Gran Turismo 6 modding discussion

Have you figured out how to assign different AES configurations / files to different exhaust parts? Including the manifolds and turbo conversions?

I've been saying this for literally years, a decade I think actually, but the potential of this system is off the charts. Just wait 'til you figure out how to get the most out of the ESGX mixing in conjunction with the AES exhausts ;)
 
Griffith500
Have you figured out how to assign different AES configurations / files to different exhaust parts? Including the manifolds and turbo conversions?

I've been saying this for literally years, a decade I think actually, but the potential of this system is off the charts. Just wait 'til you figure out how to get the most out of the ESGX mixing in conjunction with the AES exhausts ;)
Click to expand...
I'm not sure if it's possible to have them for different exhausts, but there's always the chance I'm just doing it wrong.
 
TheAdmiester
I'm not sure if it's possible to have them for different exhausts, but there's always the chance I'm just doing it wrong.
Click to expand...
I can't remember how the sound allocation works to be honest. Different exhausts have different sounds, that much is obviously true. It's possible the samples are referenced differently from the AES files, since the only cars that use AES cannot be modified (in ways that would change the sound).

My experience doesn't go very deep with all of this, I relied on more capable individuals. I do know my way around engine sound synthesis, though.
The sounds followed the engine in the case of "engine swaps", I seem to recall. So I'd say that would be a good place to look.

On that note, you should try a 350Z with one of the Red Bull V6 sounds. One of them, I can't remember which, is absolutely perfect for the job!
I also implore you to try a 5 cylinder sound.

Message me if you want any help with the signature sounds of certain engines / cars, e.g. I worked out how to do the W16 ;)
 
Griffith500
I can't remember how the sound allocation works to be honest. Different exhausts have different sounds, that much is obviously true. It's possible the samples are referenced differently from the AES files, since the only cars that use AES cannot be modified (in ways that would change the sound).

My experience doesn't go very deep with all of this, I relied on more capable individuals. I do know my way around engine sound synthesis, though.
The sounds followed the engine in the case of "engine swaps", I seem to recall. So I'd say that would be a good place to look.

On that note, you should try a 350Z with one of the Red Bull V6 sounds. One of them, I can't remember which, is absolutely perfect for the job!
I also implore you to try a 5 cylinder sound.

Message me if you want any help with the signature sounds of certain engines / cars, e.g. I worked out how to do the W16 ;)
Click to expand...
Great to see you've got some knowledge of this stuff, W16 stuff would certainly be interesting but I think I'll wait until I've experimented more myself - right now I've only properly experimented with up to 10 cylinders and a tiny bit on a V12.

350Z was actually one I was struggling to nail down but I'll check the Red Bull sounds to see if any sound like a close fit. As for the exhausts, the way it works is that in the carsound folder, you have this structure:

\aes (cars with AES sounds)
\engine (the "front of car" sounds)
\normal0 (stock exhaust for NA cars)
\normal1, normal2, normal3 (stages of exhaust tune)
\turbo0 (stock exhaust for turbo cars)
\turbo1, turbo2, turbo3 (stages of exhaust tune)

However not every car uses a "0" sound - the Cien for instance actually uses the same sound as the Zonda C12 models, but it comes with a sports or semi-racing exhaust by default, so you never hear the V12 wail of the Zonda unless you mod in a stock exhaust for it.

Also as you can see, AES doesn't have any numbered folders. I've never tried making an AES1, 2, etc yet, but my assumption is that a car with AES has the same sound for every stage of tune.

I5, V6, and V10 seem to trouble me the most as it's hard to find a good balance for them. Straight 6 is pretty easy and I've managed to get a vaguely Supra-sounding note out of one, but the rest are a bit bothersome. Regardless, here's my best attempt at an I5 so far:
 
Ah yes, I remember now. If you look in the AES files themselves, they also have entries for ESGX volume, or similar. That's the samples blended in at the same time, but the volume is usually set very low. Not sure if the specific sample set is also defined there as well.

I still think the engine definitions are the best place to look. Otherwise you're left with getting creative with the samples on top of AES and manually swapping out / modifying the sounds as and when.

Keep experimenting, it's a lot of fun! That 5 cylinder isn't far off, just a bit lumpy is all. The manifold ("extractor") numbers are too dissimilar. Also pay attention to the firing order. That is the key to the 350Z sound, and the differences in all the straight sixes. The system perfectly differentiates them all with the right numbers fed to it.

I found my old files so let me know if you get stuck!
 
Griffith500
Ah yes, I remember now. If you look in the AES files themselves, they also have entries for ESGX volume, or similar. That's the samples blended in at the same time, but the volume is usually set very low. Not sure if the specific sample set is also defined there as well.

I still think the engine definitions are the best place to look. Otherwise you're left with getting creative with the samples on top of AES and manually swapping out / modifying the sounds as and when.

Keep experimenting, it's a lot of fun! That 5 cylinder isn't far off, just a bit lumpy is all. The manifold ("extractor") numbers are too dissimilar. Also pay attention to the firing order. That is the key to the 350Z sound, and the differences in all the straight sixes. The system perfectly differentiates them all with the right numbers fed to it.

I found my old files so let me know if you get stuck!
Click to expand...
Firing order is one I’ve been paying close attention to but the inext and exext values are the ones that throw me off since I don’t really know what the abbreviations mean.

