AC Modding Questions Thread

[Lonnopo]

That looks like it did the trick! I still need to test it out in the editor but I have a feeling it will be all good now. Thank you so much man, I feel like I never would have figured that out on my own, cheers!!

 

[Ghoults]

What re the usual causes when ai path recording fails 90% of the time? I am trying to record fast and pitlanes and most of the time I just get a 1kb fast_lane file. If I try to record pitlane I just get new fast_lane candidate? My track has multiple configurations and I've created the edge csv files in blender using all the necessary tricks. The edge files work fine but I just can't record ai lines. Somtimes it just works (but rarely) and the ai lines that are recorded work fine.

 

[LilSki]

You have all the sectors working properly? If timing isn't perfect recording might be sporadic. Also if you are using the shader patch for lights might be worth disabling it. There is always a chance it is interfering in some way.

 

[Ghoults]

Thanks for response. I don't have any shader patches. It is so far bone stock ac. My timing works and lap times are recorded. I don't know if it is "perfect" but I have not had any issues with timing gates. I also forgot to mention that sometimes I don't get any ai lines recorded at all. No candidate files. I can see the purple lines changing in-game on every lap tho.

Is there any difference to which session you use? I've used practice session. Is there any difference between content manager or using ac's own launcher?

Only special thing I can think of is that my s/f line is just like 5 metres after the pit exit. The pit road merges to the track after the s/f line. I think I could easily move the s/f timing gates back a bit if that could be the cause of the issue?

 

[Fat-Alfie]

Do you ever run wide whilst recording, going outside the track edge?

I have found that AC will sometimes fail to save any kind of line if I have run wide during AI recording (annoying when it's a 26km lap )

 

[Ghoults]

I have couple of places where I have wide runoff that the car goes over about half a car widths. I'll do some laps to see if ac records ai line even when I drive some laps in the middle of the road.

e: just checked, all drivable surfaces are set valid=1. First ai line attempt was ok, got a fast_lane. Deleted it and tried again just to see if it was a fluke and the game crashed as I went over the s/f line. Log has this error:

AI RECORDING STARTED
LAP VALID: 1 CUTS:0, COUNT: 1
OnLapCompletedEvent: 116763 bls:0Updating spline splineCurrent 2808
LAP VALID: 1 CUTS:0, COUNT: 2
OnLapCompletedEvent: 109188 bls:116763Updating spline splineCurrent 3060

CRASH in:
WAS LATE, HAD TO LOOP 19 times

This is weird. I have not had any crashes before...

 

[Ghoults]

Changed car, did the same thing. Drove 3 laps, turned ai recording on after exiting pits before turn 1. No crashes. Exited the game normally. No ai files created. In ac log I see this:

