How is this production so clean??

Literally every single sound sounds so clear and distinct from every other sound. I've tried many times to get my mixes sounding even half as clean as this, but it always resulted in cutting off half of the frequency of each sample in an EQ.

Here it sounds like each sample has been EQ'd with an incredibly sharp peak at the sample's own specific frequency (like a dirac delta function).

How does one achieve this?

  • i feel like leftaroundabout said the most part, but you should also check that the track is not hard limited, there's a lot of air to breathe. But generally i didn't like the energy of the track, felt somewhat empty and like a series of boxy sterilized samples. I mean aphex twin's sound is 100 times more separated and clean in this genre , you can check Iz Us, alberto balslam, come to daddy (lord faulteroy mix)
    – frcake
    Jan 20, 2016 at 12:12

2 Answers 2


Part of the answer is the fact that the cut is almost entirely made up of short and percussive sounds. With no sustaining voices or long release times there is not much material to clutter the same frequency range at the same time. The few sustaining parts are pretty mid-range and appear to be heavily compressed. Also the use of effects is limited. Reverb is carefully used in such a way that it does not cause any one sound to sustain and drape others.

  • Totally agree. If you want to have a clear mix you gotta have a clear instrumentation. For example if you want to have a distorted and an acoustic guitar at the same time, you'll end up with some faint plucks or a fizzy noise without balls. But if these two instruments never played together, but played in each other's pauses...
    – atoth
    Mar 26, 2015 at 22:32

What's notable about this mix is how it uses the stereo width. Only the lead synth (which has indeed quite a narrow frequency range with it's flute-like sound) and the bass parts are center, everything else is spread strongly to the sides. Not in the sense that instruments are panned hard left of right – that has been somewhat of a taboo since the 70's. Rather they are all “wide” stereo sounds to begin with. For instance, the clap sound appears to come almost from behind. That is probably achieved by using two completely uncorrelated signals for the L and R components (e.g. actually record both channels seperately, with in principle the same sound but no actual phase relation). Other possibilities to spread out a signal across the stereo canvas include:

  • Boosting the S component (in an M/S basis representation) of an ordinary stereo signal, e.g of an X/Y recorded piano.
  • Introducing an artificial phase delay on one of the channels – for instance, an actual time-domain delay (1~10 ms, no feedback or clean part), or an all-pass filter.
  • Using lots of very short reverb. Perhaps ambience / gated reverb to avoid too much smearing.

Apart from that – well, EQ and compression do of course play a role, compression definitely a large one. But I think you have been taught a bit too much about “frequency separation” with EQ. It is a valuable thing to know about, and in particular it's good to develop a sense for which components of a signal aren't needed, when they interfere with other instruments, and how to rigorously prune away such components without disrupting the actual character. However, it's not in general true that you need to categorically keep apart which instrument “owns” which frequencies: some overlap is just natural, indeed essential for a proper harmony sensation. And in particular for everything percussive it's crucial that you allow for a wide frequency bandwidth, because only that actually makes sharp transients possible. (So much for Dirac delta function: the Fourier transform of that is just an infinite-lasting sine wave, quite the opposite of what's needed for a track like this!)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.