Assetto Corsa PC Mods General DiscussionPC 

  • Thread starter Thread starter daan
  • 158,126 comments
  • 52,223,928 views
Just a small correction: I think the digit marked in red is missing from the [GEARS] section of the drivetrain.ini file in the data folder.

[GEARS]
COUNT=6 ; forward gears number
GEAR_R=-3.273 ; rear gear ratio
; forward gears ratios. must be equal to number of gears defined on count
GEAR_1=3.25
GEAR_2=2.18
GEAR_3=1.60
GEAR_4=1.30
GEAR_5=1.12
GEAR_6=0.96

FINAL=4.556 ; final gear ratio

Without that 1, the jump from fourth to fifth gear feels a bit weird.
1790973992030.webp
hmmm you might be right, oddly enough 1.2 is quoted by italian wikipedia, whilst 1.12 is quoted by spanish wikipedia

thanks for catching the error, what do you think of the car otherwise by the way?
edit: also what was on the Kunos car, for whatever that's worth, if you're seeing this, it has been updated to correct it on the original link
 
Last edited:
... thanks for catching the error, what do you think of the car otherwise by the way?...
To be honest, after noticing something odd happening in your version when shifting from fourth to fifth gear, I compared the data with the original Kunos version and spotted the error.

As for my impressions, bearing in mind that I’ve never driven that marvel in real life and I’m certainly no expert driver, I can say that the Kunos version is relatively easy to drive, whereas yours adds a bit of spice. For instance, your car becomes a little more unsettled when hitting the curbs, which makes the driving experience more interesting and possibly it comes closer to reality, perhaps...

I should also mention that I turned the FFB down quite a bit and disabled Real Feel; I’m not sure why, but if it’s enabled, my steering wheel becomes incredibly stiff. This happens to me sometimes with other mods that use Real Feel, though it might just be down to one of my settings, I’m not sure...

And of course... thanks to you for sharing!
 
Last edited:
To be honest, after noticing something odd happening in your version when shifting from fourth to fifth gear, I compared the data with the original Kunos version and spotted the error.

As for my impressions, bearing in mind that I’ve never driven that marvel in real life and I’m certainly no expert driver, I can say that the Kunos version is relatively easy to drive, whereas yours adds a bit of spice. Your car becomes a little more unsettled when hitting the curbs, for instance, which makes the driving experience more interesting.

I should also mention that I turned the FFB down quite a bit and disabled Real Feel; I’m not sure why, but if it’s enabled, my steering wheel becomes incredibly stiff. This happens to me sometimes with other mods that use Real Feel, though it might just be down to one of my settings, I’m not sure...

And of course... thanks to you for sharing!
for realfeel you need to give it the exact torque your wheelbase peaks at (don't quite remember what the default is), if you set it lower it'll be too strong.
generally, a lot of cars with realfeel don't actually have power steering so they tend to be pretty hefty on the limit, I'll double check how it's set up just in case though.
 
for realfeel you need to give it the exact torque your wheelbase peaks at (don't quite remember what the default is), if you set it lower it'll be too strong...
Yeah, 10Nm in my case, with a Thrustmaster T818, but only for a short time; I'm in the process of switching ecosystems.

In any case, I don't think it's a problem with the mods; as I said, I suspect I have some other setting that's making the steering wheel very stiff all the time when using Real Feel.
 
Last edited:
Gathers Drider Honda Civic 1998 Public Release!!

Download: https://www.mediafire.com/file/ioc0kh8524ds9pb/ek9_n1_Gathers_101creative.rar/file

Converted from GT7

Custom Dashboard built from scratch

Full AO inside & outside for improved depth and realism

Race-accurate interior details

N1 Spec Physics Package by Gravetrap based on real DATA

Authentic high-rev naturally aspirated power delivery

Reworked suspension geometry for proper N1 behavior

Stock gearbox ratios

CSP Optimized Settings

Driver Animation

On this car i did my best to make the car look was close was possible to the real car That is in the museum !



The Super Taikyu Series is an endurance racing series contested by race vehicles based on production cars and production racing cars. Its roots lie in the "N1 Taikyu Series" which started in 1990, and has continued to the present day with changes in the series name and regulations, from "Super N1 Taikyu" in 1995 to "Super Taikyu" in 1998.

Since the race is contested with race cars that are based on commercially available vehicles, it is a category in which the potential of the commercially available cars is directly linked to the race results, and the Civic, which has a powerful engine in a lightweight and compact body, has been competing in this race series with successive models doing well.

The Gathers Civic began competing in 1998. It was based on the first-generation Civic Type R (EK9 model) that had appeared the previous year in August 1997, but as it was built to N1 regulations, the scope of modifications was limited, with the body being strengthened and lightened, and the suspension and intake and exhaust systems being tuned.

