I’m brand spanking new to synths in general and VCV in specific so bear with me. I’ve been looking at a bunch of things online (here, youtube, elsewhere) on generative music and patched this thing together:
I like the sound and the result. The thing is that it’s pretty, eh, cluttered. And I read here that for every thing you want to do, there’s five different solutions with up/downsides.
Can anyone shed some insight in what I could have done easier/better/smarter/… here?
Fat to cut, modules that can do on their own what I amateurishly tossed together using half a dozen others?
Looks like I’m too new to upload my rack file yet. Hopefully soon.
I wouldn’t change much here. One obvious thing is that you might need a stereo mixer (there are lots of them, choose whichever you like). And that’s pretty much all… Unless you want to experiment with polyphonic cables. That would make it a bit less cluttered, but maybe more confusing in a way, like… If you open this patch in a month, you would have to think hard about all the layers of polyphonic signals and events.
just to explain, if you are not familiar with polyphony in VCV - that’s not about the sound, that’s about bundling the signals like 2-16 thin cables or layers – into one thick cable. Like, for example, you can use only one instance of VCO, but with polyphonic CV operating it. So it would look less cluttered, but in reality, that doesn’t really change your patch (you would still have 3 VCOs, I don’t really remember if it is easier on the CPU, but that could be another reason to use polyphony. I rarely use it though), hope this makes sense.
Like this, the result would be pretty much the same for both options:
That’s also something I was curious about: Rack uses a lot of CPU. As in, I’ve given it 8 cores out of 16 (well, rather 8 threads; 8 physical cores with HT) and it uses them pretty much all. But the odd thing is:
It uses that much with ZERO modules.
Really surprised about that. Open Rack, delete everything, 8 HT cpus running 95-100%. Open my patch above which has quite some things happening - hardly any change. Wondering if that’s normal, or something on my end.
That’s something I don’t really mess with. I might be wrong, but I think VCV isn’t really this efficient when it comes to multi-core stuff. i think it’s better to use one thread and if needed, you change it… I don’t think i ever used more than 3 threads… Because the higher you go, the less optimized it gets, it feels like. So try the empty patch and change the threads to 1 (to check if it goes to unreasonable numbers). That should lower the CPU usage, I think. But I am not this smart with CPUs and all the tech things. There are much smarter people here that might help you, hopefully… Or try searching the forum, I don’t know what you should search for… maybe “high CPU usage” or something like that.
So yeah, I think that’s both on your end and the VCV fault. You changed it to the highest setting, which is not optimal, because VCV works better on 1-3 threads, in my experience.
*also try lowering the FPS (View - Frame Rate), it also affects the CPU… I only work on 10 FPS, haha. I don’t think you need 60 FPS. i think 30 FPS tops would be comfortable and reasonable. Maybe even 20. Also maybe lower a Sample Rate, if it is too high… i think 48kHz is great. But you can use 44.1kHz, that would also be fine, I think
Ha! That was unexpected. I lowered it to one, and the one process uses not even 10% of one CPU core. Compared to 8 processes, each eating 100% of a core. That’ll teach me to fiddle with defaults without a proper reason.
Yes. The VCV VCO will use the same cpu to do 4 voice polyphony as 1, so poly would save cpu. This is true of many modules from many devs. But not all of them.
There is more (much more!) over in my demo repo. This article in particular:
here
Sometimes “cluttered” is good: When you use a basic single-function module for every function in a patch, the patch will usually be easier to understand.
Once we start getting tricky and making things more elegant, it can sometimes become far more difficult to decipher and fix.
Surprisingly difficult to get a credible siren sound! I put a C and E tone saw with some minor pitch modulation over a long envelope and added some filtering. I’m not entirely unhappy with the result but it’s still very obviously off.
