- 6,948
- United Kingdom
Sounds in-game to me and it has even GT branding.
The Sound Update Thread (The Return)
- Thread starter TayeezSA
- 1,178 comments
- 147,367 views
- 1,245
Except Polyphony are notorious for dubbing their trailers with pre-recorded sound. They aren't the only developer in the world to do it, but it's all the more obvious when the in-game sound is so poor. The Bathurst trailer, for one, was nigh-on scandalous in its use of completely unrepresentative audio in almost every scene.
Thankfully, these days, the new AES audio system means that the VGTs that have been added are, in-game, generally much closer to the sounds in the trailers. Check out the Mazda LM55 or the Subaru VIZIV for some really excellent in-game audio, and logically, being new to the game, the Bugatti will use AES audio too, meaning it won't sound too far off the trailers either. I'm fairly confident these are still dubbed trailers, but I suspect Polyphony are far more happy using that audio in the trailers knowing that it is far more closely representative of what actually appears in game.
- 6,948
- United Kingdom
- 1,245
Unlikely. If AES is the sound method they've devised and used in limited quantities so far, then they're going to use it for GT7. These are clearly not sounds from AES, as they're recorded audio as opposed to generated, synthesised audio.
Did any new car in GT6 bring proper audio in the old sample format? I think most had really lazily done sounds with samples from other cars (like the Veyron, as you mention), suggesting they did not bother recording new sounds in the old format because it was inevitable they would be redone with AES. In short, that comment from Kaz about placeholder audio was correct, but only in as far as they're placeholders to be replaced in GT7.
- 6,948
- United Kingdom
Gear changes are 100% a physics / control issue. The sound engine just does as it's told; the moment you start adding fakery into the sound part to cover these physicalities, you start chasing your own tail and creating lots of special cases, when what you really need for 1000+ cars is (yes) a general model. Scipted physical gear changes fall into the category of "fakery", also, and technically exhibit the same "special cases" problem. That's maybe not a problem if you only have a handful of cars to worry about, and no modifications possible.
Using sound fakery also separates the audio feedback from the actual physical situation, adding a layer of garbled communication between the virtual car and the player. Some games fix that separation by forcing the physics to copy what the sound engine demands, but that's hardly appropriate in this case, where it should be the physics ruling all else.
So the gear changes might sound more involving etc., but to do so in this superficial way, you either have to make the cars harder / less responsive to drive (ambiguity of feedback) or you make the underlying simulation less physical. What's more important is a matter of opinion, and genre definition, but with the model you can at least get the better sounding gear changes with no detriment to the physicality or the feedback for a very large number of cars.
There are hints that PD are planning a drivetrain overhaul, as there are lists of transmission types, some of which aren't differentiated in the game as-is. The hope is that it implies a future physics change, which is greatly needed for the drivetrain modeling anyway. Once that happens, the physicality of gearchanges should be much better represented and differentiated per car.
It'll have the added benefit that PD can plug in the physical values of the cars (still) and everything should just work, with only fine tuning required, rather than hand-authoring what are basically special effects for every car (or having cars sound too similar as that unique treatment diminishes).
I never said it didn't sound good.
The main advantage with using a model instead of hand-authored content is its sheer flexibility. You cannot use samples to represent differences in cam timing, unless you have those samples to hand. You can do it with the model, and 1:1 simulation sounds as good as the real thing, if you set it up that way (Sonory).
In time, real-time implementations will have the same level of perceived quality despite the truncations and approximations, but those compromises will stay because they use less computational power than the 1:1 simulation: just as it is with graphics, and the existing sample-based synthesis.
You seem to think this is some kind of aesthete vs. technocrat issue; it isn't. It's a content issue.
I'm a bit confused about you labeling real-time synthesis of physicaly accurate engine sounds as the "easy way", though. Surely just copy-pasting pre-made sounds (a.k.a. sampling) is the "easy way"? Unless you mean in terms of content generation (once you have the tools prepared, and the relevant data isolated).
I'm no expert by any means and have no programming experience; only thinking off the top of my head here, but what about some kind of algorithm that covers the entire sound spectrum as a kind of analog/digital hybrid, which can be used by all cars?
All sound is broken down into wavelengths at its basis, and in the case of a driving game there would be subcategories of combustible engines, diesel, turbo, etc. The whole spectrum of each can be stored as hi fidelity sound data, and each car could use different ranges of it. Things like shifting can be augmented as needed in real time. Each car could have its own program data.
Just rudimentary thoughts, but I suppose if it was that simple they would've done it already haha.
