Potty Mouth (download)
dirty chorus
Potty Mouth is based on stk::Chorus, but running at 1/4 the host's reported sample rate. It has some features to make it dirtier, mostly involving running the chorus for multiple ticks per sample (either empty or recursively).
Potty Mouth: dirty chorus
-
- KVRAF
- Topic Starter
- 7343 posts since 9 Jan, 2003 from Saint Louis MO
-
- KVRAF
- 8406 posts since 2 Aug, 2005 from Guitar Land, USA
I can't get a clean chorus out of it, Savihost has it pretty high at about 30% cpu.
The only site for experimental amp sim freeware & MIDI FX: http://runbeerrun.blogspot.com
https://m.youtube.com/channel/UCprNcvVH6aPTehLv8J5xokA -Youtube jams
https://m.youtube.com/channel/UCprNcvVH6aPTehLv8J5xokA -Youtube jams
-
- KVRAF
- Topic Starter
- 7343 posts since 9 Jan, 2003 from Saint Louis MO
Hmm, I hadn't noticed any significant CPU hit when I was running it in Maschine.
It probably won't give you a totally clean chorus, but if it's really gnarly when dirt and iterations knobs are both at 0, there's probably a channel issue of some kind.
I'll give it a poke when i get a chance and see if anything unexpected turns up.
It probably won't give you a totally clean chorus, but if it's really gnarly when dirt and iterations knobs are both at 0, there's probably a channel issue of some kind.
I'll give it a poke when i get a chance and see if anything unexpected turns up.
-
- KVRAF
- Topic Starter
- 7343 posts since 9 Jan, 2003 from Saint Louis MO
Hmm, I didn't see anything like that happening. What sort of hardware are you running?
With Maschine 64-bit, I get about 2.25% CPU playing a loop from a few drum samples without Potty Mouth loaded, and about 2.6% with Potty Mouth at default settings. With dirt and iterations maxed (worst case), it runs at about 6%.
With Maschine 32-bit, I get about 2.5% playing a loop, 2.8% with Potty Mouth at default settings, and at worst case it's running about 7.5%.
In FL Studio 11 32-bit, I'm running about 0.75% CPU playing a sliced loop, 0.9% with Potty Mouth at default settings, and 3.6% at worst case settings.
SaviHost was idling at about 0.3% CPU and I couldn't get it to go above about 1.2%
With Maschine 64-bit, I get about 2.25% CPU playing a loop from a few drum samples without Potty Mouth loaded, and about 2.6% with Potty Mouth at default settings. With dirt and iterations maxed (worst case), it runs at about 6%.
With Maschine 32-bit, I get about 2.5% playing a loop, 2.8% with Potty Mouth at default settings, and at worst case it's running about 7.5%.
In FL Studio 11 32-bit, I'm running about 0.75% CPU playing a sliced loop, 0.9% with Potty Mouth at default settings, and 3.6% at worst case settings.
SaviHost was idling at about 0.3% CPU and I couldn't get it to go above about 1.2%
-
- KVRAF
- 8406 posts since 2 Aug, 2005 from Guitar Land, USA
It's the iterations that brings up the cpu. I've only tried it in Savihost so far, running a wav song into it. I'm on a quad core computer.
The only site for experimental amp sim freeware & MIDI FX: http://runbeerrun.blogspot.com
https://m.youtube.com/channel/UCprNcvVH6aPTehLv8J5xokA -Youtube jams
https://m.youtube.com/channel/UCprNcvVH6aPTehLv8J5xokA -Youtube jams
-
- KVRAF
- Topic Starter
- 7343 posts since 9 Jan, 2003 from Saint Louis MO
Iterations increasing the load makes sense. For each sample, here's how many "ticks" of chorus it will do:
dirt 0 - .2: 1 per iteration (1 - 10)
dirt .2 - .4: 2 per iteration (2 - 20)
dirt .4 - .6: 2 per iteration (2 - 20)
dirt .6 - .8: 1 per iteration (1 - 10) plus an average of about .0004 at 120 BPM at 44.1kHz sample rate
dirt .8 - 1: alternates between samples, 1 per iteration then 5 per iteration, so the average is 3-30
I looked at the stk::Chorus source code and tried to optimize that last bit, since it's "fake" ticks that empty out the delay line... but the parts I can remove from the function and still have it sound the way I wanted it to don't really help with performance.
dirt 0 - .2: 1 per iteration (1 - 10)
dirt .2 - .4: 2 per iteration (2 - 20)
dirt .4 - .6: 2 per iteration (2 - 20)
dirt .6 - .8: 1 per iteration (1 - 10) plus an average of about .0004 at 120 BPM at 44.1kHz sample rate
dirt .8 - 1: alternates between samples, 1 per iteration then 5 per iteration, so the average is 3-30
I looked at the stk::Chorus source code and tried to optimize that last bit, since it's "fake" ticks that empty out the delay line... but the parts I can remove from the function and still have it sound the way I wanted it to don't really help with performance.