I know that messing with them generally seems to throw off the cylinder timings which can be good or bad depending on what you need, but seeing as you referred to them as extractor, are you able to shed any more light on them and their effect?

I’ve noticed the best effects seem to be when they’re a multiple of the delay (e.g on a 5cyl, a cylinder delay might be 72 or 144, and 144 or 288 sounds good for exext) but other than that it’s mostly guesswork.
 
For the five cylinder, try {20, 10, 10, 20, 0}. Assuming the crank angles just increase as {0, 144, 288, 432, 576}. No guesswork needed :D

Extractor is just another name for an exhaust manifold.

The intake channel is not operative, so ignore those numbers, unless you can get it working :)


EDIT: For the 350Z, try 57998, all the other V6s are uneven firing (like an early Buick V6, Metro 6R4 / XJ220, Alfa Giulia QF, turbo-era F1 etc.)
 
Last edited:
Griffith500
For the five cylinder, try {20, 10, 10, 20, 0}. Assuming the crank angles just increase as {0, 144, 288, 432, 576}. No guesswork needed :D

Extractor is just another name for an exhaust manifold.

The intake channel is not operative, so ignore those numbers, unless you can get it working :)


EDIT: For the 350Z, try 57998
Click to expand...
Will try those on the 5cyl sound now, thanks. The main question I have is how do you know which manifold values correspond to good sound? For instance it's useful knowing that 20, 10, 10, 20, 0 is good for 5cyl, but why is that the case?

Also, for the 350Z, do you mean changing the engine ID within the AES file? I figured that was simply a duplicate of the AES file number, but if it's actually to point it to an ESGX sample then this would make a lot more sense.
 
Aha, well that's a trick of the trade shall we say. It's all to do with the manifold and firing orders, as I said.

Whatever you were already doing to swap AES sounds onto non-AES cars, just do that, but using the ca and exext data in 57998 (the AES file).

EDIT: On closer inspection of the numbers, it wouldn't yield the perfect warble. Try {22, 25, 10, 12, 0, 3}. The more files you look at the more it'll start to make sense, but I don't think PD would thank me for putting up a tutorial here...

EDIT 2: Good Skyline-like sound: {30, 0, 30, 10, 20, 10}

(Sorry for all the edits, adding stuff as I find it)
 
Last edited:
If we're talking about idle lope, that unstable "cammy" (or "porty", if we're talking two strokes and rotaries), then no. The game engine does not support it, so the idle rpm will be always be rock solid and, since the sound is synthesised directly off the engine speed, so will the resulting sound with AES.

Depending on the engine configuration, you do get some oscillations, but this is simply a direct result of the wave action in the exhaust.

Obviously, samples can have a fluctuating speed recorded into them and that's something that can be looked at with mods.
 
I've already applied to the above via DM but I'll just mention here in case anyone wonders:

I'm more than happy to release the tool, but there's no public way of editing the game's databases right now. Editing these databases is required for the events you create to show up, so without this there's no point in releasing my tool. If/when things change I'll definitely clean it up and release it.
 
A little preview of some upcoming stuff.

Gran+Turismo®6_1.png


Gran+Turismo®6_2.png


unknown2.png
 
Ha, I tried that 350Z preset out myself just the other day (after a bit of a break at it) and I'm not sure why it's not quite right, in terms of the base texture. Seems like I've always been doing VQs wrong, like it defies explanation how the real thing sounds the way it does. Yours sounds better than mine though, which has me thinking it's more a matter of resonances than harmonics 👍

I tried a McF1 sound that was just intake (like this onboard) and although the physical model can manage very raw intake-like sounds, the system is not really set up for it in terms of the filters and input to get the overall response correct. The pure exhaust note is otherwise a little dry on the real thing. Overall it's remarkably comparable to an S54 (especially in CSL trim) in this respect, just without all the rasp.


After a bit of a false start (and temporarily dashed hopes and dreams), I also found that it's quite easy to have the standard engine sound samples play alongside the synthetic exhaust :)
 
It's definitely handy not to underestimate what the exhaust parameters can do do the sound. I've had a lot of cases where I thought a sound wasn't great but it turned out to be the exhaust sound rather than the cylinder timings.

The sound sample thing is very interesting, how does it work? Blending the two, or having it change sounds for exhaust upgrades, etc?
 
If it possible to add new cars, albeit using existing meshes, is it possible to hide cars from the dealership? I'd love to see all the dupes hidden.
 
Some nice developments in the time since my last posts.

I've developed a tool that lets me input torque curve data in a certain format (from real data or other games as long as I can get the progressive numbers) and it'll automatically do all the messy database stuff for me. I've also given it the ability to add a car from some key data and copy tuning parts from a similar car, so the tuning menus and upgrades all work for added cars.

One of the latest developments I'm especially pleased with, thanks to Nenkai, is the ability to actually give them names now:
upload_2020-10-19_0-31-35.png


Also I've realised that since we can award cars for events, it would be a very handy way to add some pre-tuned (e.g. engine-swapped) cars without spoiling the stock-only nature of the dealership. For example:
GranTurismoC2AE6_1.png


I won't spoil all of them but I plan on adding cars of a similar nature and sprinkling them in as reward-only cars.

Also a little bonus for cool points:
upload_2020-10-19_0-33-39.png
 

Latest Posts

Back