- 8,723
You've got the right idea, but it really isn't so simple to create something that works for 1000+ cars without using the "shortcuts" and "static foundation" offered by a physical model.
Regarding sound generation, one method I've used is to start with a simple, but broadband sound, just as you state, and then morph it to suit any particular engine configuration you can imagine. Then a bit of resonant and / or subtractive filtering can shape the final sound closer to the intended target.
PD's new AES method is basically doing this, but I have no idea how that initial step is being performed, and precisely how the "morphing" and "filtering" parts are done. Well, I could guess, I suppose.
The problem with giving each car its own "program data", is that data is now an asset you need to manage. So the best option is to only use the real physical properties of the car (things that will never change), and then you needn't address it ever again. But the model / method of communicating the fact that a gear change has just occurred can be improved and changed with every iteration, or in any update etc., and it'll automatically work with all cars. Assuming it's a physically accurate model.
- 1,245
The XKR-S is... pretty good actually! Most other cars I can think of that were new used either rubbish samples or reused samples from other cars.
I think the big thing with AES is not just the general "quality" of the sounds (as there's proof both in GT and in other games that the sample method is effective), but the fact that it's easily tweakable to create new audio for modified cars, and easily assign appropriate audio to the car, dependent on the configuration of the engine.
Put simply, it's good for lots of cars and the modifications you can make to them that might make them sound different - or in other words, exactly what PDI want for GT.
- 2,534
- Johannesburg
- Tayeedz1
So I stumbled upon this video over in the Forza subforum.
Guess this further corroborates @Griffith500 's talk of how the sounds in Forza 2 are something for GT7 to aspire to with the new addition of AES. My ears are ready.
Guess this further corroborates @Griffith500 's talk of how the sounds in Forza 2 are something for GT7 to aspire to with the new addition of AES. My ears are ready.
- 8,723
What an excellent video; good find!
We all know we want sounds to be responsive to the changes we make, it's part of being interested in cars, especially since the engine can be considered the heart and soul of a car, in some respects.
Of course, FM2's sound is imperfect in many regards (and its successors did triumph over it, just usually on technical grounds rather than the aesthetic), and not just because of the usual stuff I go on about like sampling and dynamic range and mixing and reverb blah blah, but also because there are only really (combinations of) two sounds per car.
AES could offer virtually limitless expression, if PD were to take the time to allow us to fiddle with its controls. That's no small task, because the scope for complexity and bewilderment is vast, and so much of that space will produce very similar-sounding garbage that it would struggle to be a success.
The trick is creating controls that mimic the paths-of-least resistance of the real components on real engines (viz. packaging etc.), so that natural sounds just fall out of it. But those controls must also allow for easy, intuitive exploration outside of that "tried and true" beige-zone - so it's the combination of both aspects that's the real trick.
In the light of past successes like FM2, Project Gotham 3, plus current successes such as iRacing and (likely) future successes such as Dirt Rally, the thing that must set GT apart (in terms of inspiring by example) in the sound department is this responsiveness to player creativity. It absolutely must, because I suspect AES will not be able to compete on fidelity alone just yet.
We all know we want sounds to be responsive to the changes we make, it's part of being interested in cars, especially since the engine can be considered the heart and soul of a car, in some respects.
Of course, FM2's sound is imperfect in many regards (and its successors did triumph over it, just usually on technical grounds rather than the aesthetic), and not just because of the usual stuff I go on about like sampling and dynamic range and mixing and reverb blah blah, but also because there are only really (combinations of) two sounds per car.
AES could offer virtually limitless expression, if PD were to take the time to allow us to fiddle with its controls. That's no small task, because the scope for complexity and bewilderment is vast, and so much of that space will produce very similar-sounding garbage that it would struggle to be a success.
The trick is creating controls that mimic the paths-of-least resistance of the real components on real engines (viz. packaging etc.), so that natural sounds just fall out of it. But those controls must also allow for easy, intuitive exploration outside of that "tried and true" beige-zone - so it's the combination of both aspects that's the real trick.
In the light of past successes like FM2, Project Gotham 3, plus current successes such as iRacing and (likely) future successes such as Dirt Rally, the thing that must set GT apart (in terms of inspiring by example) in the sound department is this responsiveness to player creativity. It absolutely must, because I suspect AES will not be able to compete on fidelity alone just yet.
- 2,534
- Johannesburg
- Tayeedz1
@Griffith500 what controls that are user friendly could we the players mess around with?
For one I think GT should bring back different tuning companies parts and particularly for their exhausts, have sounds that differentiate one exhaust from 1 tuning company to another. For example, the difference in tone and loudness between an HKS Hi Power and a Kakimoto exhaust. Should be possible with AES no?
