SMC Joono

[jorge]

1000 ? Measured in which units ? clock cycles?
Does this means than anything that needs more than 1000 clock cycles is unusable?

whit a bit more of information should be easy to code a light cpu meter....

I hope the following answers your questions:

http://www.synthmaker.com/docs/uploads/2005/may/22/cpu.osm

[Nu Audio Science]

Jorge i'm a little confused but if it is SSE why is 1 note differing from 4 ?

And thanx for the great schematic

[Nu Audio Science]

Unk you say to use Stefs SVF for the filters so i have a couple of Qs

1 Should i try and clean the HP BP Notch Peak out of the code for the LP components or will it have no effect on CPU ?

2 The HP in the Junoseems to be non resonant so should i just set the RES at .5 and leave or try and create some non res code ?

[Nu Audio Science]

http://uk.geocities.com/nas_sound/Joono.zip

Update with Unks OSCs and Stefs SVFs

Unk does that aftertouch module you made replace the MIDI2poly completly ?

Has anybody had any thoughts on the 8 segment env yet ? , I'm guessing code component ?

[infuzion]

I hope the following answers your questions:http://www.synthmaker.com/docs/uploads/2005/may/22/cpu.osm

Very neat, thanks jorge!

[unkargherth]

1 Should i try and clean the HP BP Notch Peak out of the code for the LP components or will it have no effect on CPU ?

No, no. Simply use waht you want but don't touch the code

2 The HP in the Junoseems to be non resonant so should i just set the RES at .5 and leave or try and create some non res code ?

Set res to something that didn't makeup the sound

Unk does that aftertouch module you made replace the MIDI2poly completly ?

MIDI ot Poly, AFAIK, has no support for After touch. Both of them should be linked to to A MIDI Input. What basically mine do i't's to support "Channel Presure" when the keyboqard doen's support Key AfterTouch. Nothing More

Has anybody had any thoughts on the 8 segment env yet ? , I'm guessing code component ?

Same Answer as for the "mono Envelope" issuse. Nosense to doit until we can Trigger it

[jorge]

Jorge i'm a little confused but if it is SSE why is 1 note differing from 4 ?

Just imagine code that has only one instruction taking one clock cycle like Add or Mult. If such code would run inside the analyser, it would process four notes simultaneously. The analyser would say that the code took 1 clock cycle per sample but in average each note would take 1/4 of a clock cycle per sample.

[Shifrin]

Yep jorge, that cpu schematic is very useful

Cheers!

Shif.

[unkargherth]

Well. Using Jorge's schematic i tried i've done some quick test for the oscilators

On my machine ( P IV 2.4 Ghz) it gives something like 2.52% for one note
if i change the cpu speed to the correct 2.4 *10^9 then gives me something like 0.72% ( talking from memory)

1.- The increase is a bit high fro only a 30% cpu hertzs. Ha something to do that my machine supports HyperThreading?

2.- Worst case 2.52% per note means like 40% for polyphony 16. Or should i divide by four to take into account the SSE based processing? Anyway, is this too much or is reasonably for this fairly complex osc's?

[infuzion]

On my machine ( P IV 2.4 Ghz) it gives something like 2.52% for one note

Ummm, looking at Jorge's schematic, it looks like the "CPU speed" is hard-coded at 160000000. Did you change that value, or...?

[unkargherth]

Sorry. Corrections

For the Standard Hard code 1.6 Ghz cpu results is 3.9% - 4.1%
For MY CPU ( correcting the hardcoded to 2.4Ghz) result is 2.6-2.8%

( I've missed to connect the PWM pin so PWM hasn't been working)

[jorge]

For the Standard Hard code 1.6 Ghz cpu results is 3.9% - 4.1%
For MY CPU ( correcting the hardcoded to 2.4Ghz) result is 2.6-2.8%

There's always the possibility that the calculations in my schematic are wrong. It would be interesting to compare it with actual CPU useage as measured in other applications.
Maybe there's other things to take in consideration like for example my Pentium M at 1.6GHz is actually quite efficient even with a low clock frequency. My Pentium 4 with much higher clock frequency achieves roughly the same CPU useage. Although clock cycles doesnt say everything, it is still my prefered way of measuring the efficiency of a schematic.

[infuzion]

For the Standard Hard code 1.6 Ghz cpu results is 3.9% - 4.1%
For MY CPU ( correcting the hardcoded to 2.4Ghz) result is 2.6-2.8%

There's always the possibility that the calculations in my schematic are wrong. It would be interesting to compare it with actual CPU useage as measured in other applications.

Buzz has a little CPU-usage-for-each-machine popup that I use. I'm sure there's overhead for the VST wrapper, but good to compare VSTs with others.

[Nu Audio Science]

Having to wrok on an SE project at themoment and getting annoyed so i'm gonna update this a bit today

[Nu Audio Science]

OSM Here

Only a small update so it can be saved as a VSTi for messin around, I'm hoping to get more time on this this evening because it is sounding like it has a lot of potential
The envelope is crap though and doesn't work anything like an envelope should