virus TI internal specs - Page 2 - The Unofficial Access Virus & Virus TI Forum

nutekk · #11 03.11.2005, 09:53 PM

yea tk from access said its internal rate is 24/48 -depending on if your working 44 or 48

sounds gorgeous to my ears!

Zephod · #12 04.11.2005, 12:17 AM

The 56 bits size is only available as a special accumulator register on the motorola dsp series they are using.. it makes sure no bits are lost during long calculations. After the calculation some bits will have to be striped away again to fit the result in a normal register.. which is 24 bits.

Contrast this to 32bit math on pc which uses 64bit or even 80 bit accumulators inside the cpu.. 56 is nothing to go wild about

on the samplerate:

44/48 is probably right for -some- parts of the synth, but it wouldnt surprise me if some parts (like the filters) are oversampled to 88/96 to keep the resonance stable at the high end.

dr. orange · #13 04.11.2005, 04:40 PM

Quote:

Originally Posted by Zephod

44/48 is probably right for -some- parts of the synth, but it wouldnt surprise me if some parts (like the filters) are oversampled to 88/96 to keep the resonance stable at the high end.

funny, oversampled filters but horrible aliasing on higher notes... that doesn't make any sense

nutekk · #14 04.11.2005, 05:06 PM

im just not hearing anything horrible out of the TI.
some instruments actually use alaising as a defining point...microwave xt

the TI is at the top of the food chain in VA.
it sounds just as good and better then the previous viri..

MADSTATION · #15 04.11.2005, 06:07 PM

dr. orange: I understand what you mean regarding aliasing but most VA and vsts have the same issues(especially Z3TA+).

Virus might be far from the best in the high range but still an awesome all around synth. Great in many areas but not the best in every type of sound.

Zephod · #16 04.11.2005, 07:14 PM

oversampling just the filter has a lot less impact on cpu than oversampling 16 oscillators... but it has a lot more impact on soundquality in the normal playing range.

I just checked the basic osc.. it indeed does produce quite a bit of aliasing... (*frowns at access* There is no need for aliasing like this anymore!)

but.. until we see the code responsible for the sound, its all guesswork..
I was just making some observations to illustrate "samplerate" and "bitdepth" not having to be uniform across a synth.

dr. orange · #17 04.11.2005, 07:28 PM

DEAR PALS!!!

despite the horrible horrible aliasing which indeed is very horrible and a pain in the ass having it at a synth like the virus, I just have preordered it!

nutekk · #18 04.11.2005, 07:30 PM

MADSTATION · #19 04.11.2005, 07:32 PM

Thanks Zephod!
Your posts were quite informative

Do you mean that aliasing might be related to the fact that the TI doesn't have enough CPU to render the osc at higher samplerate?

What about Hypersaw. Is this aliasing free?

Dr. orange: Wow I'm actually surprised

Have fun tweaking your virus!(ti?)

Zephod · #20 04.11.2005, 08:29 PM

Quote:

Originally Posted by MADSTATION

Do you mean that aliasing might be related to the fact that the TI doesn't have enough CPU to render the osc at higher samplerate?

What about Hypersaw. Is this aliasing free?

yeah.. the cpu (I guess the DSP can be called central enough in this box?) might be a limiting factor here.. or memory. A common strategy to prevent aliasing is to store pre-bandlimited wavetables for every octave and then oversample 2x. Or store a new wavetable for every harmonic that drops from the sound.

I'll have to check the hypersaw later tonight. Its probably harder to see the aliasing with the dense spectrum caused by the detune, but I'll see what I can do.