Poll: how many cars would you trade for a better game/engine sounds?

  • Thread starter Thread starter danielc
  • 105 comments
  • 5,817 views

How many cars would you trade for better engine sounds/game/features?


  • Total voters
    282
Where is the misinformation? Blame translator san for wording, but it was a direct quote from kaz's interview.
Unless i missed the big sounds update?
 
I hope that all standard cars remain in GT7 and then say in GT series for 20 years or more. And further more removing all or some standard cars from the game to get better sound that is not going solve the problem.
 
The question makes no sense.
Getting rid of standards will not affect audio development.
Car modelers=/=audio engineers/dudes.
You're totally missing the point. If all the cars stay this would mean about 700 to 800 engines would have to be recorded, and seein as there is no hope of that ever happening, it would mean that about half the cars would still sound, as 99% of gt planets members would say, ****. If they take cars out, the consistency and quality of the game is greatly improved = better game = less crying = everyone's happy = question DOES make sense.
 
GT3 had what 150 I think we can deal with 200-300 with better quality of everything vs. again 1200 cars and a new track that looks like you're literally out in the middle of nowhere in Ascari(that's how I view it) it was pretty pathetic even if it's so famous.
People are forgetting that gt6 already has over 400 premium cars lol, and you are absolutely correct about quality over quantity. I mean forza 5 makes me drool, I'm gutted I don't have an Xbox one. the lack of cars doesn't hurt me, What hurts me is that gt doesn't have the same level of quality. Because of this, saying bye bye to half of the cars would be the easiest choice I could ever make :D
 
If they take cars out, the consistency and quality of the game is greatly improved = better game = less crying = everyone's happy = question DOES make sense.
except some of the most liked cars in the game are standards and then people will complain about "Taking my favorite car out but upgrading this car that I'm indifferent about though many people like"
 
Where is the misinformation? Blame translator san for wording, but it was a direct quote from kaz's interview.
Unless i missed the big sounds update?

You appear to have a lot of misunderstandings about how games are made in general, and specifically about sounds in games. Those misunderstandings lead to you to individual conclusions that fall just slightly off the mark, but the compound effect is to be well wide.

I'm not going to put anyone down for that, so don't think I'm attacking you, but what results is a lot of hearsay and general acceptance of frankly wishful thinking, and to comments like that from @GranTurismo guy there: that the number of cars are what is currently holding PD back with the sound update.

There is no reason to think that is the case at this late stage, although it is a reasonably large obstacle and was likely part of the motivation for them to use a "totally new" method of generation (in real time, as you play). As I said, recording and generally doing the same production processes a few times more when you've already done them hundreds of times does not constitute a "breakthrough".
 
All the rarely/barely driven hybrids/eco-cars/whatever you call 'em, the cars that don't really have a valid purpose in a racing game, those can be sacrificed for things like a premium MK.IV Supra RZ, true premium RUF's, the overdue inclusion of a BMW E30, and so many other cars. (that I could keep going on & on about, until GT7 actually hits shelves)

My quick 5-second estimate puts the number of those cars in GT6, somewhere around or above 100 that I'd be willing to sacrifice, so go with that. (voted 'under 300', not 'under 100')
 
You appear to have a lot of misunderstandings about how games are made in general, and specifically about sounds in games. Those misunderstandings lead to you to individual conclusions that fall just slightly off the mark, but the compound effect is to be well wide.

I'm not going to put anyone down for that, so don't think I'm attacking you, but what results is a lot of hearsay and general acceptance of frankly wishful thinking, and to comments like that from @GranTurismo guy there: that the number of cars are what is currently holding PD back with the sound update.

There is no reason to think that is the case at this late stage, although it is a reasonably large obstacle and was likely part of the motivation for them to use a "totally new" method of generation (in real time, as you play). As I said, recording and generally doing the same production processes a few times more when you've already done them hundreds of times does not constitute a "breakthrough".
when at any point did I say car quantity is holding sound back? I said that deleting cars improves consistency and quality....
 
You're totally missing the point. If all the cars stay this would mean about 700 to 800 engines would have to be recorded, and seein as there is no hope of that ever happening, it would mean that about half the cars would still sound, as 99% of gt planets members would say, ****. If they take cars out, the consistency and quality of the game is greatly improved = better game = less crying = everyone's happy = question DOES make sense.
eSZee has got a very good point post #3 👍.
 
