Give us better sounds - PLEASE !!

  • Thread starter steamcat
  • 4,667 comments
  • 334,179 views

Holy hell! This made me want to fire up GT 6!!!! ... If only... I'll take mono sound if that's it... You're whole head is numb in a real race car anyway... Minise the other cars.

GT would do better to ditch the entire point source audio and just run a single sound for internal and external, inside muffled a bit (not much), the outside raw.

A fine example how realistic isn't fun in games... I really care less of the mechanical motor sound, it does nothing for me as a driver... Except for telling me when something broke...
 
How many people will actually be able to appreciate these new, super-accurate sounds?

If the sounds in GT7 don't sound good enough it's because your ears clearly suck and need replacing. You all need to upgrade your ears and get cochlear implants to truly appreciate the new sounds once they arrive.
 
There's nothing to talk about GT6. You fail to see that. I said keep hope alive that GT7 will have better car sounds. You seem to be of some ridiculous belief that the PS3 hardware limitations are making the cars sound like vacuum cleaners so there can be no solution except wait until GT7. I find it hilarious you don't notice that PD has recycled a handful of car sounds that they use on every vehicle in the game.

No, I considered that from the start, and it contributed to my intense curiosity about the situation some years ago, because it didn't seem to make sense. I even mentioned it to you in an earlier post, but I'm not surprised you missed it.

The hardware-limitation should be considered in conjunction with the specific direction PD are taking, which was sign-posted with GT5P in terms of the audio engine features (which also piqued my curiosity at the time). It's interesting to note that GT5P was the first game in the series with "sound simulation" in the credits.

Your only argument seems to be that the Red Bull and Senna cars have different sounds so that's less than 10 cars total with a different sound than the vacuum cleaners used for every other car, but wait, the PS3 has a hardware limitation so therefore how can those cars sound differently? Oh that's right, you claim some sort of DSP effect so let me get this straight, you think PD is going to go back and edit the car sounds of every car in the game with some DSP effect? So 10 down, 1200 more to go. How long will that take you think?
Once again, what evidence is there to suggest that "editing" is the hold-up?

I made a demonstration video some time ago in support of my theory that the single exhaust sound (yes, there is only one!) used on all the Red Bull and Senna cars is a deliberate exaggeration to compensate for missing sources; those missing sources, as I've already mentioned to you, require extra processing power to control and perform operations on (mixing etc.) and make that 64 channel hardware limit more of a problem.

And when I say DSP, I don't just mean a wee bit of filtering; that "simulation" word applies.

This is the part I find most odd about this whole sound debate - we are all clamoring for the best/most accurate sounds, but will we even know when we've got it? Case in point Forza, whose sounds supposedly aren't very accurate, sound good to my plebeian ears.

How many people will actually be able to appreciate these new, super-accurate sounds?

Nobody knows. What you're describing is aesthetics, and that could go any way, regardless of the technology used. Do you like the Red Bull cars? What about the Senna ones? Would everyone agree with you?

Does everybody like the distorted sounds Forza has offered since the third outing? What about Shift? iRacing? Etc.

If nobody notices a difference in the new sounds (when they've arrived), you can't really call it an update. And nobody said super-accurate except you.
 
No, I considered that from the start, and it contributed to my intense curiosity about the situation some years ago, because it didn't seem to make sense. I even mentioned it to you in an earlier post, but I'm not surprised you missed it.

The hardware-limitation should be considered in conjunction with the specific direction PD are taking, which was sign-posted with GT5P in terms of the audio engine features (which also piqued my curiosity at the time). It's interesting to note that GT5P was the first game in the series with "sound simulation" in the credits.


Once again, what evidence is there to suggest that "editing" is the hold-up?

I made a demonstration video some time ago in support of my theory that the single exhaust sound (yes, there is only one!) used on all the Red Bull and Senna cars is a deliberate exaggeration to compensate for missing sources; those missing sources, as I've already mentioned to you, require extra processing power to control and perform operations on (mixing etc.) and make that 64 channel hardware limit more of a problem.

And when I say DSP, I don't just mean a wee bit of filtering; that "simulation" word applies.

