 kHz | 128 kbps MP3 | Definitely transcoded — significant quality loss |
| 16.5 – 19 kHz | 192 kbps MP3 | Transcoded — moderate quality loss |
| 19 – 20.5 kHz | 320 kbps MP3 / 256 AAC | Transcoded — minor quality loss, harder to detect |
| > 20.5 kHz | True lossless | No artificial cutoff detected — genuine lossless |
What about fake Hi-Res and upscaled files?
"Fake Hi-Res" files claim 24-bit/96kHz but actually contain 16-bit audio padded with zeros. The tool detects this two ways: by checking if the raw sample bytes have zero-padded lower bits (WAV/AIFF), and by checking if decoded float samples snap to a 16-bit quantization grid (works for FLAC/ALAC too). If detected, the verdict shows Upscaled with the effective bit depth.
Upsampled is different — the file claims a high sample rate (e.g. 96 kHz) but all spectral content stops well below the Nyquist limit, suggesting it was upsampled from a lower-resolution source.
What about the confidence percentage?
Confidence reflects the combined weight of 9 detection signals. High confidence means multiple signals agree strongly (e.g. sharp cutoff + silent noise floor + spectral holes + consistent cutoff variance = definite transcode). Low confidence means signals are mixed or ambiguous — use the spectrogram overlays as additional context. A clean horizontal shelf points to a transcode; a fuzzy, uneven fade suggests natural rolloff from an analog source.