Last edited:
Frankly, i don't care how, all i want is a multi million dollars franchise to sound better than ipad games on ipad speakers. Pd has a budget bigger than any other car game and has more years in business than most.

All you talk about is sounds this sounds that, disregard your technical knowledge for a bit and play the game. otherwise this thread , sadly, is not for you.

I have 2 companies and i am constantly evaluating my weakest links and trying to improve them, if for 5+ years the biggest complain was the same, heads would fall, and the chop would be used with no mercy.... Not even my best friend/cousin/brother would have their job if it was the case. Not to mention a potential partnership with someone to improve things.
Take yamaha for example, they know about sounds and engines, AND is a japanese company too. My secretary would be on their ass until i got what i wanted. Budget or fame is not scarce in pd's case. It would be a big plus for yamaha(or whoever) portfolio if they were the ones that fixed gt sounds.

pd won't necessarily do what i am proposing and like you said, its not a real trade.so pretend you be dumb like us for a second, cast your vote (or don't) and move on. This thread is about a ficticious trade-off using cars as currency, don't overthink it.

If by now you are not getting this, pretty please, make a new thread explaining everything you want us to know about sounds, the way pd did it, the way they are doing, the way they will do it. post lots of videos, links, etc, even make a poll...

Last but not least what is real may not work in video games, sometimes an exagerated feature translates better to simulate what driving a car at full chat means. So maybe the other car games might be right and pd wrong. Ever think about that?
 
Last edited:
eSZee has got a very good point post #3 👍.
I get what he's saying too, but the fact is you can't record 800 engine sounds it's just too big of a job. So we would still have cars sounding awful. Gt being half assed (in this area anyway) is what started all the problems years back, do we really want them to continue? But yes, the car modelling doesn't affect the improvement of engine sounds directly.
 
Frankly, i don't care how, all i want is a multi million dollars franchise to sound better than ipad games on ipad speakers. Pd has a budget bigger than any other car game and has more years in business than most.
And most would agree, as do I.
All you talk about is sounds this sounds that, disregard your technical knowledge for a bit and play the game. otherwise this thread , sadly, is not for you.
Regarding the superficial value of the sounds in the game, I have the same disappointment as anyone else - have had since GT4-ish, I'd say.

However, what "technical knowledge" I do have means that I can read into what I'm hearing and come up with apparently crazy theories, one or two of which have turned out to be true (PD's focus on simulation, a totally new sound generation method to sidestep all of the shortcomings of sampling, limitations / delay, even before the official announcement, being technically based and not logistically etc.)

Crucially, I take care to consider the evidence and to explore and also explain all the possibilities, which is why I post: sound is generally treated as a "known quantity" and "easy", when in reality it is far from either.
I have 2 companies and i am constantly evaluating my weakest links and trying to improve them, if for 5+ years the biggest complain was the same, heads would fall, and the chop would be used with no mercy.... Not even my best friend/cousin/brother would have their job if it was the case. Not to mention a potential partnership with someone to improve things.
Take yamaha for example, they know about sounds and engines, AND is a japanese company too. My secretary would be on their ass until i got what i wanted. Budget or fame is not scarce in pd's case. It would be a big plus for yamaha(or whoever) portfolio if they were the ones that fixed gt sounds.

They are most certainly working with Yamaha. That's because they pretty much announced it when talking about the LFA before GT5 released, and also because my ears, and "technical knowledge" tell me so (the Red Bull cars).
pd won't necessarily do what i am proposing and like you said, its not a real trade.so pretend you be dumb like us for a second, cast your vote (or don't) and move on. This thread is about a ficticious trade-off using cars as currency, don't overthink it.

I don't really see how it is any different to other threads, then. I thought maybe the idea was to actually discuss the implications of making compromises in large software projects, from a gamer's perspective. Instead, it's just the same stuff "I think they should do this" with no factual underpinning at all. Not a problem, it just surely doesn't warrant its own thread (except for the poll, which doesn't actually say very much, mostly due to the last option).

