The loudness, true peak and tempo of 299 finished club records, measured with the engine behind your Master report. This page is the spread they describe.
| Metric | Quietest 10% | Middle | Loudest 10% | Unit |
|---|---|---|---|---|
| Integrated loudness | -11.4 | -9.3 | -7.1 | LUFS |
| True peak | -0.3 | +0.1 | +1.5 | dBTP |
| Peak-to-loudness (PLR) | 7.5 | 9.6 | 11.9 | dB |
| Loudness range (LRA) | 1.8 | 4.7 | 7.9 | LU |
| Tempo | 120.2 | 126.0 | 139.7 | BPM |
Read a row left to right as a distribution: one track in ten measures at or below the first column, half measure at or below the middle, one in ten measures at or above the third. The last two rows are not loudness, so read those columns as lowest tenth, middle, highest tenth.
One record out of 299 measures quieter than -14 LUFS. That is the number most mastering articles quote, and it is a streaming normalisation target: it describes what a player does to a file on playback, not what a club record is mastered to. Spotify, YouTube, Amazon and Tidal sit near it; Apple's Sound Check targets -16 LUFS instead.
Master to -14 LUFS and on a platform that normalises everything, nothing happens: you are turned up to meet the rest. On a club system or in a DJ set, where nothing normalises playback level, the same file sits quieter than everything played around it.
60% of these records measure above 0.0 dBTP, 99% measure above -1.0 dBTP, and the loudest reaches +3.37 dBTP. True peak is the level the waveform reaches between the samples rather than at them, so a file can read under full scale on a sample peak meter and still overshoot once a converter or a lossy encoder reconstructs the curve. That overshoot is where inter-sample clipping comes from.
It is also why the last stage before a file leaves this tool is a true-peak limiter and not a sample-peak one. It runs at 8x oversampling and holds the output at -0.15 dBTP: across 18 test renders the worst one measured -0.16 dBTP. So the file you get back is under full scale on a true-peak meter, not just on a sample meter.
A second library, kept apart from the first and measured with the same engine: 900 commercially released tracks sorted into six genre folders, 150 in each, so no genre can outvote another by sheer count. The labels are ours, not an official taxonomy, and the line between minimal and tech house is a judgement somebody made rather than a fact. The medians are not.
| Genre | LUFS | dBTP | PLR | LRA | BPM | Over 0 dBTP |
|---|---|---|---|---|---|---|
| Techno | -8.38 | 0.00 | 8.50 | 3.29 | 135.0 | 49% |
| Dub techno | -11.62 | +0.05 | 11.71 | 4.99 | 120.0 | 56% |
| Minimal | -10.28 | -0.12 | 10.29 | 4.43 | 127.0 | 35% |
| House | -9.63 | +0.17 | 10.18 | 3.85 | 124.0 | 63% |
| Tech house | -8.66 | +0.22 | 9.06 | 5.27 | 128.0 | 68% |
| Melodic techno | -9.96 | +0.12 | 10.25 | 5.00 | 122.5 | 60% |
Every column is a median over that genre's 150 records. Over 0 dBTP is the share of the 150 whose true peak measures above full scale.
Techno and dub techno sit at opposite ends of both axes at once. Techno is the loudest of the six at -8.38 LUFS and the fastest at 135 BPM, with the lowest PLR and the lowest LRA in the table, which is what heavy limiting looks like from the outside. Dub techno is the quietest at -11.62, the slowest at 120, and the only genre where fewer than nine records in ten clear -14 LUFS.
Minimal is the one genre that keeps its headroom. It is the only median true peak under zero, at -0.12 dBTP, and only 35% of its records go past full scale where the other five run 49 to 68%. Its loudness is mid-table, so this is a choice about peaks rather than a quieter master.
Tech house is loud and wide at the same time: the highest share over -14 LUFS in the table at 99%, the highest share over 0 dBTP at 68%, and also the highest LRA at 5.27 LU. Loud does not have to mean flat, and this row is the proof inside this table.
299 commercially released club records, collected one at a time as mastering references: club electronic, tempo mostly between 120 and 140 BPM, every one of them a finished master someone put out. It is not a random sample of recorded music, not a chart, not a catalogue, and not a claim about what your own track should measure. It is what these 299 measure, and nothing beyond that.