Nobody knows. What you're describing is aesthetics, and that could go any way, regardless of the technology used. Do you like the Red Bull cars? What about the Senna ones? Would everyone agree with you?

Does everybody like the distorted sounds Forza has offered since the third outing? What about Shift? iRacing? Etc.

If nobody notices a difference in the new sounds (when they've arrived), you can't really call it an update. And nobody said super-accurate except you.


Sound Simulation...I find that very interesting! By the way, I enjoy reading your posts very much, thanks.
 
This is the part I find most odd about this whole sound debate - we are all clamoring for the best/most accurate sounds, but will we even know when we've got it? Case in point Forza, whose sounds supposedly aren't very accurate, sound good to my plebeian ears.

How many people will actually be able to appreciate these new, super-accurate sounds?

Forza has a variety of sounds for their cars, not all of them sound realistic but there's still a much larger variety than GT. PD seems to think all cars sound like vacuum cleaners. There will always be GT fanatics that see and hear what they want. Its somewhat sad honestly. I'm not saying Forza knocks it out of the park either but it is light years and dog years ahead of GT.
 
No, I considered that from the start, and it contributed to my intense curiosity about the situation some years ago, because it didn't seem to make sense. I even mentioned it to you in an earlier post, but I'm not surprised you missed it.

The hardware-limitation should be considered in conjunction with the specific direction PD are taking, which was sign-posted with GT5P in terms of the audio engine features (which also piqued my curiosity at the time). It's interesting to note that GT5P was the first game in the series with "sound simulation" in the credits.


Once again, what evidence is there to suggest that "editing" is the hold-up?

I made a demonstration video some time ago in support of my theory that the single exhaust sound (yes, there is only one!) used on all the Red Bull and Senna cars is a deliberate exaggeration to compensate for missing sources; those missing sources, as I've already mentioned to you, require extra processing power to control and perform operations on (mixing etc.) and make that 64 channel hardware limit more of a problem.

And when I say DSP, I don't just mean a wee bit of filtering; that "simulation" word applies.



Nobody knows. What you're describing is aesthetics, and that could go any way, regardless of the technology used. Do you like the Red Bull cars? What about the Senna ones? Would everyone agree with you?

Does everybody like the distorted sounds Forza has offered since the third outing? What about Shift? iRacing? Etc.

If nobody notices a difference in the new sounds (when they've arrived), you can't really call it an update. And nobody said super-accurate except you.


Jeez man, what is the hold up then? PD has been using the same vacuum cleaner sounds for GT for a decade and you still think that they're still going to release the proper car sounds that they've been keeping in a vault somewhere since the beginning?

GT has been the laughing stock of car sounds and damage modeling for years now, I don't know how this piece of very well known information has passed you by all these years.
 
Jeez man, what is the hold up then? PD has been using the same vacuum cleaner sounds for GT for a decade and you still think that they're still going to release the proper car sounds that they've been keeping in a vault somewhere since the beginning?

GT has been the laughing stock of car sounds and damage modeling for years now, I don't know how this piece of very well known information has passed you by all these years.

Vacuum cleaner is not a very technical term. We're having a serious discussion here. Some sounds are quite o.k, just missing some details (very important ones) such as intake sound. The Ralt RT3 sound is really good too and from what some of us can tell it's GENERATED (important word) in a different way.

God, this vacuum cleaner joke is old and annoying, go drive the stock Cizeta or the Ralt RT3 and tell me if it sound as a vacuum cleaner.

Peace.
 
Nobody knows. What you're describing is aesthetics, and that could go any way, regardless of the technology used. Do you like the Red Bull cars? What about the Senna ones? Would everyone agree with you?

Does everybody like the distorted sounds Forza has offered since the third outing? What about Shift? iRacing? Etc.

If nobody notices a difference in the new sounds (when they've arrived), you can't really call it an update. And nobody said super-accurate except you.

