What kind of compressor would you put into a delay feedback chain?

MvKeinen

I'm doing this with my Eurorack and there are very nice results. Because it avoids the resonance catastrophe at a point where the sounds get very interesting. Normally it is next to impossible to reach an equilibrium by tuning the feedback level, but if there is a self correcting stage inside the feedback there is much more space to wiggle and with that a lot more sonic variation, especially when the delay time is very short (karplus). I'm using the negated output of an envelope follower that plus an offset that dampens a VCA ( = simple compressor) and also a tan function. https://www.instruomodular.com/product/tanh/

sounds like this: (minute 0:00 - 1:00) and especially 3:45

https://youtu.be/21ORZczjDd4

Any ideas? Do you know of any examples in the user lib with a compressor inside the delay feedback chain.

Thanks in advance!

colB

A couple of old ones that IIRC both use compression. It makes a huge difference!

https://www.native-instruments.com/de/reaktor-community/reaktor-user-library/entry/show/7525/

https://www.native-instruments.com/de/reaktor-community/reaktor-user-library/entry/show/6015/

There is an envelope follower macro somewhere in the factory library that is a good starting point to roll your own (in core pane, hit enter and type Env, there are two versions). There also used to be a good relatively simple compressor (again in a core pane, just hit enter and type comp and it will be there at the bottom of the list).

I find it best to keep the attack very short, like half a millisecond, otherwise you get unwanted build up of transients when the feedback it high. Of course, it also makes sense to get some saturation in there, so some of the compression is from the compressor, and some from the saturator. Then you get nice slow evolution into collapsed layers of soft distortion.

Another thing to consider is what you use for the side chain. How much of the feedback vs the dry input, so more or less ducking. Both options are good but the result is different. Get some EQ in there too, and it can get really nice.

My current delay project has a fade between saturation and compression, HP, LP, and some 'degradation' FX too. Lots of fun :)

MvKeinen

Awesome thanks!

Will check it out now!

And also thanks for countless little and big helpers throughout the years @colB

Studiowaves

Generally if you put a compressor in the feedback loop you end up with an expander on the output.

colB

A compressor/expander setup is used to help mitigate noise in things like BBD and tape delays.

This is different, here the compressor is doing the job of a limiter to tame or kill exponential amplitude growth into nasty distortion with high feedback settings. That does not require a matching expander.

MvKeinen

I had a look into this one. It seems perfect as a starting point.

but I wasn't able to find the thread that was linked in the comments. Tried wayback machine, but it didn't work. Lets see how far I'll get

https://www.native-instruments.com/de/reaktor-community/reaktor-user-library/entry/show/11811/

colB

https://community.native-instruments.com/discussion/comment/67984#Comment_67984

Hmm, That one is very simple and is actually more appropriate for what Studiowaves is talking about. It only gives a moderate compression, ideal to compare with an expander for simulating the noise reduction companding in older FX units. Maybe not so useful for feedback limiting. It might still work ok though with the right settings.

It depends what you are going for. If you want something that makes it easier to find the sweet spot on the feedback dial, but doesnt have to completely squash the potential howling feedback at the extremes, then that will maybe work ok, but if you want to really limit the amplitude to get total control, maybe that won't be enough. The factory one should work pretty good though: