MeldaProduction

Satch1 · Post by **Satch1** » Fri Aug 17, 2012 11:44 pm

Hi Vojtech,

I've just discovered that the longer the syncronisation length is in the feedback processor, the higher the cpu hit, and when you get up to 32 and above, it goes through the roof. At 64 using your feedback demo preset, my cpu is reading 82%. Is this normal behaviour, or a bug? I don't remember it being like this before. I thought the shorter the feedback delay was, the more cpu it used. Is that correct?

farfadetfarfelu · Post by **farfadetfarfelu** » Sat Aug 18, 2012 12:47 am

It works fine here, less CPU with bigger feedback delays.

I've noticed weird behaviour with the feedback module in the past like cpu staying high even after a feedback module is removed... i did not encounter this recently though.

Maybe if you restart MXXX it will come back to normal.

Satch1 · Post by **Satch1** » Sat Aug 18, 2012 1:34 am

Hi mate, thanks for replying. I've restarted my computer but I've still got the same problem. In Reaper, if I create a single track with MXXX on it, load the feedback preset, and put the syncronisation to 64 bars, I'm getting a 56% CPU reading in the track fx window, and 26% on the main CPU display. If I do the same in FL Studio 10, I can only get up to 16 bar sync before the CPU bar is maxed. Not good! I'm on Win XP with a quad CPU. Are you Win7?

farfadetfarfelu · Post by **farfadetfarfelu** » Sat Aug 18, 2012 2:07 am

Wow actually you are right, i was not trying with the sync ON.
Reaper and win7, i have the same problem as you, very high cpu on 32 and 64 length !

Satch1 · Post by **Satch1** » Sat Aug 18, 2012 8:16 am

Ok, thanks for testing and confirming. At least my system's not knackered then. Over to you Vojtech... sorry!

MeldaProduction · Post by **MeldaProduction** » Sat Aug 18, 2012 10:30 am

It's the feedback delay implementation. I'll check it out. Never though anyone would do anything like this

.

MeldaProduction · Post by **MeldaProduction** » Sat Aug 18, 2012 10:31 am

Btw. do you know how much time it is right? In 120bpm 1/4 is 0.5s, so 1 is 2s and 64 is then 128s, more than 2 minutes delay!!!

Satch1 · Post by **Satch1** » Sat Aug 18, 2012 12:24 pm

Yes I realise that, and I'd probably never use such a big delay, but I first noticed this behaviour when I had some processors inserted within the feedback loop, and consequently I could only get up to 1/2 bar sync before my CPU maxed. Obviously having processors in the feedback loop is going to be a lot more demanding, so I tried it without, and was able to use a much larger delay. But the point is that CPU is increasing with longer delay times when I thought it was supposed to be the opposite. Is this correct?

MeldaProduction · Post by **MeldaProduction** » Sat Aug 18, 2012 2:24 pm

True, I'll fix that hopefully

Satch1 · Post by **Satch1** » Sat Aug 18, 2012 3:46 pm

Good man, thanks. I hope you can fix it without too much of a problem.

MeldaProduction

KVR Audio

[MXXX] Feedback CPU - (Fixed)