Outside people with your knowledge (a very, very small percentage of people) nobody really knows what they are talking about, or hearing for that matter when it comes to the sound issue. To me, the Red Bull/ Senna cars sound better. Accurate? Haven't a clue. It's for this reason that in my opinion, Polyphony should go down the easier Forza/distortion road rather than the new method you have been describing which they may or may not be pursuing. I think that in the end, all there hard work and innovation will be wasted on people that will not be able to appreciate it (unless you tell me about it of course :sly:)
 
Outside people with your knowledge (a very, very small percentage of people) nobody really knows what they are talking about, or hearing for that matter when it comes to the sound issue. To me, the Red Bull/ Senna cars sound better. Accurate? Haven't a clue. It's for this reason that in my opinion, Polyphony should go down the easier Forza/distortion road rather than the new method you have been describing which they may or may not be pursuing. I think that in the end, all there hard work and innovation will be wasted on people that will not be able to appreciate it (unless you tell me about it of course :sly:)
People care about the result and not how its done most of the time. How does an engine in a car work? How do i repair my car? How does the electronic products in the kitchen work? Of course, some people who like those kind of stuff want to now how it works, but at the end the result counts. It is the same in gaming. Most people don't care how developers are doing rain for example, they see it and look how realistic it looks and don't worry about how the developers achieved this. Same with sound. Almost no one cares, if PD has a secret method every other developer can't match, if the result are sounds that are less realistic as others.

If you have a complex problem, write simple code. If you have a simple problem, write simple code. Easy to remember
Christian Gyrling from Naughty Dog
Because they don't. 💡
The new recordings have not been processed into samples. Let that sink in for a second.
Are you busy in PD office or can you still post without Kaz seeing that you don't do your work? They have a new method since 2003 or something like that according to their blog. How is it possible that they didn't use it on the new cars? I believe you have said the hardware limitations, but then read the twitter post. Other developer worked within their limitations from the hardware and got much better results than PD.
 
Last edited:
Vacuum cleaner is not a very technical term. We're having a serious discussion here. Some sounds are quite o.k, just missing some details (very important ones) such as intake sound. The Ralt RT3 sound is really good too and from what some of us can tell it's GENERATED (important word) in a different way.

God, this vacuum cleaner joke is old and annoying, go drive the stock Cizeta or the Ralt RT3 and tell me if it sound as a vacuum cleaner.

Peace.
"Vacuum Cleaner" sound is a way for people with poor sound vocabulary (not a slight - we as a species are very visual oriented and as a whole are much better at describing pictures than sounds) to describe things like looped samples that have been pitch-shifted to the point of sounding unpleasantly synthetic or "stretched", as well as having a very smooth, sinewave-esque rhythm (as opposed to a gurgly or burbly sound in your typical muscle car, for example), as well as having over-emphasized high frequency noise (typically in the 2-5KHz range) (this can be a result of the stretched loops) (this can be caused by using a low sampling rate and high compression algorithm). Re: the rhythm, the reason you suggest the Ralt and the Cizeta sound good is because there is some pronounced rhythm to the samples used (you could call this "pulsey"), and I think you'll find that all the cars people suggest sound good have this type of rhythmic nature versus the cars that sound bad which are unnaturally smooth.

Cars sound exciting and dynamic and visceral because the character of the car changes based on what the engine is doing. For example, taking an audio capture of a car at 2000 RPM, and taking a capture of the same car at redline (let's say 8000 RPM), and then pitching one to match the frequency of the other, will still produce two different sounds, because the sounds the car itself makes changes - and that's just looking at RPM. To combat this, other games record the RPM at intervals of say 500 RPM. That way a lot of that changing character is still captured (plus the native pitch of the loops is relatively preserved, negating any "stretching" effects). Another way that character changes is based on load, IE a car at 8000 RPM at full throttle up a hill sounds different than when it is in neutral at 8000 RPM, which sounds different to when the car is going down a hill off-throttle at 8000 RPM (ie "negative load"). While there are many ways of recreating these effects, the most common is to either have an off-throttle sample set (memory intensive, CPU light) which is cross-faded onto and off of based on load or throttle or whatever. The other way is to manipulate DSP like equalizers or compressors or what have you (CPU heavy, memory light) to change the sound of one sample set to sound differently based on load.

GT does clearly have an issue at its audio content level, but I hope you can see by my post that there is also a great dependency on the parameters driving the audio content, which I would call "audio physics" but is usually just "physics." One of the key ways I can instantly identify when listening to GT is when a car shifts - none of the physics interactions of a real car are represented at all (clutch in -> no load, rpm dips, clutch out -> negative load if downshifting or spiked positive load if upshift is not rev matched. This process takes a minimum amount of time depending on the type of transmission.) Fixing the content would not fix the physics driving the content.