One thing I can congratulate GT on is that they actually got the sound of how an RB26DETT sounds right.
GT6 sound
The real sound (Although with an HKS exhaust it's close to the factory sound)
I don't know how to describe it well but the RB26 just has this soft whaling sound that other games can't seem to replicate all that well. Even Forza doesn't sound quite right let alone NFS. So Kudos to GT there.
For one I think GT should bring back different tuning companies parts and particularly for their exhausts, have sounds that differentiate one exhaust from 1 tuning company to another. For example, the difference in tone and loudness between an HKS Hi Power and a Kakimoto exhaust. Should be possible with AES no?
One thing I can congratulate GT on is that they actually got the sound of how an RB26DETT sounds right.
GT6 sound
The real sound (Although with an HKS exhaust it's close to the factory sound)
I don't know how to describe it well but the RB26 just has this soft whaling sound that other games can't seem to replicate all that well. Even Forza doesn't sound quite right let alone NFS. So Kudos to GT there.
- 8,723
It would fundamentally depend on the (performance) tuning options.
Take the current setup, just looking at exhaust for now:
Stock / Sports / Semi-Race / Race exhaust (effectively cat-back)
Stock / Racing cats
Stock / Racing manifold
I'll start with the manifold; PD will have the stock manifold "parameterised" already for the game and then, supported by collecting a few bits of info about the physical size and (perhaps) the packaging of the engine in the car, the "Racing" manifold could be a totally custom item - at least sound-wise. I personally would have the player pick from a few visually recognisable standard configurations per engine type (and aspiration etc.), then add a couple of sliders that control the specific shape / "size" of the manifold itself (procedural geometry). This can be done with the synth running, so players can hear the changes they're making and will intuitively pick up a technique that works for them. This would also ideally affect the engine performance in a meaningful way, but that requires a tuning overhaul.
The cats can be bundled in with the exhaust tract, in terms of their sound treatment. This part is more free-form, and there are a few ways to tackle it but they come under two main groups: abstract and physical. Either let the player hand-tune a few filters to get the sound timbre they want (using intuitively labelled sliders that control parts of the sound that players are "aware" of, say), or confine the tuning to the actual physical parts on the car.
With the first option, it'd be too easy to make one car sound like another (and also to find that vast "weird zone"), but it's far easier to create content for. With the second approach, it might be more intimidating for beginners (plenty of presets / parts / constraints will help), but each car will more definitely have its own nuances. A mountain of data would be required, though, potentially adding complexity to content creation (adding new cars).
The physical aspects you might change in the exhaust tract are its diameter(s), exit point on the car, resonator / cat / muffler locations and dimensions, tip type and size etc. - maybe eventually even branching and merging (true dual, x-pipe, LFA / F355 etc.) These would probably have to be created by PD and saved as presets for each upgrade "stage" first, with some limited scope for modification (tone, volume etc.).
A good place to start with full customisation is on motorbikes...
Perfectly possible, but the issue is collecting all the necessary data and tuning any approximations in the model to fit each part separately. It also means the above presets are no longer enough, and full control is required (say, after motorbikes; GT8?). This is a lot of work, and would probably favour the second approach I mentioned above (physical) so that we can mix and match from a relative few parts that interact believably.
I think half of PD's existing sample database is taken up with recordings of various RBs in various states of tune!
- 1,492
- Germany
- KiroKai
- KiroKai
I always felt the closest sound to an RB is oddly the 300SL's sports exhaust. But then I haven't heard real RBs often, heard many more VR38s in person and on video so I might be mixing them up. Still, it's a strangely unique sound that seems to be taken from a different car, could you give it a listen and tell me your thoughts @Griffith500 ?
- 8,723
I'd say that it's quite likely to be an RB26DETT with a "sports exhaust", yes. It has that soft wail combined with the taut growl from the "twin triple" configuration of the exhaust manifolds and downpipes. Sort of half Ferrari F12, half E36 M3 (exhaust only). Speaking of E36 M3s, here's a great video of some subtle and not-so-subtle changes to the exhaust on the same car under similar circumstances.
The VR / VQ is fundamentally a different sound, more warbly due to the packaging / configuration / firing order etc. differences between a V6 and an inline 6. It's possible to get either to sound identical, of course, especially if there are no constraints on packaging.
There is a softness to the stock VR38 that emulates the softness of its inline six "ancestors" - this is all in the tuning of the exhaust parts downstream of the manifolds, though. I believe that was deliberate in this case.
