EDM final bounce - compressor before sum or sum then compres

Just would like to get some feedback for anyone that creates EDM music and is more along the lines of an advance producer and there theory about when they apply final compression to an overall mix. For example lets say you are summing, do you like applying compression to the outs of your sum before it hits the final limiter or do you prefer to have compression in play before it hits the summing stems?

Thanks

If the premise is that you're mixing, then slapping on anything at the end of the mixing process is pointless. This type of processing is better left to mastering.

If you like to work with some compression on the mix bus while you're mixing, then do so fairly early in the process, e.g. when the basic beat is up and running.

Any such processing should be kept on when bouncing the mix for mastering since it's integrated with the mix and you've made decisions in the mix based on how the compressor reacts. Removing and later trying to replicate such as mix bus compressor during mastering is possible, but rarely recommended.

Sum compressing sub groups in the mix is a creative and sometimes technical choice. It's a different proposition to mix bus compression and isn't necessarily an either/or scenario, since sum compressing sub groups and mix bus compression can be combined.

However, compression in EDM is mostly about creative shaping, rarely about control, and almost never about loudness. So think before you compress.

To sum things up (see what I did there?):

· Adding a compressor at the end of the mixing process = pointless, better left for mastering

· Adding a compressor fairly early in the mix process = keep it on when bouncing for mastering

· Compressing groups and the mix bus aren't mutually exclusive

· Compression in EDM is about shaping, rarely about level

What do you mean by "shaping"?

Do you mean like fx. giving the sounds more attack, by compressing everything else but the attack?

And when saying compression in EDM is almost never about loudness, is it because the sounds already are static? Is it because its sample based, and therefore doesn't have variable peaks like in recorded music?

Would you prevent using subcompression and mixbusscompression in EDM?

What do you mean by "shaping"?

Levelling compression means manipulating the dynamics so the instrument track appears more even in terms of loudness. While levelling also affects the character of the sound, you're trying to avoid audible compression. When levelling, you're usually concerned with macro dynamics, i.e. getting different segments or phrases at the approximate same level.

Shaping is all about audible compression. Audible in this case doesn't mean artifacts such as pumping and breathing, but that you're trying to manipulate micro dynamics, e.g. the transients or tails of individual notes or parts of the signal. Shaping could mean controlling the tail of a kick drum, adding snap to a snare drum, controlling the attack of a guitar, etc.

Some of the techniques used in levelling and shaping are the same, but they're used for different reasons - and the devil's in the detail. Think about the intention when you're compressing.

A lot of people are in fact doing shaping compression with the intention of levelling, which means they're missing the point and getting the worst of both.

When dealing with vocals and real recorded instruments it's not uncommon to use serial compression, e.g. compressing the signal twice, with one doing levelling and the other doing shaping duties. One of these compressors could be a hardware tracking device. The best of both worlds, as long as you know the when and the how. But that's a different story.

And when saying compression in EDM is almost never about loudness, is it because the sounds already are static? Is it because its sample based, and therefore doesn't have variable peaks like in recorded music?

Anything programmed is level controllable in other and usually better ways, such as MIDI velocity.

The same is actually also true for shaping, if you are the producer and not just the mix engineer on an exported project: sample/waveform manipulation, enveloping (ADSR), etc. Compression can be used as yet another tool for shaping the sound, but rarely does it make much sense as a pure levelling tool when it comes to EDM.

Would you prevent using subcompression and mixbusscompression in EDM?

No, compression can be a great tool. The trick is to know when and how to use it.

Would you prevent using subcompression and mixbusscompression in EDM?

No, compression can be a great tool. The trick is to know when and how to use it.

But talking solely about bus- and mixbusscompression in EDM, will it make any sense to use something like The Glue (SSL style buscompressor) on the mixbuss?

I know many people using that one to "glue the mix together" whatever that means.

I mean, if you can level out everything by adjusting the levels on the individual channels, why use a comp on the mixbuss if you want to "glue everything together"?

"Glue" in this case refers to any process that makes the elements of the mix seem less disparate. By adding some overall compression you're forcing what might otherwise sound like independent elements in a mix react to each other.

It's just one small part of the equation, not a magic bullet or a general problem solver. And as mentioned, not always needed or should instead be performed during mastering.

Mix bus compression is often most audible on the parts of the mix that take up a large amount of energy, such as the kick and bass, or on the snare hits. Loud vocals will also make a mix compressor react quite a lot. To avoid pumping, a high pass filter is sometimes applied to the sub frequencies in the detection circuit, or the kick and bass is sum compressed together earlier. External side chain compressing the bass with the kick is another option, and like sum compression, also a quite creative one.

Another mix glue tool is distortion. An effect such as the Sonnox Oxford Inflator is in fact a steady state harmonic saturation effect. By imparting a similar overdriven characteristic to alle elements in the mix, it can be very similar to what happens when all elements have passed through the same analog stage.