Page 1 of 1
Mono vs poly cpu spikes
Posted: Thu Jun 05, 2014 5:03 pm
by Xtinct
I've noticed an odd behaviour in my new synth when loaded into Cubase (cue the collective groan) I play 1 note in "mono" (M to V number of voices 1) I get 10% average cpu and 10% realtime cpu , then I switch it to poly in this case 8 voices but still just play 1 note I get 10% average cpu but a 70% spike in realtime.
Anybody got any ideas whats going on does the M to V primitive act differently when set to 1 voice
Re: Mono vs poly cpu spikes
Posted: Fri Jun 06, 2014 4:43 pm
by KG_is_back
That is truly very weird. So if I get it right, in both cases you play just one single voice - only thing that is different is the max-voices setting on midi to voices?
Re: Mono vs poly cpu spikes
Posted: Fri Jun 06, 2014 7:51 pm
by Xtinct
I tried it with a stripped down version and now both are the same so it must be another part of the synth
Re: Mono vs poly cpu spikes
Posted: Fri Jun 06, 2014 8:34 pm
by KG_is_back
Xtinct wrote:I tried it with a stripped down version and now both are the same so it must be another part of the synth
does the no of voices parameter affects other parts of the synth directly?
Re: Mono vs poly cpu spikes
Posted: Sat Jun 07, 2014 1:27 am
by Xtinct
Only a gain reduction for poly, though some of the processing is volume dependent.
Re: Mono vs poly cpu spikes
Posted: Sat Jun 07, 2014 8:00 am
by KG_is_back
Maybe denormals are happening somewhere.
Re: Mono vs poly cpu spikes
Posted: Thu Jun 12, 2014 3:06 am
by Xtinct
No, no denormals had lots of fun finding and wiping them out, I'm starting not to trust Cubase's new meter because if I load enough instances I can get the real time indicator to constantly hover between 90 and 100% while the average never goes above 20%.
Probably just a small bug or still SM unfriendly but 7 is worth it for the EQ alone.
