Mix Analyzer

The report shows tempo movement, an energy arc, integrated loudness (LUFS) and loudness range (LRA), true-peak, per-track key where detectable, and quality checks like clipping and dead air.

It marks likely track changes where spectral content, energy, and tempo shift at the same moment. These are soft markers, not judgments — a seamless blend may leave no clear boundary, and a single continuous track shows none at all.

No. Auto-detect infers the style from the set's tempo range and you can override it, but the measurements — tempo, loudness, and key — are read the same way for every genre. The setting is optional context, not a filter on the numbers.

Where a track's key is clear enough to detect, each track change is labelled with its Camelot codes — for example 8A to 9A. The report shows the key movement and leaves the call to you; it does not rate transitions as good or bad.

A single-track tool reads one file. The mix analyzer reads a whole set — tempo movement across every track, the energy arc, loudness for the full file, and where the track changes fall. For one track, use the BPM Finder or Key Finder instead.

No. The full analysis pipeline runs directly on your device. A 120-minute mix stays on your machine — no upload, no waiting for server processing, no file size restrictions.

A DJ set is one audio file with many tracks blended together. The analyzer measures the whole file end to end and marks where the tracks change. Everything runs on your device in a handful of passes over the audio.

Track-change detection

The analyzer computes spectral flux — the per-frame change in frequency content — across the full set. Sharp flux spikes mark moments where the spectral content shifts, which usually means a new track entering. It cross-references those with energy-envelope shifts (sudden RMS changes) and tempo-grid breaks (BPM deltas above a threshold). When two of the three line up within a one-second window, that timestamp becomes a soft track-change marker — a candidate, not a judgment.

Tempo timeline

Tempo comes from onset detection and autocorrelation of the onset envelope, measured segment by segment rather than as a single number. The timeline shows how tempo moves from track to track across the set. Where a segment is too ambiguous to call — a breakdown, an ambient passage, a beatless intro — the cell is greyed instead of showing a fabricated number.

Loudness and key

Loudness follows the EBU R128 standard: integrated LUFS for the whole set, loudness range (LRA) in LU, and true-peak in dBTP. Per-track key is detected where the tonal content is clear enough, and labelled with its Camelot code. For a single track instead of a full set, use BPM Finder or Key Finder for one-file BPM or key readings.

The waveform panel shows the energy arc — low-, mid-, and high-band RMS across the whole set — so you can see where it builds and where it drops. Under it, each detected track change is listed with its timestamp and, where both keys are clear, the Camelot codes on either side (for example 8A to 9A). These are labels, not judgments: the report shows the movement and leaves the interpretation to you.

Quality checks

• Clipping — samples reaching 0 dBFS, flagged with a timestamp. A true-peak over 0 dBTP can also distort on some systems even when the raw samples do not clip.
• Dead air — long silent gaps, usually an export or concatenation slip. Flagged so you can trim before publishing.
• Loudness range — a very low LRA means the set is heavily compressed and flat; a very high LRA means big level swings between tracks. Neither is wrong, but both are worth knowing before a club or streaming upload.
• No track changes — can mean one continuous track, or blends smooth enough that no clear boundary exists. A lot of markers on a short file usually points to busy, cut-heavy mixing.

Every reading here is a measurement — there is no pass or fail. To dig into the harmony behind the Camelot labels, the Chord Detector analyzes individual tracks, and the Camelot Wheel shows how the codes relate.