The highlight of the Gathers Civic's full-scale entry in the 1998 season was its success in the 15th Tokachi 24 Hours race held that same year. Although the Tokachi 24 Hours is a race in the Super Taikyu series, it is a special race in which it competes alongside higher-category GT cars.

The Gathers Civic, with car number 77, was driven by three drivers: Yasuyoshi Yamamoto, Satoshi Tsujimoto, and Kaoru Shimamura, and competed in Class 4. In contrast to the cars in the higher classes that struggled with problems in the final race, the car showed off a fast run that surpassed the Skyline GT-R (BCNR33), a top-class car in the Super Taikyu Series.

Three GT cars took first, second, and third places overall, but the Gathers Civic took the checkered flag in an astounding fourth place overall, achieving the overall victory in the Super Taikyu class. Up until that point, the Skyline GT-R had achieved 50 consecutive wins in the Super Taikyu series, and this race embodied the Civic's unique style of taking on larger-displacement vehicles with a small-displacement engine.

As of 2023, the Gathers Civic is stored in the Honda Collection Hall within the Mobility Resort, and although it is not available year-round, it can still be seen in its original condition.

Unfortunately, photography of the engine bay and interior was not permitted at this Gathers press conference, but like other vehicles housed at the Honda Collection Hall, it has previously been seen for running confirmation tests and appears to be preserved in working order.

Honda Access currently supports Modulo Nakajima Racing, which competes in SUPER GT under the "Modulo" sports brand, and runs the Modulo NSX-GT. As it celebrates its 35th anniversary, we look forward to seeing another brand, "Gathers," racing on the circuit once again, as it continues to forge ahead in its history.

__custom_showroom_1790972314.webpezgif-7da85468494c8107.gifSem nome.webp__custom_showroom_1790972349.webp__custom_showroom_1790972341.webp__custom_showroom_1790972328.webp__custom_showroom_1790972319.webp
 
While i finally captured a proper lua coder to help me on my projects, i got caught off guard.

Meet Cobal, cause we're not joking anymore


1790979717406.webp



i love this one particularly, because
1790979743727.webp
is just :

"to compare if more overtakes took place in race A than in race B, we do (overtakes A) - (overtakes B)". Repeat for sidings and probably pace... What a


And there are pages of those. I love this. There's also quotes from psychology experts, statisticians, formulas from Mann-Whitney / Vargha-Delaney (researcher in quantitative psychology & psychometrics, and professor in psychology & statistic methodology) and many, many others.

The scientific paper is here, the main page Cobal

I'm not falling for that one - i still hesitate, out of curiosity mainly, but i've had my safe amount of desillusions for the year, and i want to keep some stamina for 2027.
 

Attachments

  • 1790980245322.webp
    1790980245322.webp
    32.3 KB · Views: 1
Last edited:
While i finally captured a proper lua coder to help me on my projects, i got caught off guard.

Meet Cobal, cause we're not joking anymore


View attachment 1568781


i love this one particularly, because View attachment 1568783 is just :

"to compare if more overtakes took place in race A than in race B, we do (overtakes A) - (overtakes B)"


And there are pages of those. I love this. There's also quotes from psychology experts, statisticians, formulas from Mann-Whitney / Vargha-Delaney (researcher in quantitative psychology & psychometrics, and professor in psychology & statistic methodology) and many, many others.

The scientific paper is here, the main page Cobal

I'm not falling for that one - i still hesitate, out of curiosity mainly, but i've had my safe amount of desillusions for the year, and i want to keep some stamina for 2027.
In other words, this seems pretty much to be Assetto Corsa's answer to Gran Turismo 7's Sophy AI.
 
In other words, this seems pretty much to be Assetto Corsa's answer to Gran Turismo 7's Sophy AI.

I wish. This is a complete hallucination where the Median Absolute Deviation formula is applied from nowhere, then discarded because it doesn't fit the normality hypothesis and you'd rather use non-parametric tests from Vargha Delaney/Hodges-Lehmann.

Am i familiar with half of those? Nope, but google is.


In the alternate reality (the one you and i are living), we're comparing 11 Kunos BMW Z' GT3s racing on a few laps with a lua script.

there's a difference between "oh on this test we got more overtakes with the script running and it kept the spacing tight"


and

"For race r, let Yr = (Yqr)q∈Q collect outcomes q, such as duel time, overtakes, physical spacing
and pace. A candidate c is described by the vector of race-level medianŝ
Y(c) =
(
medianr∈R(c) Yqr
)
q∈Q . (2)
The outcomes remain separate: a longer battle, an additional pass and a smaller physical gap
measure different contributions to racing. Reading them together connects competitive gains
with pace and safety and allows a candidate to be judged against its module’s objective and the
complete race"



Such a cosmic trip. Wouldn't bother me out of a test app on OT, but this is a patreon commercial.
 