...
...
AI RECORDING STARTED
DSP: Car0 sp:184.0 pos:0.000 G:4 RPM:6413 pRPM: 7659 , limitRPM: 7500 dec: -32.38 -15.25
DSP: Car0 sp:226.8 pos:0.000 G:5 RPM:6323 pRPM: 7593 , limitRPM: 7500 dec: -68.31 -1.89
LAP VALID: 1 CUTS:0, COUNT: 1
OnLapCompletedEvent: 121982  bls:0Updating spline splineCurrent 2869
DSP: Car0 sp:223.3 pos:0.799 G:5 RPM:6244 pRPM: 7592 , limitRPM: 7500 dec: -25.21 -21.11
LAP VALID: 1 CUTS:0, COUNT: 2
OnLapCompletedEvent: 105046  bls:121982Updating spline splineCurrent 3063
WAS LATE, HAD TO LOOP 19 times
DSP: Car0 sp:181.1 pos:0.495 G:4 RPM:6199 pRPM: 7557 , limitRPM: 7500 dec: -7.61 -7.61
LAP VALID: 1 CUTS:0, COUNT: 3
OnLapCompletedEvent: 105845  bls:105046Updating spline splineCurrent 3057
WAS LATE, HAD TO LOOP 20 times
SETTING PAUSE MODE: 1
StartingLights pause mode changed: 1
Saving replay: C:\Users\...\Documents/Assetto Corsa/replay/temp/AC_191118-183143_O_ks_glickenhaus_scg003_problemtrack
closing pluginManager
DELETING PY PLUGIN tyre_temp
DELETING PY PLUGIN MiniChat
DELETING PY PLUGIN fuel_usage
km_per_liter: 1.655158 lap_eval: 2.904196  , evaluator: 2.898870 (guess: 30.000000), length:4806.902344
PCI: ks_glickenhaus_scg003 = 236.4
Saving telemetry channel LAP_BEACON with 33808 samples @ 100
Saving telemetry channel G_LAT with 6762 samples @ 20
Saving telemetry channel ROTY with 6762 samples @ 20
Saving telemetry channel STEERANGLE with 6762 samples @ 20
Saving telemetry channel SPEED with 6762 samples @ 20
Saving telemetry channel THROTTLE with 6762 samples @ 20
Saving telemetry channel BRAKE with 6762 samples @ 20
Saving telemetry channel GEAR with 6762 samples @ 20
Saving telemetry channel G_LON with 6762 samples @ 20
Saving telemetry channel CLUTCH with 6762 samples @ 20
Saving telemetry channel RPMS with 6762 samples @ 20
Saving telemetry channel SPLINE_POSITION with 6762 samples @ 20
Saving telemetry channel SUS_TRAVEL_LF with 67615 samples @ 200
Saving telemetry channel SUS_TRAVEL_RF with 67615 samples @ 200
Saving telemetry channel SUS_TRAVEL_LR with 67615 samples @ 200
Saving telemetry channel SUS_TRAVEL_RR with 67615 samples @ 200
Saving telemetry channel WHEEL_SPEED_LF with 6762 samples @ 20
Saving telemetry channel WHEEL_SPEED_RF with 6762 samples @ 20
Saving telemetry channel WHEEL_SPEED_LR with 6762 samples @ 20
Saving telemetry channel WHEEL_SPEED_RR with 6762 samples @ 20
SAVING SPLINE
BEST LAP POINT COUNT:3057
PITLANE POINT COUNT:3057
FILENAME: to trk folder/fast_lane.ai.candidate
Calculating radius
Building sides

CRASH in:
OS-Version: 6.2.9200 () 0x300-0x1
d:\dev\dev_pc_master_race\ksphysicsac\ai\aispline.cpp (701): AISpline::calculateNormals
d:\dev\dev_pc_master_race\ksphysicsac\ai\aisplinerecorder.cpp (218): AISplineRecorder::save
d:\dev\dev_pc_master_race\ksphysicsac\ai\aisplinerecorder.cpp (38): AISplineRecorder::~AISplineRecorder
ERROR: SymGetLineFromAddr64, GetLastError: 487 (Address: 00007FF713234B74)
00007FF713234B74 (acs): (filename not available): AISplineRecorder::`scalar deleting destructor'
d:\dev\dev_pc_master_race\ksphysicsac\track.cpp (44): Track::~Track
d:\dev\dev_pc_master_race\acs\trackavatar.cpp (218): TrackAvatar::~TrackAvatar
ERROR: SymGetLineFromAddr64, GetLastError: 487 (Address: 00007FF713166D14)
00007FF713166D14 (acs): (filename not available): TrackAvatar::`scalar deleting destructor'
d:\dev\dev_pc_master_race\ksgame\game.cpp (314): Game::deleteGameObjectRec
d:\dev\dev_pc_master_race\ksgame\game.cpp (314): Game::deleteGameObjectRec
d:\dev\dev_pc_master_race\ksgame\game.cpp (86): Game::~Game
d:\dev\dev_pc_master_race\acs\acs.cpp (478): wWinMain
f:\dd\vctools\crt\crtw32\dllstuff\crtexe.c (618): __tmainCRTStartup
ERROR: SymGetLineFromAddr64, GetLastError: 487 (Address: 00007FFFCE6C13D2)
00007FFFCE6C13D2 (KERNEL32): (filename not available): BaseThreadInitThunk
ERROR: SymGetLineFromAddr64, GetLastError: 487 (Address: 00007FFFCF6954F4)
00007FFFCF6954F4 (ntdll): (filename not available): RtlUserThreadStart

This was without the edge csv files. If I understand this correctly ac crashes on exit during ai line file creation because it can not create the edge files. I'll try with those edge files next.

I hope people don't mind if I spam this thread a little. If yes I can make new thread for this issue or move to different thread.

