A 192Khz PM DSP unit dedicated to brass/wind/reed modelling?
-
- KVRAF
- 5520 posts since 6 May, 2002
What if someone released a Hardware DSP Unit that was dedicated to 16 realtime Physically Modelled Orchestral (Brass/Wind) voices running at 192Khz (Sort of like a polyphonic VL-1)? It could connect to our PCs via firewire/and or USB2 and allow us to run the Orchestral VSTI like a DSP cards plugins. Would you buy it and how much would you pay for those 16 voices of polyphony at 192Khz?
Intel Core2 Quad CPU + 4 GIG RAM
-
- KVRist
- 133 posts since 9 Apr, 2004
Not nearly enough polyphony is it? Then again, it probably wouldn't need the polyphony of Gigastudio if the samples are modelled, right? All articulations/releases for an instrument, even in stereo, would be one voice (?)
Also, who needs 192hKz sample rate?
If it were "all that" and could sound as realistic as anything out there (best of VSL Pro/Synful orchestra/EWQLSO Platinum/et al), and not take any of my puter's CPU, HD, or RAM that would really be something. I suspect people out there would pay a lot more for it than I could possibly afford.
fizbin
Also, who needs 192hKz sample rate?
If it were "all that" and could sound as realistic as anything out there (best of VSL Pro/Synful orchestra/EWQLSO Platinum/et al), and not take any of my puter's CPU, HD, or RAM that would really be something. I suspect people out there would pay a lot more for it than I could possibly afford.
fizbin
-
- KVRAF
- Topic Starter
- 5520 posts since 6 May, 2002
I am not talking about sampling at all. This would be a 100% Physical Modeling synth like the Yamaha VL-1.fizbin wrote:Not nearly enough polyphony is it? Then again, it probably wouldn't need the polyphony of Gigastudio if the samples are modelled, right? All articulations/releases for an instrument, even in stereo, would be one voice (?)
Also, who needs 192hKz sample rate?
Any unit being designed in 2005 had better be able to run at 192Khz. The UAD does and that is a big plus over competitors. As always lower samplerate means more voices. I think that 16 channels of 192Khz audio is probably hefty for 1394b.
Intel Core2 Quad CPU + 4 GIG RAM
-
- KVRAF
- 8079 posts since 9 Jan, 2003 from Saint Louis MO
Any hardware I would even consider buying would need to be much, much more versatile than that. I myself wouldn't even buy a $10 VST plugin that only did orchestral brass and winds... 
-
- KVRian
- 954 posts since 15 Dec, 2000 from NY,NY,USA
Same here,simply beause I have no use for woodwinds/brass in my music for the most part,but if I did it would have to sound better and more realistic than the current sample based offerings,otherwise no matter how much cheaper it would be than a sample library,it's wasted money if it's not better than that option.foosnark wrote:I myself wouldn't even buy a $10 VST plugin that only did orchestral brass and winds...
"Technological progress is like an axe in the hands of a pathological criminal." - Albert Einstein
-
- KVRist
- 167 posts since 7 Apr, 2001
I for one, am still waiting for either Yamaha, or someone else to make a VL1 VSTi. I currently use 2 VL70m's and of course my trusty Korg Oasys (PCI). I seem to recall Arturia sending out an email a year or so ago asking if people would be interested in PM wind VSTi...
"The conclusion, when all has been heard, is: fear God and keep His commandments, because this applies to every person. For God will bring every act to judgement, everything which is hidden, whether it is good or evil." Solomon.
-
- KVRAF
- 2315 posts since 11 Mar, 2003
I really don't think 192kHz is a really big selling point for anyone. Sure make it futureproof if you're bothered (surely 96kHz is enough?), but I wouldn't try and use it as a marketing point.
Last edited by Mr Arkadin on Tue Mar 03, 2026 4:18 am, edited 1 time in total.
-
- KVRian
- 954 posts since 15 Dec, 2000 from NY,NY,USA
Realistically speaking,users of the UAD are lucky to get 5ms at 44 (and that's just with plugs and no VI's)or 48 and Virus PoCo can't seem to hit 96 without problems if at all,so unless there's a new or more powerful DSP proc availible this might be an unrealistic forcast to begin with.16 voices at 192hz is going to require some serious DSP Proc power(at least 4 times the FW PoCo if not more) and a hefty amount of card memory.And if the physically modeled instruments aren't as good as current sampling library's some developer is going to be left holding a very expensive unbought box. Still game?electro wrote:16 realtime Physically Modelled Orchestral (Brass/Wind) voices running at 192Khz (Sort of like a polyphonic VL-1)? It could connect to our PCs via firewire/and or USB2
"Technological progress is like an axe in the hands of a pathological criminal." - Albert Einstein
-
- jaaathmaster
- 2690 posts since 1 Jun, 2001 from Marlow, S. Bucks, UK
It *is* better than that option, at least in terms of expressivityAlan wrote:Same here,simply beause I have no use for woodwinds/brass in my music for the most part,but if I did it would have to sound better and more realistic than the current sample based offerings,otherwise no matter how much cheaper it would be than a sample library,it's wasted money if it's not better than that option.foosnark wrote:I myself wouldn't even buy a $10 VST plugin that only did orchestral brass and winds...