"Modulo wheels" on the wings straightened out a bit for the Gathers civic, not changed anything else

rename the .txt to .dds and drag over

no need for a photo you get the point

i will probably go over this and do a 4k version soon
 

Attachments

Last edited:
While i finally captured a proper lua coder to help me on my projects, i got caught off guard.

Meet Cobal, cause we're not joking anymore


View attachment 1568781


i love this one particularly, because View attachment 1568783 is just :

"to compare if more overtakes took place in race A than in race B, we do (overtakes A) - (overtakes B)". Repeat for sidings and probably pace... What a


And there are pages of those. I love this. There's also quotes from psychology experts, statisticians, formulas from Mann-Whitney / Vargha-Delaney (researcher in quantitative psychology & psychometrics, and professor in psychology & statistic methodology) and many, many others.

The scientific paper is here, the main page Cobal

I'm not falling for that one - i still hesitate, out of curiosity mainly, but i've had my safe amount of desillusions for the year, and i want to keep some stamina for 2027.

Well, well, well, very interesting.
The idea of an AI that actually reads situations rather than just sticking to an ideal line is incredible.
Brilliant coding here for sure. But until it supports custom mods, tweaks or he eventually releases an automated end-user calibration tool...
 
I wish. This is a complete hallucination where the Median Absolute Deviation formula is applied from nowhere, then discarded because it doesn't fit the normality hypothesis and you'd rather use non-parametric tests from Vargha Delaney/Hodges-Lehmann.

Am i familiar with half of those? Nope, but google is.


In the alternate reality (the one you and i are living), we're comparing 11 Kunos BMW Z' GT3s racing on a few laps with a lua script.

there's a difference between "oh on this test we got more overtakes with the script running and it kept the spacing tight"


and

"For race r, let Yr = (Yqr)q∈Q collect outcomes q, such as duel time, overtakes, physical spacing
and pace. A candidate c is described by the vector of race-level medianŝ
Y(c) =
(
medianr∈R(c) Yqr
)
q∈Q . (2)
The outcomes remain separate: a longer battle, an additional pass and a smaller physical gap
measure different contributions to racing. Reading them together connects competitive gains
with pace and safety and allows a candidate to be judged against its module’s objective and the
complete race"



Such a cosmic trip. Wouldn't bother me out of a test app on OT, but this is a patreon commercial.
yeah, I'll believe it when I see it.

when you parse through the jargon and mathematical formulas, there's not a lot of substance here in terms of "what does it do". the videos on his site are not really convincing, frankly.

and it's an interesting decision, to say the least, to not release a public test version and just immediately paywall it.

Put me down as skeptical.
 
I have two questions that I'm seeking answers to today, if anyone can help.

1. Does anyone know how to get rid of the sound of the car being dropped off the jack when leaving the pits?
2. Is anyone else having video stuttering issues with today's new CSP build?
 
A small heads up for those unaware, Racedepartment has had a pair of bots reupload mods with viruses active at the moment:

image-34.webp


One community member explained the virus in detail:

"Do NOT download and try to install anything from this account! All of this person's "mods" have an autoinstaller.exe inside the .zip file which contains obfuscated code and is most likely an information stealer, but it could also be another type of malware. There is also another .zip file inside which is locked behind a password which is done to make it undetectable by antivirus software."

The infected files have been up for some 8+ hours now, keep reporting the profiles so hopefully something gets done!
 
time to break silence

here is some WIP stuff i'm doing rn

Honda NSX Concept-GT
Is undergoing extensive visual and shader refurbishment (still having 1.7m polys (what do you expect from gt models))
Some minor fixes and upgrades on physics
Remapped body uv (again) now it should be somewhat inviting for skinners, i hope some one would make a skin some day.

BMW M6 GT3
Bonny pointed out that my representation of his model lacked any soul investment and i agree i barely touched any shader stuff
So i'm working on rectifying that
Car is already looking way better than it was, perhaps i'd do the same to my classic gt3s

WIP Lexus RCF GT500
All booleans and subdivisions are complete, now it's time for material work which i find the most boring process of converting GT models

Also i made 2.0 992 GT3 R EVO available to download on my boosty
Screenshot_9nepe_honda_nsx_gt500_2016_tochigi_racing_ring_2-9-126-13-41-46.webp
Screenshot_9nepe_honda_nsx_gt500_2016_tochigi_racing_ring_2-9-126-13-42-44.webp
Screenshot_9nepe_honda_nsx_gt500_2016_tochigi_racing_ring_2-9-126-13-43-41.webp
1791017087787.webp
Screenshot_bm_bmw_m6_gt3_tochigi_racing_ring_1-9-126-17-50-45.webp

1791018086723.webp
 
Last edited:
X90
yeah, I'll believe it when I see it.

when you parse through the jargon and mathematical formulas, there's not a lot of substance here in terms of "what does it do". the videos on his site are not really convincing, frankly.

and it's an interesting decision, to say the least, to not release a public test version and just immediately paywall it.