That said, there is a lot of room for improvement in the audio content for GT7. While I don't ascribe to Griffith's theories about some synthesizer-based system, I would expect to see a drastic improvement in audio fidelity (ie the sampling rate and compression scheme) at a bare minimum, and possibly some updates to granularity (ie number of loops). But then again, I also expected "standard" cars to die, and if rumor is true that isn't going to happen either.

btw, the red bull cars that Griffith claims to pinpoint some change in system to me just sound like some standard GT-esque loops that have been processed in-game with a distortion algorithm, and not a particularly pleasant one either. But the distortion does mask a lot of the negatives of the looping system (ie covers some of the stretchiness) as well as adds upper high frequency (12khz+) content that was otherwise dropped by the sampling rate and compression scheme, which is a noticeable improvement over the average car.

The reason he uses technical words is because this is a technical topic. It's easy to say "sounds good" or "sounds bad" or "sounds accurate" or "sounds inaccurate" but to describe how one gets to that conclusion and why a certain feeling is invoked requires a bit more precision that complex vocabulary provides. I try my best to make things as simple as possible, or at least as relatable as possible, but it's hard to immediately jump to "oh the sample rate causes this effect" if you guys don't already understand what sample rate is and how it can affect sound, for example. Griffith does a good job at trying to fill in those gaps which is why his posts are so long (as mine now is as well.)
 
Last edited:
Well, the big sound update not coming to the PS3 was pretty much a given, I honestly don't know a lot about sound but I have a good ear and Griffith's posts always make sense in relation to what I'm hearing. Just "observe" the sound and you'll see all the pieces coming together.
I edited my post now, because i want to give Kaz one last chance, but if GT7 on PS4 won't have realistic sound i'll lose my faith in Kaz completely and as sad as it is for me, because i grew up playing GT move on and play only other racing games. What will you do, if sounds don't get an huge improvement in GT7?
 
Last edited:
Well, the big sound update not coming to the PS3 was pretty much a given, I honestly don't know a lot about sound but I have a good ear and Griffith's posts always make sense in relation to what I'm hearing. Just "observe" the sound and you'll see all the pieces coming together.
 
How long do we wait for the third part of Kaz's blog post before we say that he's trolling us? A month? Two?

Two weeks on now, and it started so well but I'm still waiting for some actual new information out of it. He hasn't told us anything that the smart sound guys in here hadn't already figured out, as far as I can tell.

On the plus side, at least this time we also have access to the raw Japanese that he's writing so any misunderstandings of terminology can be cleared up.
 
How long do we wait for the third part of Kaz's blog post before we say that he's trolling us? A month? Two?

Two weeks on now, and it started so well but I'm still waiting for some actual new information out of it. He hasn't told us anything that the smart sound guys in here hadn't already figured out, as far as I can tell.

On the plus side, at least this time we also have access to the raw Japanese that he's writing so any misunderstandings of terminology can be cleared up.

To be fair it's been a pretty busy couple of weeks. I imagine that we'll get part 3 sometime in the next week. But that's just speculation.
 
To be fair it's been a pretty busy couple of weeks. I imagine that we'll get part 3 sometime in the next week. But that's just speculation.

Admittedly so, and I'm not getting on his case yet. But I think it's sensible to decide early what's a reasonable timeframe, because otherwise it's very easy to just say "we'll wait another couple of weeks". Three months later...you get the idea.

I would have thought that he'd write the entire thing in one go and then they'd split it up into sections, because that's what I'd do. I'd want to make sure that I got all the answers in that I wanted to, and that it fit together in a sensible narrative. I wouldn't want to get to part 4 and realise that there was more I wanted to flesh out about the early days of sound recording in order to make the future plans seem sensible.

But that's just me being a planner. Maybe Kaz can wing it, or maybe he has the whole thing written in his head already and just hasn't gotten it down in electrons yet.
 