If the idea is to let our imaginations run wild and come up with our ideal game, then sure. But imagining "compromises" as though they might actually have been feasible and lamenting PD's not having taken them surely is not the same thing. Either we discuss the practicalities as well, or this is just a fantasy thread.
If by now you are not getting this, pretty please, make a new thread explaining everything you want us to know about sounds, the way pd did it, the way they are doing, the way they will do it. post lots of videos, links, etc, even make a poll...
There is already a sound thread.
Last but not least what is real may not work in video games, sometimes an exagerated feature translates better to simulate what driving a car at full chat means. So maybe the other car games might be right and pd wrong. Ever think about that?

Yes, psychoacoustic concerns are very important. I use them to make shortcuts all the time, either in terms of "development" time or processing power or simply because nobody knows / understands the underlying physics. But there are still massive improvements to be made for games overall, and they will start trickling out over the coming years thanks to GPGPU and heterogeneous computing paradigms in general. PD were playing an incredibly long game with this aspect, so in time it should all make much more sense.



The quality of the entire game I mean.

Ah, but your original post, the one I referenced, was only about sounds. I agree that there is something to be said about the level of quality PD chose with the Premiums, and the knock on effects of that. However, it's clear they were playing the long game, and so only time will tell as to whether it was the right choice (unfortunately that doesn't help us now).

As for removing the Standards, that doesn't change much in terms of any trade on the final gameplay result, since they cost PD nothing (ignoring licensing effort etc.) to include in GT5, virtually.

GT6 is different with the improvements in some of them, which can only be a response to the complaints, such that those which have been earmarked for Premiumification will roll over into GT7 (or beyond) whilst everything else will be dropped, I'd wager. (But, then again, I thought they'd drop the Standards for GT6, but that was assuming they'd make new cars only, and not touch up the Standards at all).
 
I cannot make this clearer, FANTASY THREAD HERE. ok?
But anyway... How better or worse is an ipad comparing to the ps3? Graphics and sound wise?
And could you share wich games have you worked on, so i can see(hear) some of your magic?
You touched a good point, in time, its past the time...this should be in the pipes by gt4 time, they had the whole ps3 development cycle and more. With the physics they have, worse graphics and better sound would have made this one of the most well balanced games of all times(engine wise)
Ps3 is dead, Ps4 is now born and we still got what we got. I end up having more fun playing other games, ones that cost as much as a dlc pack(and it sounds better). Gt6 just got my money this time because they said the 2 magic words, Ayrton and Senna... After less than a month, despite awesome physics its gotten so boring, i am spending much more time in the forums than playing! Wich means they will probably not see my money next time around, (unless many things happen)
This is a forum, and it sure does allow me to say almost whatever. If is against aup, and gtp forum rules, the mods would have closed it long ago. Here is where we can play games(not gt, think long trip in a car "games" ) talk serious, talk crap, news and so on...
It differs in the reverse way of my proposal, instead of what they should do, what would you be willing to loose and for what.


The trade in final gameplay(i know its meant for someone else) is inconsistency on the eye and ears of the person who bought the game. And to some its a big hurdle to overcome. This game sells world wide and local cultural aspects influence on this matter. So respect other peoples opinions like we should extend the same courtesy... I don't know if intentional or not, your posts sound very authoritary and patronizing, since i have not heard the new sounds, you sound like a pr for the sound department. Try writing as if you were teaching a class or maybe a letter to yourself.
Redbul jr sounds good indeed but its not nailed and still is well under par.
Edit
I forgot to ask, with this new real time method, do they have to do this for each car or is a one-size-fits-all deal? Because if they have to do it individually, less cars means they would spend less time or more time in each perfecting it further.
 
Last edited:
In that case, I think all 1200 cars should be Premium and unique, with trim options, and Sony should upgrade our PS3s to PS4s for free, just for being such awesome customers, so the new sounds will just work as they are without further optimisation. This is the fantasy where I'm 100% right, rather than merely "confident", of course.


There is only one video of "my work", it's in the sound thread; I'm almost loathe to link to it, but it's here. I'm a hobbyist (literal "amateur"), and I make that abundantly clear. I've spent years, on and off, in my spare time and only achieved a fraction of the functionality and the fidelity that PD already have in their incomplete models, as demonstrated by the Red Bull cars (which are just exhaust sounds).

There is unlikely to be any difference in making loops from recordings and tuning their new system to match recordings, in terms of time spent. The main advantage is that you can tune outside of your recording library and get believable, more or less "correct" results with great ease (think sounds for modified cars, where there is actually less pressure to match a given recording - it could even be automated, to an extent). My favourite thing is being able to look up pictures of engines and getting my synth to sound like recordings (texturally speaking) based on those pictures alone.

It's a powerful way of looking at sounds, but one that requires a great deal of work to unlock properly and have decent fidelity, at least initially, in terms of developing the systems. Making sounds for individual cars is, as I said, not really that different, complexity-wise, but it is very different, more like tuning physics for the cars.
 
I don't know how hard it was to make it, but sounds "computerized" , switching between notes(is that the correct term?) doesn't sound alive, if you can relate/understand my feeling. I also wouldn't like to have gt cars sounding like that.
Why is it like that? Just because of the sample size being stretched and compressed to climb as rpms would in diferent gears? How would you(or someone) solve it for "aliveness" feel?
Nevermind what the graphs are about, just ilustrating...
Feels like its
sloped-output_graph.jpg


Instead of

Line.h4.jpg
 
Last edited:
Double post but for a reason.
Sometimes i think you have a little add... You never answer the on topic question i am asking you.
In that case, I think all 1200 cars should be Premium and unique, with trim options, and Sony should upgrade our PS3s to PS4s for free, just for being such awesome customers, so the new sounds will just work as they are without further optimisation. This is the fantasy where I'm 100% right, rather than merely "confident", of course.

So i would like you @Griffith500 to answer me this.


What are you willing to throw out
in the game for
Something you want in game?
 
I don't know how hard it was to make it, but sounds "computerized", switching between notes(is that the correct term?) doesn't sound alive, if you can relate/understand my feeling. I also wouldn't like to have gt cars sounding like that.
Why is it like that? Just because of the sample size being stretched and compressed to climb as rpms would in diferent gears? How would you(or someone) solve it for "aliveness" feel?

If you're talking about my video, it's band limited and poorly mixed for a start, has awful aliasing (noise) from "quantisation error" and other things I'm inexpertly doing to butcher the sound via initial processing, before synthesis (I cleaned up the real-time part a lot, but it's still heavily coloured). Only the exhaust has any effects applied (and is deliberately "flat", like a "perfect" exhaust manifold, which pretty much never exists), and they're constantly changing as I tweak and try new things (there's no resonance in that version, for example, I broke it...); the intake is "raw", so the quantisation and interpolation are very audible as a sort of fizzy square-wave sound (there's also some "numerical dispersion" in there, too, probably :dopey:).

There are no samples, as such, certainly not in the usual sense. To solve "the problem", I would write bespoke software that can give me the precursors I need (clean, full-fat) and "solve" my serious generation problem (which causes the volume fluctuations you're hearing, as well as I/O glitching) for better / smoother revving - currently a sticking point, then I'd spend a lot of time learning how to harness advanced physical modeling and DSP techniques to get "better" / nicer real-time processing. I'd also like to get control data at a higher resolution than 60 Hz (limitation of tools provided with LFS), which implies writing my own interfaces at the very least. I don't have any of that ability, and I only taught myself what I can do.

The Red Bull cars' sounds show evidence of PD having solved the "generation problem" (they're so smooth!), and they have much nicer sounding precursors and processing going on that should sound interesting when not "overdriven". That is the point of the video, ultimately, to show how (in my opinion) PD have overdone the sounds to compensate for there only being exhaust sound. From the description: "this video is not intended as a quality demonstration of the synthesis method itself".



To answer your question, what I really want is a completely different game. But chances are, I'd have to make it myself - I don't have that ability, either. In that sense, GT can be GT, and other games can be other games, and I'll continue to mix and match them to get closer to my actual wishes.
 
Gotcha!
I just know when someone can make an amg v8 , lexus v10 or a zonda v12 sound like it does in real life, it will be a great day for gamers around the world... Meanwhile in a quick google search, i found
 
Gotcha!
I just know when someone can make an amg v8 , lexus v10 or a zonda v12 sound like it does in real life, it will be a great day for gamers around the world... Meanwhile in a quick google search, i found


It will indeed!

