WIP Sears Point Raceway (Infineon Raceway) Sonoma

[mianiak]

There are 2 types, there are breaking bumps before each corner, there are some sharp crack bumps where the different asphalt meets. Then there are physical bumps modelled into the mesh. ie, I think the most pronounced one is the bump where the indy track turns off the main. You can see it here,

http://youtu.be/5_UJKA8AwlA?t=6m35s

[ed] sorry for pasting the link as a code but the forums were forcing the video to be embedded and I just wanted to link it to start at a certain point.

 

[mianiak]

I uploaded a new build with all the fixes and sent you guys pm's with the link. (at least I think I fixed all the fixes )

 

[R Soul]

Tiny gap where the pit exit meets the track:


Here follows a list of things:
I looked at the Full layout.

Still some DX8 problems. This time when it reported a missing texture I put in a dummy file and tried again till I got the full list:
Jumbo screen material problem. It doesn't mention a specific problem just "Error loading global material SCREEN"

Error loading texture GRASS_ADD for material GRASS00

Error loading texture INF_TRACK_DRAGSTRIP_S for material DGST

Error loading texture INF_TRACK_DRAGSTRIP_B for material DGST

Error loading texture INF_TRACK_DRAGSTRIPTAPER for material DRAG_SKIDS

Quick fixes:
Your inf_wall objects are all HAT targets, but they shouldn't be.

PitBoxes are still HAT targets.

ptent objects (trucks) are HAT targets.

ice cream van: HAT target, not lit at night.

inf_bridge_py: HAT target.

inf_buildings06, inf_buildings03: HAT and collision target.

dodgemarena: HAT and collison target

LightPole objects: HAT and collision targets. Some of them are hidden behind walls. For those that are exposed I think you should use invisible poles for the collisions.

inf_outterrain00 and inf_outterrain01: HAT and collision targets.

* I don't know how the game deals with these things but I think there should be as few HATs and collisions as possible.

Object suggestions:
Your inf_tires objects don't use the Collision tag. I recommend you place invisible low-poly collision walls over them. That should stop cars going through the tyres without inducing physics bending behaviour.

The inf_infence objects dont have the collision tag. I recommend more invisible low-poly collision walls.

inf_terra12 could be split up so that the chequered flag floor and the concrete (material rdcp00) are separate objects.

inf_outterrain01 is a big object. Not as many polys as I thought, but it still might be a good idea to split it up.

 

[gwaldock]

Hi mianiak and krivjur,

I agree with R_Soul regarding unnecessary HAT targets. They may be the primary cause of your track's somewhat low FPS average.

That said; I ABSOLUTELY LOVE THE "FEEL" OF YOUR TRACK! Yes, the bumps are present; there are plenty of them, they are not too intense and they seem to be in all the right places. IMHO once the "eye candy" issues have been addressed yours may be the DEFINITIVE version of Sears Point / Infineon Raceway simulation.

Thanks again for your efforts!

 

[mianiak]

Ok, I'll fix the material issues because that's crucial, but the rest I will leave for now. I have a good reason for doing this I just can't say what that reason is. My only goal with this for now with rf1 is to just get it out in the state it is in so people have the track to play on.
I know that's not the right thing to do but I have to stop working on this for a little while, It's either I wait for another few months or release it now how it is. As long as it is in a state where it works without any loading errors, then I will be happy with that for now. At a later date I can come back and fix up all these other things.

[ED] I uploaded a patch

 

[mianiak]

Ok hang on a minute, I went back and checked the hattarget thing and saw some big mistakes so I will sort that out too. There is a reason why I set the walls to hattarget=true, that's because at other times I have had problems getting stuck in wall mesh if its not a hattarget. But I will give it a go and see how it works out.
This fix will be only a small update, just all the scn files.


[ED] I did a test with fraps.

As it is
Frames, Time (ms), Min, Max, Avg
7523, 60000, 102, 163, 125.383

With walls and such set to HatTarget=False (after deleting the hatfile to be sure a new one was made)
Frames, Time (ms), Min, Max, Avg
7495, 60000, 102, 164, 124.917

There is not much difference.