Admittedly so, and I'm not getting on his case yet. But I think it's sensible to decide early what's a reasonable timeframe, because otherwise it's very easy to just say "we'll wait another couple of weeks". Three months later...you get the idea.

I would have thought that he'd write the entire thing in one go and then they'd split it up into sections, because that's what I'd do. I'd want to make sure that I got all the answers in that I wanted to, and that it fit together in a sensible narrative. I wouldn't want to get to part 4 and realise that there was more I wanted to flesh out about the early days of sound recording in order to make the future plans seem sensible.

But that's just me being a planner. Maybe Kaz can wing it, or maybe he has the whole thing written in his head already and just hasn't gotten it down in electrons yet.

Yeah that's a good point. Personally, I'd start to worry around the 1 month mark (1 month since sounds Pt. 2 that is) especially since now with Goodwood and the N24h over there really isn't anything else I can think of that they would post on Pitstop...
 
Vacuum cleaner is not a very technical term. We're having a serious discussion here. Some sounds are quite o.k, just missing some details (very important ones) such as intake sound. The Ralt RT3 sound is really good too and from what some of us can tell it's GENERATED (important word) in a different way.

God, this vacuum cleaner joke is old and annoying, go drive the stock Cizeta or the Ralt RT3 and tell me if it sound as a vacuum cleaner.

Peace.

So you have two cars to offer as example that don't sound like a vaccuum cleaner? So add those to the Red Bull and Senna cars and PD only has 1200 more to go.

People care about the result and not how its done most of the time. How does an engine in a car work? How do i repair my car? How does the electronic products in the kitchen work? Of course, some people who like those kind of stuff want to now how it works, but at the end the result counts. It is the same in gaming. Most people don't care how developers are doing rain for example, they see it and look how realistic it looks and don't worry about how the developers achieved this. Same with sound. Almost no one cares, if PD has a secret method every other developer can't match, if the result are sounds that are less realistic as others.

Christian Gyrling from Naughty Dog Are you busy in PD office or can you still post without Kaz seeing that you don't do your work? They have a new method since 2003 or something like that according to their blog. How is it possible that they didn't use it on the new cars? I believe you have said the hardware limitations, but then read the twitter post. Other developer worked within their limitations from the hardware and got much better results than PD.

Winner winner, chicken dinner.

"Vacuum Cleaner" sound is a way for people with poor sound vocabulary (not a slight - we as a species are very visual oriented and as a whole are much better at describing pictures than sounds) to describe things like looped samples that have been pitch-shifted to the point of sounding unpleasantly synthetic or "stretched", as well as having a very smooth, sinewave-esque rhythm (as opposed to a gurgly or burbly sound in your typical muscle car, for example), as well as having over-emphasized high frequency noise (typically in the 2-5KHz range) (this can be a result of the stretched loops) (this can be caused by using a low sampling rate and high compression algorithm). Re: the rhythm, the reason you suggest the Ralt and the Cizeta sound good is because there is some pronounced rhythm to the samples used (you could call this "pulsey"), and I think you'll find that all the cars people suggest sound good have this type of rhythmic nature versus the cars that sound bad which are unnaturally smooth.

Cars sound exciting and dynamic and visceral because the character of the car changes based on what the engine is doing. For example, taking an audio capture of a car at 2000 RPM, and taking a capture of the same car at redline (let's say 8000 RPM), and then pitching one to match the frequency of the other, will still produce two different sounds, because the sounds the car itself makes changes - and that's just looking at RPM. To combat this, other games record the RPM at intervals of say 500 RPM. That way a lot of that changing character is still captured (plus the native pitch of the loops is relatively preserved, negating any "stretching" effects). Another way that character changes is based on load, IE a car at 8000 RPM at full throttle up a hill sounds different than when it is in neutral at 8000 RPM, which sounds different to when the car is going down a hill off-throttle at 8000 RPM (ie "negative load"). While there are many ways of recreating these effects, the most common is to either have an off-throttle sample set (memory intensive, CPU light) which is cross-faded onto and off of based on load or throttle or whatever. The other way is to manipulate DSP like equalizers or compressors or what have you (CPU heavy, memory light) to change the sound of one sample set to sound differently based on load.

