KVR Audio

orangesuite · Post by **orangesuite** » Wed May 31, 2006 11:22 am

"I once got my CPU meter to 25 % in Cubase. But then I had 8 instances playing 128 instruments simultaneously. I could probably optimize it a bit more, but right now, CPU usage is the least of my worries."

bullshit

benjamind · Post by **benjamind** » Wed May 31, 2006 11:27 am

That bit really got me thinking. 128 synths and only 25% CPU. My God, what in the f***ing hell does this guy have...a secret AMD Unicorn Magic Triple Quadcore 64Ghz with 64MB of on-die cache!!!

*Cough cough* ehem...I expect to see the thread sorta die down and maybe even the thread on Northern sounds will sort of die down too. I thought this might have been true, but uhm, I'm starting to think a bit about this. I think this little CPU usage bit was a kinda giveaway. Yeah...I'd say bullshit too...unless his CPU is like the one I just mentioned.

orangesuite · Post by **orangesuite** » Wed May 31, 2006 11:32 am

lets burn him

ObsoleteAcc99 · Post by **ObsoleteAcc99** » Wed May 31, 2006 11:38 am

[DELETED]

.jon · Post by **.jon** » Wed May 31, 2006 11:42 am

orangesuite wrote:"I once got my CPU meter to 25 % in Cubase. But then I had 8 instances playing 128 instruments simultaneously. I could probably optimize it a bit more, but right now, CPU usage is the least of my worries."

bullshit

Originally Posted by JT3_Jon
I can field this one. What arwa is doing is creating synth waveforms based on real instruments. As a fellow WX5 owner, I'm assuming you also own a VL series tone generator. Playing a VL tone generator is the same concept, for when you play the sound, you are not playing back a recording (sample) but instead manipulating a waveform BASED on the instrument you are playing. This allows for much more control and expression in the instrument, as you are now basically playing an instrument, rather than triggering recordings.

I hope that makes sense. I'm not the best at explaining technology

Correct. What makes the synth stand out is that I've done some sophisticated modeling of the behaviour of the instruments. The sounds are NOT generated through Physical Modelling (wave equation) but the instrumental behaviour have been modelled using physical equations. The tone and behaviour (and performance) of true Physical Models just can't compete with sample technology (yet). It's basically just a question of what level you choose to model an instrument. Samples are high-level, Physical Modelling is low-level, I'm in between.

low CPU does not sound bullshit if the above is true... this synth might actually be real IMO.

orangesuite · Post by **orangesuite** » Wed May 31, 2006 11:48 am

c'mon, there is no way (at the moment) that you can get a synth that sounds as good as the demos and have as many instances running.

.jon · Post by **.jon** » Wed May 31, 2006 11:57 am

Did you read the above?

It's not PM, he's only calculating real-time the performance behaviors and manipulating a waveform according to them.

I'd think that even basic substractive synthesis takes more resources, and I can run something like 30 xhips on my humble centrino notebook.

[EDIT: With asio4all. With a strong AMD64 and a proper interface that really does not sound impossible at all.]

Yossarian · Post by **Yossarian** » Wed May 31, 2006 11:58 am

The same guy posted about this orchestral synth back in January with an mp3 demo. I remember thinking that it was interesting and showed great potential but the sound wasn't quite there yet. To my ears he's made quite a bit of progress since then.

I definitely think this is for real. If it's a hoax, he's being extremely patient and elaborate about it - and for what? It just doesn't make any sense to me.

I believe this guy has made some genuine innovations here and the results are very impressive. My only (very minor) gripe is with the vibrato but it seems this is highly tweakable. It also appears to be a more mature technology than e.g. Synful Orchestra (which still has a lot of potential). If (when?) strings are included, this will be a very powerful orchestral package.

/Yoss

whyterabbyt · Post by **whyterabbyt** » Wed May 31, 2006 12:02 pm

orangesuite wrote:c'mon, there is no way (at the moment) that you can get a synth that sounds as good as the demos and have as many instances running.

Says who?

orangesuite · Post by **orangesuite** » Wed May 31, 2006 12:05 pm

me. actually there's no point in discussing this as we dont have anything to try out. hijack over

snooky · Post by **snooky** » Wed May 31, 2006 12:05 pm

this looks promising

djsubject · Post by **djsubject** » Wed May 31, 2006 12:23 pm

sounds good, speck looks good, wondering if its real aswell

Subz

Yossarian · Post by **Yossarian** » Wed May 31, 2006 12:48 pm

I did a bit of sleuthing and it turns out this guy is a student at the Royal Institute of Technology (KTH) in Stockholm, Physics dept. His masters thesis was in music acoustics and two weeks ago he held a seminar at the institute entitled "A multi-dimensional interpolation scheme for improved control of sample-based synthesis".

I think it's pretty safe to say this guy is the genuine article.

Of course, there's also the possibility that I'm part of the scam.

/Yoss

Meffy · Post by **Meffy** » Wed May 31, 2006 12:55 pm

So tell us, orangesuite, what software have you written? What experience have you in this field that gives you such deep insight into a program you've never even seen? I'm assuming from your tone that your credentials are far more impressive than a masters in musical acoustics from RIT, Stockholm, and you've a whole lot of real-world, professional programming experience under your belt.

Yossarian · Post by **Yossarian** » Wed May 31, 2006 2:38 pm

It would be very cool to hear what a combination of this and Synful would sound like. Synful is great for strings but a bit weaker for brass and woodwinds so they seem to be a good match.

/Yoss

What the !!! Has anyone even seen this!!??