enguetedl.ch is a small, ad-free Swiss tool for DJs: reference mastering, six-stem separation, loudness analysis and format conversion.
It measures a reference track you pick (its loudness, its tonal balance, the shape of its EQ), then moves your track toward it. No reference on hand? Auto-match does the same against an in-house target built from club records we measured with this same engine.
If the mix underneath has a problem the reference does not share, matching will not find it.
Pick one whose shape you like, not the closest-sounding track you can find. Matching brings your loudness and tone into the reference's range; it does not clone it, so stop once you are in range rather than once it is identical.
Because we check it rather than trust it: the top end is inspected for the hard shelf a lossy encoder leaves, so renaming an .mp3 to .wav does not get it past. A dark master with little above 16 kHz still passes, and your own track is never held to this check.
Six WAV files in one zip: vocals, drums, bass, guitar, piano, and other (synths, percussion loops, FX). They come from htdemucs_6s, the six-source variant of Demucs, the one place on enguetedl.ch where a trained model runs.
htdemucs_6s is Hybrid Transformer Demucs, published by Simon Rouard, Francisco Massa and Alexandre Défossez at Meta AI (ICASSP 2023). It reads the waveform and its spectrogram at once, and learned what an instrument sounds like from MUSDB18-HQ, a research set of 150 tracks, plus roughly 800 further songs.
The model's own authors say so about this variant: guitar is okay, piano is not, with a lot of bleeding and artifacts. Vocals, drums and bass were the original four-source job and come out best.
Yes. A whisper of hi-hat under the vocals or a shadow of bass in the guitar stem is normal, not a failed job. Expect stems clean enough to loop a vocal or mute the drums, not a studio multitrack.
Integrated loudness, true peak, loudness range, peak-to-loudness, tempo, musical key, and how the energy is spread across sub, bass, mids and top end. It reads the track once and keeps the numbers; nothing is written back to the file.
The list of every track you have analysed. It sorts on any column, files tracks into playlists, and puts two side by side so you can see whether one will jump 4 dB louder than the other in a set.
The terms the Master report and the Library print on screen, each one defined on its own.
We measured 299 club records with the same engine the app uses and published the whole distribution, so you have something to compare your own against.
Only the container and the codec. Bit depth, sample rate, tags and artwork come out the way they went in, and the finished row tells you exactly what came out. The one setting is the format you want.
You can, but it brings nothing back: the result is a bigger file carrying the same lost detail. Convert lossless to lossy when your rig needs it, and keep the lossless original.
Your upload is processed and then destroyed, and the file it produces stays available for you to collect for 60 minutes before it goes the same way. Everything else we hold is on our privacy page.