GT does clearly have an issue at its audio content level, but I hope you can see by my post that there is also a great dependency on the parameters driving the audio content, which I would call "audio physics" but is usually just "physics." One of the key ways I can instantly identify when listening to GT is when a car shifts - none of the physics interactions of a real car are represented at all (clutch in -> no load, rpm dips, clutch out -> negative load if downshifting or spiked positive load if upshift is not rev matched. This process takes a minimum amount of time depending on the type of transmission.) Fixing the content would not fix the physics driving the content.

That said, there is a lot of room for improvement in the audio content for GT7. While I don't ascribe to Griffith's theories about some synthesizer-based system, I would expect to see a drastic improvement in audio fidelity (ie the sampling rate and compression scheme) at a bare minimum, and possibly some updates to granularity (ie number of loops). But then again, I also expected "standard" cars to die, and if rumor is true that isn't going to happen either.

btw, the red bull cars that Griffith claims to pinpoint some change in system to me just sound like some standard GT-esque loops that have been processed in-game with a distortion algorithm, and not a particularly pleasant one either. But the distortion does mask a lot of the negatives of the looping system (ie covers some of the stretchiness) as well as adds upper high frequency (12khz+) content that was otherwise dropped by the sampling rate and compression scheme, which is a noticeable improvement over the average car.

The reason he uses technical words is because this is a technical topic. It's easy to say "sounds good" or "sounds bad" or "sounds accurate" or "sounds inaccurate" but to describe how one gets to that conclusion and why a certain feeling is invoked requires a bit more precision that complex vocabulary provides. I try my best to make things as simple as possible, or at least as relatable as possible, but it's hard to immediately jump to "oh the sample rate causes this effect" if you guys don't already understand what sample rate is and how it can affect sound, for example. Griffith does a good job at trying to fill in those gaps which is why his posts are so long (as mine now is as well.)

Thank you sir. I don't see the need to get all technical about poor sound quality when the bottom line is that most people don't care about all of that technical jargon. Bottom line is that GT has poor car sounds based on what the majority of people hear through their speakers, the majority doesn't really cares why, they just want PD to fix it which isn't going to happen with GT6.
 
Last edited by a moderator:
Outside people with your knowledge (a very, very small percentage of people) nobody really knows what they are talking about, or hearing for that matter when it comes to the sound issue. To me, the Red Bull/ Senna cars sound better. Accurate? Haven't a clue. It's for this reason that in my opinion, Polyphony should go down the easier Forza/distortion road rather than the new method you have been describing which they may or may not be pursuing. I think that in the end, all there hard work and innovation will be wasted on people that will not be able to appreciate it (unless you tell me about it of course :sly:)

I think Kaz has been testing the water on the low-fidelity front. Because the Red Bull cars etc. are pretty low-fi, and he did say they might need to "sex up" the sounds. LFS has incredibly low-fi sound, but the quality of the physics driving it (and the expressivity of the method used) more than makes up for that.
...

Are you busy in PD office or can you still post without Kaz seeing that you don't do your work? They have a new method since 2003 or something like that according to their blog. How is it possible that they didn't use it on the new cars? I believe you have said the hardware limitations, but then read the twitter post. Other developer worked within their limitations from the hardware and got much better results than PD.
There are a few examples of cars we know they recorded, and yet the samples they use in the game are clearly from another car. In fact, they're identifiable from earlier games. One example that I remember is the Audi R8 (V8), which was in GT5P; PD specifically mentioned they'd recorded it at Audi's own anechoic rolling road. The car in the game does not reflect this, and I heard another car (a Standard) recently that sounded exactly like the in-game version, but annoyingly I can't remember what it was. The HSV-010 went through three iterations, all using recycled samples from GT, and the latest version even borrows from TT; we were lucky it got such attention.

When you listen closely to practically all the cars, the sound they make is recognisable from earlier games; especially the modified sounds. The vast majority of sounds used date from GT2. The sampling format itself (schematically, not in terms of data storage) is the same as it was for GT2, broadly (we now have separate engine and exhaust sounds, and the bit rate is increased - still compressed, although only lightly and psychoacoustically-motivated). So there's still only idle, low and high speed samples, and no dedicated on- and off- throttle samples.