I'm primarily interested in something capable of expressive solo brass and reed stuff... And I would definitely pay much more than $10 for it. Then again, I have a sax playing pal, so it would have to be better than the existing Tassman or Reaktor PM ensembles, and it would never replace the real McCoy, but rather compliment it. I'd love to be able to do the bizarre morphing the VL-70m is capable of, with the reed/brass tones it can produce, but in VST integrated format.
A hardware unit is a nice idea, as long as it has VST integration as described, but I'd rather it was a native plug. If the 'hardware' aspect persuaded Yamaha or Korg, thus far reticent to implement their PM technologies, to implement a VST integrated PM instrument, that would be a big jump forward. Personally, not bothered about high sample rates, 44.1khz is plenty for me.
Music with dinner is an insult both to the cook and the violinist.
-
- KVRist
- 181 posts since 22 Aug, 2003 from Olympus Mons
to spread some rumours....WhiteNoise Audio has a universal PM Synth in the making...sounds promising 
.
Cheers, Richard
. Cheers, Richard
-
- KVRAF
- 8711 posts since 24 May, 2002 from Tutukaka, New Zealand
Likewise, having a 192KHz sample rate doesn't impress me at all ... it may as well be 768KHz or 44.1KHz - there are many a h/w unit that make do with 44.1 and still sound good, so it's not a selling point at all. To make that type of hardware would mean more power inside, probably raise the costs, and thusly put me off it completely.
Generally I like the idea. But also I don't think a unit dedicated only to brass/woodwind would sell that many - it wouldn't be enough to tempt me either.
Don't forget that the modelling synths never sold that well...mainly because the main bulk of potential users also wanted access to synthetic sounds as well.
If the price were right, I maybe would be tempted if it could also do strings (including plucked and strummed...not just bowed), but realistically I would also want access to some of the more esoteric parameters to enable me to make weird noises - not just orchestral.
One of the real problems with any unit like that is the realism that has always been lacking in any polyphonic play. By that I mean...it may sound incredible one a single note - but inevitably when you play chords you are simply repeating the same patch parameters - i.e. you get 4 notes all playing exactly the same patch. Which is not at all how a group of brass/violin/oboe players sound when playing together.
That is also a problem with sample based orchestral units, and huge sample volumes for VSTi samplers. So to get anywhere near realistic ensembles you need to lay down single notes from different patches, or single notes from the same patch, but tweaked differently. I do this to get even something like a brass section of 2 trumpets and a couple of saxes in my dub stuff (which doesn't necessarily have to be ultra-realistic - but I know if I simply play chords, it sounds crap)....
OK....so this is not a h/w unit problem only - sampling hits the same problem....BUT we know we need to do this to get realistic instrumentation. If we've got to go to all those lengths anyway, what possible benefit does a polyphonic h/w unit have over a s/w sample volume? Unless this h/w unit can play 16 notes each note with its own discrete patch. Is this in your plans for the modeller unit?
If not, then the cost is going to be alot more than a sample volume, and yet will be no more user-friendly.
In some ways I can see why the VL1 was mono (not just because it was new technology), but it actually would be no more useable if it were poly. If I had a poly VL1, I would still be laying down single note by note.
So.....sounds like a good idea on the surface, but it would have to be very attractive in its capabilities to move me. I almost know from the start that it's going to be only effectively useable as a mono unit, so what's the point of high samplerate poly? And if I were to play it in poly mode, there has to be access to alot of parameters and it has to sound f**king fantastice for it to have any benefit over sampling.
Generally I like the idea. But also I don't think a unit dedicated only to brass/woodwind would sell that many - it wouldn't be enough to tempt me either.
Don't forget that the modelling synths never sold that well...mainly because the main bulk of potential users also wanted access to synthetic sounds as well.
If the price were right, I maybe would be tempted if it could also do strings (including plucked and strummed...not just bowed), but realistically I would also want access to some of the more esoteric parameters to enable me to make weird noises - not just orchestral.
One of the real problems with any unit like that is the realism that has always been lacking in any polyphonic play. By that I mean...it may sound incredible one a single note - but inevitably when you play chords you are simply repeating the same patch parameters - i.e. you get 4 notes all playing exactly the same patch. Which is not at all how a group of brass/violin/oboe players sound when playing together.
