pdxindy wrote: ↑
Sun Nov 28, 2021 8:47 am
WasteLand wrote: ↑
Sun Nov 28, 2021 8:28 am
pdxindy wrote: ↑
Sun Nov 28, 2021 6:19 am
masterhiggins wrote: ↑
Sat Nov 27, 2021 8:32 pm
Can't wait to be thoroughly disappointed by the crippling CPU usage.
I stopped using Aalto and Kaivo because of the cpu use. More specifically, that they use full cpu whether they are playing notes or not so they use up lots of cpu when idle.
They are only practical for me to record straight to audio and then deactivate the plugin to save the cpu.
i think it is quite system dependent, not very low cpu/sample buffer, but as expected from these kind of synths, even with MPE. and very managable. but 1 system very new, second system 2 years old... so...
If you have an instance set to 4 voices, it uses the full CPU of 4 playing voices even when it is not playing any voices. So if I have a 6 minute composition, and a Kaivo part that lasts only 30 seconds, it still uses the full cpu during the entire 6 minutes. Quite inefficient!
i don't want to go into a debate, they are semi-modular, and i think they work even on sample rate, i assume it, and i think i even asked it. but my memory. assume: well that is the nature of modular...
but i get mixed results; system 1; reaper 5 instances of Kaivo kills it. no MPE.
system 2; bitwig, 11 instances, MPE, heavy modulation. all tracks armed, 8 voice. no stress...
on system 1, the double buffer of the soundinterface kicked in, for the first time i saw that i cán work, mostly it doesn't... the double buffer...
on system 2, no double buffer, but extreme low sample buffer load. for MPE, and these kind of synths. extreme low, it is an overstatement, but i could add more...
i can kill system 2 with Pigments 3, heavily modulated, MPE, with Bitwig, with only 4 instances...
i thought that on system 1 it could perform better, perhaps i sold the wrong soundinterface... well no prob, system 1 is meant for sketching, although i make also tracks with it.
system 2, well, it is my desktop/studio built, so, it is built for performance, also the soundinterface seems to influence the 'score'.
i use on both system a sample buffer of 512 samples, on system 2, i went to 256 samples, the result in percentage; sample buffer fill was the same.
is this interesting? it seems that Kaivo/Aalto indeed draw a lot, but it also depends on the system you have.
all done under Windows 10, by the way.
Reaper of course, all tracks armed, why choose bitwig, on system 2? well bitwig isn't that optimized, or it is.. it depends on the plugin..
EDIT: i only want to supply more info, although i don't share the specs of the system, perhaps worth it?
on my previous studio desktop, i had also no problems, with Kaivo. Aalto i just recently, some months, work with it.
but your remark is quite usefull; and you can demo Madronalab synths, without time restrictions (but with noise...), it seems quite worthfull to check if your system can handle them in way that is comfortable.
and perhaps, but that is question, for the developer, Randy, there are perhaps some optimizations necessary. the V2 versions are coming. well, also later...