If they already have a real time solution, why are we waiting for it?
No idea, in honesty, but it's already on the RB cars as an exaggerated exhaust-only version. There is only one exhaust preset in use, but it's "driven" differently for the V6s versus the Junior's 4 cylinder engine - so already you should be able to see its flexibility. Based on that, they're wanting to add extra layers and introduce a new mixing scheme to cope with them, maybe a robust virtual channel switching, source grouping and per-source LoD scaling algorithm, or any combination / omission, is the holdup.
The point was that whatever PD does or doesn't have, it's not functional yet. Whatever Sonory has is, and if it's their livelihood (or his livelihood) then they likely know a lot about it.
I'm sure the model itself is functional on superior hardware, the exhaust already is on PS3 - again, I think it's a mixing issue of squeezing so many more channels onto the SPUs and doing so in a way that sounds don't disappear noticeably. I'm sure PD know more for simply having engineered their own solution running in realtime, as opposed to putting numbers into a commercial simulation software package. Don't be so simplistic.
Trying to learn everything from scratch is a bad way to work when there are already professionals in the field. Hire experts, learn from them, and then improve on their work yourself. Otherwise you're just repeating work that someone else has already done. Waste of time and money.
Just going by the RB cars it's plain to see that no-one is doing what they've already done. And there's more to come. Doing it from scratch actually is the only way to make progress, samples are a dead-end - always re-assess prior assumptions, it's good engineering practice because things change quickly and often. Plus, they'll have the professionals that actually matter on-board already, no doubt. People with real experience in real-time sound synthesis techniques and tricks. GT has always had sound designers, but it wasn't until GT5 Prologue that it got sound simulation credits, too.
I don't have as much faith as you that PD have a working system. Whatever the RB Junior is indicative of, it's not a final solution.
No it's not a final solution, but as with GT5 Prologue hinting that they were working on this system, it's the gaps and discrepancies in those cars that point to what's really going on.
I wasn't suggesting that they get Sonory to make samples. I was suggesting they use Sonory's physical model, which is demonstrated to actually work in a commercial environment. If they can't get a cut-down version of that running in real time, then using it to make highly optimised samples would be an acceptable stopgap.
They can't do that in realtime, because their model is not a real time model! If it were, they'd offer it as such! (This has happened in the past). Stop being so simplistic!
But working from a functioning, successful physical model that can simulate real engines would seem to me to be a good place to start. All we've seen so far is entirely fictional engine notes on fictional cars, if indeed the Red Bulls were generated that way.
Time will tell as to how accurate they can get it to sound on a per-car basis, and will hinge just as much on the usability of the dev tools they made for the existing sound artists as it will the synthesis method itself. But really the hard work's been done: getting the real-time synthesis working to such exquisite responsiveness and detail in the exhaust alone, as demonstrated by the RB cars.
That exhaust sound can easily be tweaked and filtered etc. to create any range of colours and qualities, which will interact dynamically with whatever is driving it - again, as demonstrated by the RB cars. It's the "driving" part that's the real challenge, at any rate, and that seems fine; they appear to have solved what I considered to be unsolvable in that respect, going by the sheer sense of continuity in the smoothness of the sound progression over the rev range. You don't get that with samples.