Put me down as skeptical.
Oh it's absolute nonsense. Skeptical is the way, but i also have a true form of ... hate and fascination? for the academico buls****o -marketing used here. I've seen similar jobs on skin care studies, but it's the first time it applies to AC, so of course it grabbed my attention.

You...

Me...

View attachment 1568854

(Your message practically BEGGED for this meme to appear... 😙)
The whole paper just says "out of 300 tests with 11 Kunos Z4, there was 2,8 more time spent side by side for the AI than with vanilla ac, and the overtake success rate improved by 83%".


No need to bring social science experts or bloated mathematics in the process, this is gold paint on a band aid sold as surgery.


Now i've been playing with ac's ai for long enough to know that the Kunos Gt3s are not problematic : you can send them 25% gaspedal all lap round and they'll gain 2 seconds a lap with no issues. Add a random 25% offset to the left and they ll overtake themselves 15 times a lap : that's a 50kb lua anyone can try at home. It looks good, it feels better, but it's far from an actual scripted strategy - it just exploits vanilla ac slightly out of its safeties.


The hard part comes to maintaining a really competitive pace AND an actual racing behavior with, let's say , a URD Darche 992, or a lotus 49. Or both on the same race, on a faster pace than vanilla, while using their specific driving style. That would be some true achievment.



Yet, i'll get my hand on the script and see by myself.
 
Last edited:
... Skeptical is the way, but i also have a true form of ... hate and fascination?...
That’s the right attitude: it is thanks to it that humanity moves forward, venturing down paths less traveled, even if they sometimes seem a bit crazy.

Then again, I have to admit that if it were up to me, given how much I can't be bothered the hassle of trying new adventures, we’d still be living in caves... 🙄
 
why bother? got 1 already, same goes to the geezer from Russia. Pointless neither of you are beating RSS.
Who wants to beat the RSS? That’s not the point—and if you take a look, it’s a different model.
 
Who wants to beat the RSS? That’s not the point—and if you take a look, it’s a different model.
only win win for rss in that situation
is amount of skins for m6
 
why bother? got 1 already, same goes to the geezer from Russia. Pointless neither of you are beating RSS.
And what is the problem, even if is the same car/model, how many mods we have out there similar/equal!? And probably this car give another type of good emotion.
 
Thought I would post these for anyone interested. This is my version of a fictional Gazoo Racing Toyota_gr_gt3 factory team skins for multiple GT3 series, IMSA GTD Pro, Bathurst 12 Hour, GTWC, WEC & N24. I've applied what I believe to be the appropriate numberplates and decals, tires etc. as required by each series as well as a few graphical improvements on my previous version.
Now you can race in any of these series with this awesome new car.
https://mega.nz/file/DkhBlIzQ#bV4H_ir9CO9G9kgfdgryBOvxuEpAT7DMb4MtGKATRN0

View attachment 1568527View attachment 1568530View attachment 1568531View attachment 1568532View attachment 1568533View attachment 1568534
What car mod?
 
rfactor 2 conversation Sepang
Converted from rfactor 2.
Rainfx, Grassfix and lights, csp 1.79 or higher needed.
GP layout only.
View attachment 1473922

View attachment 1473923
View attachment 1473924
Download:https://mega.nz/folder/z8Il0KCK#sWqQ4VdeEDJGOVOJmhyJ8w
File:sepang_rf2_20250821.rar
Bug fix for Azure Circuit (Monaco) 2019
Fix some collision mesh bugs, change some texture, new ai line in hairpin.
View attachment 1473925

Download:https://mega.nz/folder/z8Il0KCK#sWqQ4VdeEDJGOVOJmhyJ8w
File:azure_2019_v2_20250821.rar
Screenshot_formula_extreme_26_v2_malaysia_3-10-126-21-25-20.webp

Straight mod is added to Sepang(rfactor 2 conversion). DRS mod is also remain as another layout.
Download:https://mega.nz/folder/z8Il0KCK#sWqQ4VdeEDJGOVOJmhyJ8w
File:malaysia_26.rar
 
Sharing a mod? Host it on GTPlanet Downloads. Free, public hosting for files up to 10GB in size.
Back