The track is 'next gen' its not a low poly 512x512 textured track from 2005, it uses large textures and quite high detail objects. this would be why the fps seems low, on my system I'm over 100fps, my goal was to keep it around 100, I think that is good enough. If it gets too low for some people then Auto Detail FPS would be in need.

But at any rate, there are some objects that are a hattarget that shouldn't be, I will fix them up and leave it at that

 

[Emery]

I still have the peculiar garage framerate slowdown. Otherwise, as before, framerates are okay. I've tried always rebuilding HAT, not rebuilding HAT, and Auto Detail FPS without an improvement. In checking other intensive tracks, the problem does not appear. ATI 5870 w/2 GB GDDR5, 2560x1600 resolution, 3 GHz Core2 Duo.

So far have only noticed AIW problems:

- No lap/sector times in Test Day.

- Apparently you have to maintain pit lane speed until T2 now? I get fined for speeding if I romp on it anywhere between the lights and T2.

 

[mianiak]

Emery, can you try commenting out inf_bld_suncogas and test it (cause i never got around to building a collision and shadow object for that), then try commenting out dragskids and test it (cause that uses a big shader), also try inf_outterrain01 (cause thats the hill in the background). See if any of those removes the fps hit.

I don't mean to get you working .but I cant seem to get it to drop, it all is fine for me.

Thanks

 

[Emery]

mianiak - I'll do that tonight. This is such a wonderful rendition of Sears Point that I enjoy ensuring it is as good as it can be. If no one else is experiencing this problem, I suspect it is a hardware/software/settings combination at my end, but it's weird that only Sears Point shows this effect.

 

[krivjur]

So far have only noticed AIW problems:

- No lap/sector times in Test Day.

- Apparently you have to maintain pit lane speed until T2 now? I get fined for speeding if I romp on it anywhere between the lights and T2.

Not really an AIW issue. ;-)

There's small issue on SEARSPOINTGMT.mas on the 0.2.4.4 patch which by error contained Drift sectors. On 0.2.4.3.7 the main sectors on that .mas are for the main 3 layouts, plus overriding xpitin on Full and IndyCar folders. There will be small change on how they will be packed in future.

 

[Emery]

Not really an AIW issue. ;-)

True enough! I get into the habit of thinking of these things (start, finish, sectors, xpitin/xpitout) as they are in the BTB workflow rather than how rF treats them as objects.

 

[gwaldock]

Don't fret too much about fps mianiak. The high level of performance and immersion that I experience on your track would not be possible with a "slide show" frame rate; your graphics refresh quickly enough to maintain a very smooth flow (on my system).

 

[Emery]

Sadly, the fps issue remains for me even after commenting out the objects.

[had an idea...]

Hah! I figured it out (sorta). No fps issues when running windowed, so it's got to be something with either full-screen mode or the graphics options (resolution, ansio, or AA) when I run it on my computer. Weird. Hadn't thought to try it in windowed mode before.

 

[mianiak]

There is one thing that could be possible. I have found a drop when I round the hairpin, at that point the whole track swings into view. It's a hard track to lod things out of because you can see nearly everything all the time. But I could lod some things out around T4 and adjoining areas for that stretch between the hairpin and T1.

 

[R Soul]

Have you considered making low poly versions of some objects so that you only get the full versions at, say, 400m?

The long distance objects could also use materials with simpler shaders.

Oh and the gass_add texture is still missing for material grass00.

 

[Erik Tveit]

Hmm, the track is fantastic, I haven't discovered a bug yet. I found FPS drops when going out of the hairpin, yeh. But I thought that was normal with my laptop.

 

[mianiak]

All High poly objects have a low poly lod which kicks in between 200 and 400m for different things, sofar I have broken almost everything up into 1000 poly sized objects (thats 2000 tri's) except the tires which I still have to do. What isn't broken up are the garages which are 12k tri's each and the grandstand which is only 4k tri's. I work on a comparison with car models, most lod_a car models are 20k tri's (which is generally referred to as 10k poly's), more modern models can go up to 40k tri's. But its not always 40 cars with 40k tri's each, because mostly you see the lod's of the cars when your on the track. So I kind work on an estimated average around that when it comes to poly sizes of objects.

