VSTis as good as hardware? Similar to Access Virus Quality?

VST, AU, AAX, CLAP, etc. Plugin Virtual Instruments Discussion
Post Reply New Topic
RELATED
PRODUCTS

Post

there's no such thing as zero latency, so if you mean unnoticable latency, obviously with today's 2ms to 10ms ASIO latencies

Post

Its indeed offcourse only 80voices total, but try running for instance a capable comparable synth (routings and options wise) like z3ta+ at 96khz (2x oversampling + high mode?) And then a nice patch, and then try pressing 80 keys, and you won't even have the unison option. I have a fast computer myself, but I still think its either just not efficiently programmed enough, or cause of all the extra things computers run, that its not possible. I could be very wrong offcourse ;)

Post

ericj23 wrote:are you sure your pc can do this ???
Probably. The dsp-chips they use are not that great at all. As aciddose says, they probably use some effing great dirty tricks to get stuff done quickly. I know some peeps at Access and all they say is that the guy has had some brilliant ideas to do thinks cheap and easy.

Btw. forget about 0 latency. No DAC nor any midi connection can do this, and if they mean the digital connections via usb whatsoever, it's just as 0 latency as any other plugin (that is, 0 latency when bouncing to disc, while using the soundcard's latency for monitoring).

;) Urs

Post

tony tony chopper wrote:there's no such thing as zero latency, so if you mean unnoticable latency, obviously with today's 2ms to 10ms ASIO latencies
i can notice 10ms latency - the virus runs at the 2ms latency end fo the spectrum - so again folks 80 voices with 70 fx at 24/96 at 2ms latency

are you sure your pc can do that ???
I believe every thread should devolve into character attacks and witch-burning. It really helps the discussion.

Post

ericj23 wrote:
tony tony chopper wrote:there's no such thing as zero latency, so if you mean unnoticable latency, obviously with today's 2ms to 10ms ASIO latencies
i can notice 10ms latency - the virus runs at the 2ms latency end fo the spectrum - so again folks 80 voices with 70 fx at 24/96 at 2ms latency

are you sure your pc can do that ???
yes!!

shut up and go play with your crapass virus - let the big boys talk now...

Post

Here's what they have to say:
HyperSaw is a multi-sawtooth oscillator. With a single parameter the user can alter the amount of parallel oscillators from 1-9, smoothly and in realtime. You can literally add more oscillators to fatten up the sound whilst playing a chord. The relative tuning and spread can be controlled as well. Using the highly efficient HyperSaw engine with 9 parallel oscillators per voice (each with it's own parallel sub oscillator) the VIRUS TI series can calculate over 1800 oscillators in realtime!
That's pretty good. My synth won't spit out more than 1400 oscillators in the best case at 44khz on my AMD 3500.
Of course it's only for oscillators.

Edit: ok I can top it with 1890 osc realtime on my AMD3500 if I cheat a bit (all osc in tune & in phase to work better in the cache :)

Post

vvanrij wrote:Hmm, would it really be possible to run a Virus TI on a modern PC? I mean if you sum up the specs, I think it would get a modern pc on its knee's. 16multi, 80poly, the high quality sound, all the effects..
Yes it would, even the new T1 runs on old chips, nothing that powerful really. With lots of stuff (mod and effects) switched on the T1 is supposed to give upto 40 polyphony with about 4 parts (according to SOS review). It isnt anything above many softsynths in terms of pure audio quality either.

I remember that a few years ago some effects software was done as a native version that was usually run on something that uses the motorola chips - heres the link well worth reading

http://sound-on-sound2.infopop.net/2/Op ... 5653031655
As I said, a single Athlon 1800xp can run 23 Waves Linear EQs. A £2500 Digi HD card with NINE 100mhz top-of-the-range Motorala DSPs can run 9. A TC Powercore (which has only 4 of these same DSPs) would handle 4. And the algorithms are IDENTICAL
this is from june 2002, computers now are several times more powerful than this example. Maybe the DSP versions werent that well optimised but it still gives a reasonable indication for comparison. When you buy most hardware you are not paying the extra for some all mighty DSP performance.

Post

>>the virus runs at the 2ms latency end fo the spectrum

ok but it's useless if you don't.. sequence it. Your sequencer will introduce more latency, and even more if you stream the audio back, than if it was a VSTi right in the sequencer.

So it's good when you play it live but.. do you press keys with a 2ms latency?

Post

ericj23 wrote:
tony tony chopper wrote:there's no such thing as zero latency, so if you mean unnoticable latency, obviously with today's 2ms to 10ms ASIO latencies
i can notice 10ms latency - the virus runs at the 2ms latency end fo the spectrum - so again folks 80 voices with 70 fx at 24/96 at 2ms latency

are you sure your pc can do that ???
the voice count is dynamic depending on complexity, it doesnt run at 96k internally and 2ms (do you even know for a fact that the Virus does 2ms) latency is possible but really overkill.

Post

The Virus has high-resolution DACs, but internally it still runs at 44khz.

Post

its all about latency folks - the waves linear phase example from sound on sound - yes your hd card can only run 9 - but it runs them at as close to zero latency so that you can track through them

set your nice asio card at 1 or 2 ms latency and then see how many waves eq's you can track through
I believe every thread should devolve into character attacks and witch-burning. It really helps the discussion.

Post

Still, 1800 osc realtime is pretty good and matches today's PC's.

I can't do more than that in Sytrus with linear interpolation, maybe without any interpolation and ramping.

Wait, gonna try with the 3xosc, because Sytrus is using double float phases, and the integer conversion is costy.
Is the Virus using 24bit integer? That's more doable. 3xosc uses 32bit int phase, but only produces stereo osc (not interpolated and not bandlimited).. let me check..
nope, the 3xosc will produce around 1800 osc too, but stereo this time. Something tells me the limit is more the mem accesses than the operations.

Post

Keep in mind you're only going to get near that 1800 figure if you use strictly the Hypersaw algorithm, which means the waveforms can only be saws or squares (or anywhere in between.)

Post

only synth that I know of that might be capable of that is Synth1 :D .. or Subtractor

Fots

Post

Thank you projectdan, your post explains practicly everything :D I thought the virus would have about 80voices no matter what, but going to 40 at 4multi, well thats not what I suspected :) My own hardware synths don't lose voices (unless unison offcourse), but this again proves how much the virus is just a vst in a box :) Anybody knows much about vaz modular? How does it compare up to Reaktor, Nord G2?

-Victor

Post Reply

Return to “Instruments”