Even if they were authoring samples to these same light-weight, GT2 specs (so as to stay within the same hardware limitations), why haven't PD used the new recordings they clearly have? My theory is that their aim is to not use recordings directly in the game any more, in favour of some kind of "simulation" (at some arbitrarily low level). I'll mention LFS again, because it was my inspiration for looking into this in detail, and resulted in my building such a system for myself. If you use an external program to change the engine configuration of a given car, the game automatically changes the sound to suit, without samples in the usual sense; the same settings on different cars sound different. I've no doubt some of the cleverest minds in the world, who have a lifetime's professional experience with sound generation, should be able to come up with something like it in function, and to a much better quality.

...

btw, the red bull cars that Griffith claims to pinpoint some change in system to me just sound like some standard GT-esque loops that have been processed in-game with a distortion algorithm, and not a particularly pleasant one either. But the distortion does mask a lot of the negatives of the looping system (ie covers some of the stretchiness) as well as adds upper high frequency (12khz+) content that was otherwise dropped by the sampling rate and compression scheme, which is a noticeable improvement over the average car.

...
Have you driven them yet?

When all I had to go on were videos / recordings of the cars, I thought the exact same thing, and posted as much in this thread.

Incidentally, how do you use "filtering" to change the pulsetrain pattern for on- and off-throttle tones, and how do you animate that according to the control inputs in an emergent fashion?
Also, if it's just a "DSP effect", why is it only applied to those cars, and only when not in the data logger?
It's that "the simplest answer is usually right one" thing, except that you have to consider the whole system. If the simplest possible explanation for the part involves considerable personal invention to explain the whole, it's not so simple any more. The sound in GT has been something of a conundrum for some time.

Whether you use samples or not, it's still synthesis, a point you made yourself in the past. So the gamut of methods available are not constrained within any classical definition.

Have you played with LFS' sound modding capabilities? If not, I recommend getting the demo. ;)
 
Also, if it's just a "DSP effect", why is it only applied to those cars, and only when not in the data logger?
It's that "the simplest answer is usually right one" thing, except that you have to consider the whole system. If the simplest possible explanation for the part involves considerable personal invention to explain the whole, it's not so simple any more. The sound in GT has been something of a conundrum for some time.

Whether you use samples or not, it's still synthesis, a point you made yourself in the past. So the gamut of methods available are not constrained within any classical definition.

Have you played with LFS' sound modding capabilities? If not, I recommend getting the demo. ;)
If I am right in that it's simply a DSP distortion effect, it would be applied only to certain cars only for artistic effect, and dictated probably by a time limitation as to why other cars don't also have it. I am not privvy to how things work at PD so I can only assume based on what my ears tell me and what my experience in the industry suggests. As to why some gamemodes have it versus other, that's much simpler - for performance reasons, the DSP is turned off in data logger so that the data logging has more CPU to do what it needs to. This is analogous to visual "Level of detail" or LODding. I suspect that in the datalogger a lot of other CPU-intensive systems are reduced or removed as well, since running real-time logging and analysis is much more complicated than the systems used for replays (which are just saving a position and orientation of the car every so often).

I have played with LFS for a little bit but have never been able to produce, nor have I ever heard anyone else produce, something that sounded significantly better than what traditional looping systems produce. Perhaps there is an argument for switching to a traditional "synthesized" model from a cost or authoring time standpoint, but certainly not yet for a qualitative one. The technology for driving sinewave generators and noise filters has not yet gotten to a point that it can quickly and efficiently match what a real car makes.
 
If I am right in that it's simply a DSP distortion effect, it would be applied only to certain cars only for artistic effect, and dictated probably by a time limitation as to why other cars don't also have it. I am not privvy to how things work at PD so I can only assume based on what my ears tell me and what my experience in the industry suggests. As to why some gamemodes have it versus other, that's much simpler - for performance reasons, the DSP is turned off in data logger so that the data logging has more CPU to do what it needs to. This is analogous to visual "Level of detail" or LODding. I suspect that in the datalogger a lot of other CPU-intensive systems are reduced or removed as well, since running real-time logging and analysis is much more complicated than the systems used for replays (which are just saving a position and orientation of the car every so often).