I think the issue with making a lod with different material will bring up 2 problems. 1, an extra material and 2, surfaces that don't match up properly. If your using bump and spec then lod it into just a diffuse mat, it wont look the same.

With everything I have done to fight this fps thing, I have successfully gained 10fps at the exit of the hairpin, that might not seem like much but on my system, that puts the lowest frame rate at about 90 fps (haven't done a proper test yet). So if I was running vsync on my monitor that runs at 60hz, I wouldn't notice any drop at all.

That bloody grass! haha . What your seeing is just a small part of this problem, it has haunted me throughout this entire project, but it's ok, I know exactly what the issue is. One thing with 3dmax that can cause issues with the materials is if you edit a material, you have to export every single object that uses that material, if you miss one object that uses that mat then it wont work. In comparison, simed is very forgiving when it comes to editing materials, in simed you can edit the mat on one object and it will work on all. This issue is only apparent because this project has been a learning project. At the start I did things differently to how I did things in the end. The next time I build a track, these sort of issues wont come up because from doing this track I have learned the way to do it right from the start.

Just a note..
Earlier I fell into the trap of rushing things and as your all aware, I made mistakes. So I have taken half a step back and am going through all these issues. I have decided that there is no rush to release this.

I'm working on an update, I should have it ready for you guys soon.

Thanks!

 

[gwaldock]

Cut yourself a little slack mianiak. I mentioned earlier in this thread that I own GPL, rFactor, GTL, GTR2 and Race07/GTR-Evo; I also have copies of all of the Nascar series, the NFS series and others. My point is that I have NEVER known of a dev to release a track with no defects! You blame yourself for making "mistakes" but the truth is that ALL programmers (even those that are making money for their work ) release products that need to be "debugged"; many "finished" tracks have more severe issues than the minor bugs in your "beta".

 

[mianiak]

It's all good. It's not really hard to do it right, i was just rushing too much and not thinking about where I was going

Here are the latest benchmark results. I set fraps to benchmark for 90 seconds, which is almost a lap if you start it as you are going along pitlane. I let the AI drive so it would be a constant.

Started just as I went over the SF line in pitlane, ended just before the SF line.
Nascar circuit.
Max settings in RF.
All the bells and whistles on in the Nvidia control panel.
I used the Corvette mod from Niels and Siim using the hood view.

Frames,    Time (ms),    Min,    Max,    Avg
12081,    90000,        99,        216,    134.233

Here is the recording for every second.

01: 114
02: 115
03: 120
04: 128
05: 131
06: 127
07: 124
08: 113
09: 114
10: 115
11: 109
12: 126
13: 139
14: 150
15: 163
16: 166
17: 152
18: 141
19: 134
20: 123
21: 117
22: 119
23: 125
24: 128
25: 133
26: 154
27: 151
28: 154
29: 150
30: 140
31: 142
32: 143
33: 143
34: 144
35: 153
36: 159
37: 161
38: 167
39: 154
40: 138
41: 151
42: 189
43: 210
44: 201
45: 183
46: 149
47: 120
48: 108
49: 107
50: 108
51: 105
52: 107
53: 103
54: 102
55: 104
56: 111
57: 116
58: 113
59: 122
60: 116
61: 115
62: 118
63: 122
64: 119
65: 119
66: 128
67: 135
68: 124
69: 115
70: 101
71: 106
72: 115
73: 119
74: 136
75: 154
76: 159
77: 164
78: 167
79: 152
80: 196
81: 207
82: 152
83: 137
84: 122
85: 119
86: 120
87: 118
88: 115
89: 113
90: 109

 

[mianiak]

Just an information update guys, I have all the great feedback you guys have given me. I am going through all those things and decided that there are some other things that I would like to do too, like rebuild the walls and fences because in real life they are different to what I have. In real life, they are mostly all temporary barriers and are not all aligned perfectly. (ignore the bad stitch in the pit entry pic)




Depending on when rf2 is released, which should be very soon, I might switch this project over to rf2, then maybe come back to it for rf1 after the rf2 version is done.

Thanks for all your help, the 4 of you have done really well in finding the issues you found and I very much appreciate it.