- 2,534
- Johannesburg
- Tayeedz1
Thanks for giving us answers & hope as always @Griffith500
I remember you said that this strictly relates to the drivetrain model but Forza with a manual-clutch wheel setup sounds so right! The gear changes sound on point and I believe I heard a bit of gear grinding too if I'm not mistaken.
AES not only pertains to the engine, induction and exhaust but to other factors no?
I remember you said that this strictly relates to the drivetrain model but Forza with a manual-clutch wheel setup sounds so right! The gear changes sound on point and I believe I heard a bit of gear grinding too if I'm not mistaken.
AES not only pertains to the engine, induction and exhaust but to other factors no?
Last edited:
- 8,723
I personally think drivetrain stuff should be kept separate from the sound stuff, because the drivetrain is gameplay and the sound is just a representation of that gameplay that we can interpret better than the raw numbers stored in memory.
The player is the input; the sound and the graphics are the feedback; everything in between is the game.
In its current form, AES could not be used to model the performance of an engine, I don't think; a different model could be used for that and that could then modify the parameters of AES in real time to reflect any variation, but that's a separate system. At the point that AES is the engine simulation, then it ceases to be sound and is simply a physical simulation, for the purposes of interaction. You still have to choose where to take your outputs from that and render them to the game's audio scene as feedback, by analogy with the relationship between physics and graphics.
It's just shifting the boundaries of interaction, changing where new information is generated in an interactive sense (gameplay), and when it's just being converted from one form to another (feedback).
GT also sounds much better with H-pattern control, because the commands sent to the sound renderer result from the simulated physical reaction of the car based on the player's input via physically-representative controls. The input of a DS3 / DS4 is not a great representation of a real car's controls (particularly in the way we use them), so this difference affects the physics first, then the sound. You're just pressing a button to change gear: how is that representative of what actually happens, if you don't model what that button press is meant to represent?
The fix is to make the game respond to the controls as though they were real controls controlling real things, first and foremost. Then all the feedback systems should just work as they always did (and as they currently do in the H-pattern case).
The player is the input; the sound and the graphics are the feedback; everything in between is the game.
In its current form, AES could not be used to model the performance of an engine, I don't think; a different model could be used for that and that could then modify the parameters of AES in real time to reflect any variation, but that's a separate system. At the point that AES is the engine simulation, then it ceases to be sound and is simply a physical simulation, for the purposes of interaction. You still have to choose where to take your outputs from that and render them to the game's audio scene as feedback, by analogy with the relationship between physics and graphics.
It's just shifting the boundaries of interaction, changing where new information is generated in an interactive sense (gameplay), and when it's just being converted from one form to another (feedback).
GT also sounds much better with H-pattern control, because the commands sent to the sound renderer result from the simulated physical reaction of the car based on the player's input via physically-representative controls. The input of a DS3 / DS4 is not a great representation of a real car's controls (particularly in the way we use them), so this difference affects the physics first, then the sound. You're just pressing a button to change gear: how is that representative of what actually happens, if you don't model what that button press is meant to represent?
The fix is to make the game respond to the controls as though they were real controls controlling real things, first and foremost. Then all the feedback systems should just work as they always did (and as they currently do in the H-pattern case).
- 2,534
- Johannesburg
- Tayeedz1
Sorry for the double post but this is worth posting.
@Griffith500 I think you might like this and the bit at the end.
@Griffith500 I think you might like this and the bit at the end.
- 8,723
It echoes some of the things I'm often saying:
It's not easy.
Physics first, sound second.
Computational resources can be scarce.
Interactive sound is at its best at its most interactive.
Game audio is not something many people grasp intuitively; it's technical.
I think a little app using PD's new tech (or something similar) would be pretty cool.
Imagine tuning an engine and tweaking all the intake and exhaust components to get a particular sound, and then being able to play with that bespoke sound, or sync it up with something real...
- 218
- Ontario, Canada
- MaxPrez90
Their AES method seems very good, but they should really get an X360 and play PGR4 for an example of stunningly accurate virtual engine sounds. The turbo spools, blowoff valves, induction, and exhaust sounds are among the best, if not THE best, in any racing game ever made.
That Ruf Supercar, though...eargasmic.
That Ruf Supercar, though...eargasmic.
- 2,534
- Johannesburg
- Tayeedz1
I saw that one, wanted to post it but was looking for a good reason behind it.
In relation to what @Griffith500 has said before, it's getting there but sounds a bit off because there's no induction noise along with the other parameters in AES.
- 4,992
- Nowhere special
- SweetLilPoppet
- SweetLilPoppet
- 13,913
- Adelaide
- Neomone
I'm afraid it's not. The interior filter makes it sound even more echoey, but even from the external view it still sounds very odd.