That is also a problem with sample based orchestral units, and huge sample volumes for VSTi samplers. So to get anywhere near realistic ensembles you need to lay down single notes from different patches, or single notes from the same patch, but tweaked differently. I do this to get even something like a brass section of 2 trumpets and a couple of saxes in my dub stuff (which doesn't necessarily have to be ultra-realistic - but I know if I simply play chords, it sounds crap)....
OK....so this is not a h/w unit problem only - sampling hits the same problem....BUT we know we need to do this to get realistic instrumentation. If we've got to go to all those lengths anyway, what possible benefit does a polyphonic h/w unit have over a s/w sample volume? Unless this h/w unit can play 16 notes each note with its own discrete patch. Is this in your plans for the modeller unit?
If not, then the cost is going to be alot more than a sample volume, and yet will be no more user-friendly.
In some ways I can see why the VL1 was mono (not just because it was new technology), but it actually would be no more useable if it were poly. If I had a poly VL1, I would still be laying down single note by note.
So.....sounds like a good idea on the surface, but it would have to be very attractive in its capabilities to move me. I almost know from the start that it's going to be only effectively useable as a mono unit, so what's the point of high samplerate poly? And if I were to play it in poly mode, there has to be access to alot of parameters and it has to sound f**king fantastice for it to have any benefit over sampling.
-
- KVRAF
- 3096 posts since 3 Nov, 2002 from Kettering UK
kritikon, you make some excellent points regarding Polyphony.. I wanted to make the same point but you explained it perfectly.. The reason why ensemble pieces are so hard to acheive on a polyphonic instrument!
-
- KVRAF
- 1644 posts since 18 Mar, 2004 from Lincoln, CA
I think if you're going to make the user go to the extra trouble of connecting to an external device via USB or Firewire, then you should make sure you have a product that is significantly better than something like Synful or Edirol HQ Orchestra. Both of those are small in size and great in sound quality/playability, and neither makes you have to connect to something external. The whole point of using external DSP power is to have something that would be too taxing for your DAW's CPU, and that suggests something very highend and powerful.
-
- jaaathmaster
- 2690 posts since 1 Jun, 2001 from Marlow, S. Bucks, UK
IMO for solo brass/woodwind, physical modelling is better than both.Lunatique wrote:I think if you're going to make the user go to the extra trouble of connecting to an external device via USB or Firewire, then you should make sure you have a product that is significantly better than something like Synful or Edirol HQ Orchestra.
Well, this is how it's been marketed to the public, although I could be cynical here and suggest the appeal of DSP cards for the developers is more to do with intellectual property than raw computing horsepower - harder to copy a PCI card or Firewire boxLunatique wrote: Both of those are small in size and great in sound quality/playability, and neither makes you have to connect to something external. The whole point of using external DSP power is to have something that would be too taxing for your DAW's CPU, and that suggests something very highend and powerful.
Still, if that's what it takes to get those cool PM algorithms into VST-land, so be it
Music with dinner is an insult both to the cook and the violinist.
-
- KVRAF
- 6478 posts since 16 Dec, 2002
Electro, you might want to study some AD/DA basics before requesting 192khz. Even the chief designer of Lavry AD/DA converters (just about the best in the whole world) units says even 96khz might be too much.
Why?
Study the phrase "There's only one way to draw a sine wave thru two discrete points in time". While only the most high-end converters and clocks meet this specification, it's enough for AD/DA designers to strive for something else than mad samplerates. You will find that the best units in the world sound just about exactly the same at, say, 48khz and 192khz. (and no, protools HD will not sound the same, guess why?)
192khz is thoroughly prosumer driven marketing. Like most of the time, prosumers don't quite know what they're requesting.
Now before you get mad at me,
DSP *internal* samplerate is a whole different issue and has everything to do with the output quality of whatever is being modelled. Here 192khz suddenly isn't such a big number, and it is usually filed under "oversampling". Math accuracy is the goal here.
That doesn't mean the unit has to output this ridiculous rate.
Why?
Study the phrase "There's only one way to draw a sine wave thru two discrete points in time". While only the most high-end converters and clocks meet this specification, it's enough for AD/DA designers to strive for something else than mad samplerates. You will find that the best units in the world sound just about exactly the same at, say, 48khz and 192khz. (and no, protools HD will not sound the same, guess why?)
192khz is thoroughly prosumer driven marketing. Like most of the time, prosumers don't quite know what they're requesting.
Now before you get mad at me,
DSP *internal* samplerate is a whole different issue and has everything to do with the output quality of whatever is being modelled. Here 192khz suddenly isn't such a big number, and it is usually filed under "oversampling". Math accuracy is the goal here.
That doesn't mean the unit has to output this ridiculous rate.