I won't claim to understand what is going on in that video, and I've seen it before, too. I don't really know what fuzzy logic is; I expect it's something to do with "shades of grey" between the two "binary" (on/off; true/false etc.) options you get with "ordinary" logic (yeah, no clue). I know, I'm on the internet, but I'm feeling particularly lazy right now.

Anyway, it's still sample based, and I really don't like samplers any more.


This is good:



They're still samples, but there's a ton of them, which goes a long way to alleviating some of the shortcomings, but there are diminishing returns in terms of control complexity and variety / "expression". It's also a lot of data, with quite a bit of redundancy when you really look into it; like "panoramic" photographs stitched together with excessive "overlap".

Most interestingly, they are not taken from recordings - they are 100% synthetic samples. That is more-or-less the level of quality that will eventually be possible to generate in real time, only with more expression and fewer actual controls than any equivalent sampler. They're probably using an automotive-industry simulation tool, something like Ricardo WAVE, possibly with a custom renderer, or other hot-rod options - maybe they wrote their own solver (but validating that is a royal pain) since, e.g. WAVE, is waaay too "feature rich" for just sound-production duties. That's what I'd like to (be able to) do.


And this:



It uses a combination of "pure synth" techniques, i.e. only generative rather than what I would call "reactive", or "physically based". It's likely that PD have combined the two approaches with their new technique in some way, again going by the Red Bull X-cars and my own experiments.

And now your thread is massively polluted, sorry! :dopey:
 
What i liked about the fuzzy logic thing is the lower graph, showing when and what was being played. But the bad thing it needs 10 samples for a single car...
But can i use one of these (1st video) in my ps3? All ferrari sounds would be funny as hell... Lol (jk, i know its obd2 to sync rpm to the sample)
Edit, i saw a obd2/cigarrete lighter one, wich is easier to use
 
I would trade all of the standard road cars for premium interiors in all of the standard race cars... and premium sounds.
 
What i liked about the fuzzy logic thing is the lower graph, showing when and what was being played. But the bad thing it needs 10 samples for a single car...
But can i use one of these (1st video) in my ps3? All ferrari sounds would be funny as hell... Lol (jk, i know its obd2 to sync rpm to the sample)
Edit, i saw a obd2/cigarrete lighter one, wich is easier to use

That graph could be drawn for any sampler. The real meat of his innovation is in dynamically controlling the blending based on only a few inputs, apparently "translated" using fuzzy logic. I prefer to use the physical inputs only, rather than fudging unrelated "inputs", e.g. "load", from other unrelated inputs. Physical models take more readily to this reduction in controls also, for obvious reasons.

The number of samples is about right for today's games, but aesthetically that's actually too few. That's why granular techniques have become popular, despite the fact they sound weird at constant rpm - they only use one long sound sample (less RAM overall). Games get around the constant speed weirdness by preventing such situations from happening - in GTA 5, for example, you can only be accelerating or decelerating, it won't let you keep a perfectly constant speed.


This is a granularly re-synthesised engine sound from a single acceleration "ramp" recording.


It's unlikely that you could hijack the sounds in a PS3 game, even with some clever tricks and access to the data files etc. Sonory makes the sounds for the Automation game, however, they're still only samples (and their audio engine wasn't very well set up last I checked - it also serves to demonstrate just how important the control is, e.g. no game accurately models tiny rpm fluctuations, except iRacing, and it's massively important). If they can find a way to convert / translate their process to real time, they'll be onto something massive. That's what's got me excited about the sound update for GT.
 
. I prefer to use the physical inputs only, rather than fudging unrelated "inputs", e.g. "load", from other unrelated inputs. Physical models take more readily to this reduction in controls also, for obvious reasons.
Aup requires english posts only... Lol

What do you mean physical imputs? Like throttle position?
 
Aup requires english posts only... Lol

What do you mean physical imputs? Like throttle position?

Exactly that, yes. I use throttle position, boost and rpm for the main controls. One is player controlled, the other two should fall out of the physics engine (if it's separated from the sound "simulation" - one day they will merge for some games, e.g. in the way iRacing's drivetrain simulation affects rpm).
 
Back