Trying to save pit_lane.ai and it creates fast_lane candidate which is same file as fast lane. Log shows this:

LOADING CSV: content/tracks/track/config/data/side_l.csv
LOADING CSV: content/tracks/track/config/data/side_r.csv
CALCULATING: 3050
DISPATCHING 2 threads
THREAD 0 from 0 to 2000
THREAD 1 from 2000 to 3050
FINISHED COMPUTING SIDES IN  103629.160108
SAVING SPLINE
BEST LAP POINT COUNT:3050
PITLANE POINT COUNT:3050 (WHY not use the lane I just recorded?!)
FILENAME: content/tracks/horsmav8/track/ai/fast_lane.ai.candidate (creates fast_lane...)
Calculating radius
Building sides
DEBUG TIMER: AISpline::calculateNormals:28.048247
BUILDING GRID FOR PERFORMANCE
SAVING VERSION 7
SAVING GRID DATA
FINISHED SAVING SPLINE

 

[Stereo]

Is there any "flat_nm" that's actually flat?

I was using 128 128 255 but knew it was wrong, went into the sdk and the Kunos provided one is 129 128 255 which is even worse.

There's a UV seam down the middle of this object due to how I unwrapped it (as it's underside, low priority to be seamless) and regardless which one I use it spreads the normals apart slightly.

I feel like mathematically the correct answer might be 127.5 which is obviously hard for a dds to contain.

 

[Fat-Alfie]

128 128 255 is correct, I think. A normal pointing directly at you would have an xyz value of 0, 0, -1

If x varies between -1 and +1, and Red 0 to 255, 0 would equal Red 128.
Same for y/Green.
z varies between 0 and -1, so Blue is between 128 and 255

128 128 255 seems right to me.

 

[Stereo]

It's not 128 though, cause the transform you do is 2*(value/255)-1. Which produces -1 when value=0, +1 when value=255. But produces 0.003921 from 128. (-0.003921 from 127)

To make 128 come out to half you'd need to do 2*value/256-1, which goes to 0.996 instead of 1 maximum (not really a major issue for normals; 0.996 0 0 normalizes into the correct 1 0 0)

 

[Amir Vodokotlic]

In process of creating simple test track.
and blender or ks editor is screwing me up.
I used in blender preset units meters.
Created everything to accurate measurements.
Exported and in game car was tiny like super super small.
Its only when I set everything to 0.010 scale in blender did it export correctly.
And everything was in proportion.

So how do I work in correct meter scale in blender without needing to work in 0.010 scale.

 

[Stereo]

In blender's fbx export you need to change its scaling by a factor of 100. Slightly depends which version of blender you use but usually it's just replacing 0.01 with a 1.0

 

[Amir Vodokotlic]

In blender's fbx export you need to change its scaling by a factor of 100. Slightly depends which version of blender you use but usually it's just replacing 0.01 with a 1.0

Its 2.79b.
If I leave the scaling factor at 1 and build everything in blender in meters, everything in game is massive.
Also do I leave that little thing beside scale in fbx export checked or unchecked?

 

[Stereo]

Yeah if you put 0.01 there you should be set. I'm not sure what the thing on the side does offhand.

 

[AccAkut]

Yeah if you put 0.01 there you should be set. I'm not sure what the thing on the side does offhand.

I never change the scale number, but I put the lil check thing, and it exports fine that way

 

[Amir Vodokotlic]

The 0.01 on export works great.
Allows me to model in meters in blender without needing to scale anything.

Now question on getting elevation.

Without LIDAR what other alternatives is there?

I tried Blender GIS plugin.
Good for google satellite image plane.
But Lacks precision when grabbing height data.

 

[Ghoults]

I'd try not to use settings where you need to multiply by 0.01 or 100 to get correct scale. Any time you change scale it is a possible issue that will hurt you later on. Here are my settings that work fine in metric:

 

[LilSki]

The units you work in have no effect on anything you export in blender. I work in imperial myself. Export settings do not matter which road you take. Whether it's 0.01 with Apply scale on all local or FBX All with it set to 1. Those two roads end at the same place without issue.

 

[GianmarcoFiduci]

Hi guys, I was wondering if anyone knows how to change tires size.
In some mods it happens that tire size in the .kn5 file looks right, but then ingame (or with the “align using data” tool in CM) they look smaller/bigger, wider/narrower.
How to fix this?

 

[Stereo]

Probably means there's some scaling on the WHEEL_FR etc. nodes. Should all be set to 1.0 before making the fbx.