Why make such a marked (dramatic, even) change in the aesthetic effect for so few cars? They really do stand out, especially in terms of the interactivity.

How could they not have time to twiddle with a few settings for a distortion effect on all the cars, when they haven't even made new samples for them? There are a lot of sound designers credited, and that number has grown since GT5P.

Why would the datalogger be more demanding than the in-race situation? It's just a standalone telemetry screen displaying replay data, and playing sound for two cars. Load up a replay of the 97T, and see for yourself how that distortion algorithm magically turns a flat-plane V8 sample into a V6 sound.

I have played with LFS for a little bit but have never been able to produce, nor have I ever heard anyone else produce, something that sounded significantly better than what traditional looping systems produce. Perhaps there is an argument for switching to a traditional "synthesized" model from a cost or authoring time standpoint, but certainly not yet for a qualitative one.

No, it never will sound like sampled sounds, because it uses filtered noise bursts. It's not really "traditional" synthesis, either, which is what I was getting at with the gamut thing.

Have you ever asked yourself how LFS can reproduce the sound of different engine configurations without changing anything in the game but the engine configuration? That's its party trick. The timbre issue is incidental and easily side-stepped (at great computational expense, for now) once you work out how it achieves that little trick (ingeniously logical and straight-forward). You can use it to modify samples to the same effect, too, although you retain the samples' general timbre if you don't create your own.


It's a trick that GT6's Red Bull cars seem to achieve also. The X2014 Standard and Fan sound exactly the same in the data logger, but in-game they clearly have very different exhaust manifold configurations. Of course, you can only hear that below 9000 rpm because the rasp effect saturates above that, and that part is identical between all the Red Bull and Senna cars. Have you driven them yet? :)
 
Vacuum cleaner is not a very technical term. We're having a serious discussion here. Some sounds are quite o.k, just missing some details (very important ones) such as intake sound. The Ralt RT3 sound is really good too and from what some of us can tell it's GENERATED (important word) in a different way.

God, this vacuum cleaner joke is old and annoying, go drive the stock Cizeta or the Ralt RT3 and tell me if it sound as a vacuum cleaner.

Peace.
Sadly, the vacuum cleaner thing is not a joke, it's real.



Very few cars sound barely ok in GT, the majority are a joke. Otherwise, the sound question in Kaz Q&A wouldn't get over 700 likes.
As much as I like GT, I'm not blind enough to put it on the throne, kneel and praise it for the sake of praising.
 
Does anyone still have the link of the video about samples being too stretched? From what I recall (and understand) a big issue was that the sound sample used at lets say 5k rpm was being stretched to be used at even 9,000, so what you hear at 9,000 is pitch distorted 5k rpm, not using two separate ones or more in order to save resources.

Idea is to post it in the comments section of the most recent blog post, if the said video was correct of course.
http://pitstop.gran-turismo.com/en/article/634

On a separate note, Griffith what do you think is the very best sounding game/sim out there? Is it raceroom, project cars?
 
Does anyone still have the link of the video about samples being too stretched? From what I recall (and understand) a big issue was that the sound sample used at lets say 5k rpm was being stretched to be used at even 9,000, so what you hear at 9,000 is pitch distorted 5k rpm, not using two separate ones or more in order to save resources.

Idea is to post it in the comments section of the most recent blog post, if the said video was correct of course.
http://pitstop.gran-turismo.com/en/article/634

On a separate note, Griffith what do you think is the very best sounding game/sim out there? Is it raceroom, project cars?
The video is in this thread somewhere, search for posts by Chippy569.

The Hudson is the only car where that isn't an issue: it has samples every 1000 rpm.

It depends on what you mean by "best". pCARS and Raceroom are just evolutionary arms of the standard approach, and are pretty boring as far as my interest is concerned. I really like what iRacing is doing, what with the drivetrain simulation really livening up the sounds it makes, and I also really like what the Red Bull and Senna cars in GT6 represent; on a similar note, I think LFS's sound is outstanding in its elegance and insight (not bad for a graphics programmer).

If your yardstick is absolute fidelity, PGR4 is still king. Nick Wiswell works for T10 now, but Forza doesn't sound like PGR yet. What's up with that?
 
Back