Log in

View Full Version : Access Virus TI reduced aliasing?


DIGITAL SCREAMS
09.11.2004, 04:15 PM
On another forum someone said to me that aliasing in the TI model will be less audiable because its using 192khz dacs. Can someone confirm this?

Cheers

DS

WorkHard
09.11.2004, 04:26 PM
I dont think that has anything to do with aliasing, but I could be wrong.

I mean, that wouldn't explain why softsynth's alias.

Juho L
09.11.2004, 05:07 PM
Higher resolution makes smoother "blending" possible. So I think it reduces aliasing.

DIGITAL SCREAMS
09.11.2004, 06:50 PM
This is the original post...

'So, the new Virus TI, which uses 192Khz DACs will have less 'audible' aliasing than the old Virus B/C, as the nyquest frequency is now 96Khz. so you need to pick waveforms with harmonics higher than 96khz to get aliasing. So if you go for a waveform with a 100Khz harmonic, you hear a 92Khz harmonic (or rather, don't hear it). But this does mean an increase in processing power (remember you have samples to fill and at 192Khz, a *LOT* more samples to fill)'

So Juho does that make sence to you?

DS

Juho L
09.11.2004, 07:58 PM
So Juho does that make sence to you?

Yup. As I said, more resolution -> Smoother blending -> higher frequencies.

DIGITAL SCREAMS
09.11.2004, 09:33 PM
Hurrayyyy..... :D :D :D

DS

jasedee
10.11.2004, 07:52 AM
The thing is, depending on how steep the filters are to stop frequencies above the Nyquist, we get a phenomena where a "ripple-back" effect takes place, causing some 'distortions'.....

Aparently, 'they' are now saying you can go too far with sample rates, and we should stick to 96kHz!!!

Which is real bad for me, seeing as my site name is 192k24bit.com :(

If you use more gentle slope filters that cut gradually, this effect is minimised. So if we start the roll-off from 44.1k all the way to 96k, it will be better??? So, the question is:

"How good/gentle are the anti-aliasing filters on the Ti??"

unease4u
10.11.2004, 01:05 PM
I'd be really surprised if the internal sample rate was upped to 192kHz. My guess is that the internal sample rate of the sound engine is the same as before, just that the converters can support 192kHz.

Juho L
10.11.2004, 01:12 PM
I'd be really surprised if the internal sample rate was upped to 192kHz. My guess is that the internal sample rate of the sound engine is the same as before, just that the converters can support 192kHz.

And there's no sense in that. It would be waste. The internal processing has to also be 192kHz or the high resolution DA is useless because on lower sample rates than 192kHz it would just be idle on most time in form of transforming data that has already been transformed.

unease4u
10.11.2004, 03:06 PM
Well, I'd love to be proven wrong but running the audio engine at 192kHz would take 4 times the processing power compared to 48kHz.

Juho L
10.11.2004, 03:15 PM
Well, I'd love to be proven wrong but running the audio engine at 192kHz would take 4 times the processing power compared to 48kHz.

Indeed it does, but I don't see a reason why they would use 192kHz DAC's if the original data is 96kHz.

The DSP's on the TI are way more powerful than on prrevious Viruses and there's two of them so raw CPU power shouldn't be an issue.

Gopal
10.11.2004, 10:02 PM
I think you'll find that 192 requires exponentially greater power than 44.1 rather than just 4 times. as 24 bit is exponentially greater than 16 bit. [/quote]

Juho L
10.11.2004, 10:22 PM
I think you'll find that 192 requires exponentially greater power than 44.1 rather than just 4 times. as 24 bit is exponentially greater than 16 bit.

Increasing sampling frequency only means that the calculations have to be done faster. The "sampling frequency/processing power needed" relation is linear, which means that 192kHz requires twice as fast processing than 96kHz.

The bitdepth is different issue. On bitdepth the main issues are the memory and bus structure of the processor. For example 16-bit system needs to do extra cycles in addition to increased memory consumption when moving to 24-bit system because it can't handle the data at once (has to be processed in two or several parts: The least significant 16-bit part and the remaining 8-bit most significant part. When doing calculations this way the processing can get really "non-economic" and we have a bottleneck.

But all this technical speculation is quite trivial on this whole issue.

grs
15.11.2004, 01:12 PM
I'd like to know if you can run the virus TI drivers (asio?) at any given rate ie 44.1, 48, 96khz etc.
Given that assumtion, would the virus still keep making all its calculations for its synth sounds at 192khz if you had set you sequencer at 44.1khz?
Also if the answer is yes, what happens when it sends to the vsti/usb busses. Downsampling?
That would be what I think would happen. It would sound worse if the virus produced sounds at 44.1khz IMO.
Basically all that Intergration needs way more explanation than the pretty brocures!
I'll be happy if there is a manual floating around for the TI, anyone seen it?

blay
15.11.2004, 01:25 PM
no manual - prob just what youve seen already.
theres some pretty 3d animations on their site now :roll:
guess youll have to sit tight with the rest of us until access gives us some more specs to drool over :wink:

grs
15.11.2004, 01:33 PM
can you give me the link, I havn't seen them there?

blay
15.11.2004, 01:54 PM
www.access-music.de/3d/virus_ti_keyboard.html

the check the relevant items page and look in the column menu :wink: