New version 0.8.9 is out!
- Thread starter frank11
- Start date
OK, the default HDR/CSM fs_filter's level of brightness change is HUGE!
What is pushing the brightness up vs a simple pass-through?
The bloom effect is added, so I guess it is the HDR > LDR calculation that is causing it to get so bright... any reason why everything is now brighter than before for like for like settings?
I'm just trying to understand the logic so I make thoughtful changes, rather than just drop my sky/sun diffuse/ambient track settings aimlessly!
Dave
What is pushing the brightness up vs a simple pass-through?
The bloom effect is added, so I guess it is the HDR > LDR calculation that is causing it to get so bright... any reason why everything is now brighter than before for like for like settings?
I'm just trying to understand the logic so I make thoughtful changes, rather than just drop my sky/sun diffuse/ambient track settings aimlessly!
Dave
Dave, I believe I've said it before, but one pretty good way of exploring what has changed, is using a file comparison utility, such as WinDiff (http://www.grigsoft.com/windiff.zip). Just select the old racer folder and new racer folder and do a diff. It will list all files that are new or have been changed and then you can view the differences in changed files.
Yeah, that kinda works ok.
As I suspected, it seems the HDR > LDR process has changed a little...
There is still lots of development type stuff in the shaders which makes reading it hard. There are about 7 methods in hdr.cg that could be used/referenced from the fs_filter cg file to do the HDR > LDR scaling... so I'm not sure what changed.
I'll have to ask Camsimmy because he wrote it, but there is a shader I use lots that is not working so nicely in this latest version. It's based off the burn_f/v for the track or tarmac etc, but instead of using the red channel to burn, it uses an alpha channel, or DDS's 4th channel.
We will have to try ask one of the Racer devs to add it, or create it, and maintain it, because using the red channel for specular is really not very nice at all. Ideally I'd prefer an rgb DDS with a pair of secondary 8 bit channels, one for specularity strength, the other for specularity sharpness, then we can really get some decent looking surfaces to our tracks!
Dave
As I suspected, it seems the HDR > LDR process has changed a little...
There is still lots of development type stuff in the shaders which makes reading it hard. There are about 7 methods in hdr.cg that could be used/referenced from the fs_filter cg file to do the HDR > LDR scaling... so I'm not sure what changed.
I'll have to ask Camsimmy because he wrote it, but there is a shader I use lots that is not working so nicely in this latest version. It's based off the burn_f/v for the track or tarmac etc, but instead of using the red channel to burn, it uses an alpha channel, or DDS's 4th channel.
We will have to try ask one of the Racer devs to add it, or create it, and maintain it, because using the red channel for specular is really not very nice at all. Ideally I'd prefer an rgb DDS with a pair of secondary 8 bit channels, one for specularity strength, the other for specularity sharpness, then we can really get some decent looking surfaces to our tracks!
Dave
EoD, you're clearly not understanding my point, I understand what you're saying and I did read all your posts, but if Ruud and Mitch won't let the people who are pushing the boundaries of what we are doing have an alpha, I doubt they'll give it to someone with an ATI card to iron out kinks.
In the end it's really ATI's fault anyway, they're the ones who left OGL dev behind in favour of DX.
I'm glad to see you've cheered up a bit, it's not nice to have so much negativity at these forums
Anyway,
Dave, that shader shouldn't be too much to worry about, you'll have to catch me pretty early tonight though (my time).
I'm still a little worried about the sky rendering. Ruud, how far away is that procedural sky you mention on racer.nl? Is it just a dream still or are you working on it yet?
The smoke particles should NOT be that colour, they look fine at midday but that's about the only time, ambient light is screwed up in the current calculations. This NEEDS to be fixed BEFORE we can get any decent, repeatable outcomes in racer, there's no point mucking around trying to get things looking right if it's in fact wrong for when the correct sky/environment comes out.
In the end it's really ATI's fault anyway, they're the ones who left OGL dev behind in favour of DX.
I'm glad to see you've cheered up a bit, it's not nice to have so much negativity at these forums
Anyway,
Dave, that shader shouldn't be too much to worry about, you'll have to catch me pretty early tonight though (my time).
I'm still a little worried about the sky rendering. Ruud, how far away is that procedural sky you mention on racer.nl? Is it just a dream still or are you working on it yet?
The smoke particles should NOT be that colour, they look fine at midday but that's about the only time, ambient light is screwed up in the current calculations. This NEEDS to be fixed BEFORE we can get any decent, repeatable outcomes in racer, there's no point mucking around trying to get things looking right if it's in fact wrong for when the correct sky/environment comes out.
The smoke doesn't seem to match the scripted "time" command whence the confusion. It does match the time when the track first loads. I don't know why it doesn't get the right lighting after that. But what looks ok in the evening looks funny when you change the time to noon and the smoke stays bright orange. Maybe I'll take a stab at the smoke shader.
Intuitively it does seem wrong to me to ignore regular lighting completely because "particles are flat". It should at least use the normal atmosphere.cg routines so it's consistent. Since the smoke is billboarded maybe its normal direction can still be used to check diffuse, since you're seeing one side of the smoke cloud. Just increase the relative weight of ambient light since it's softer. But something I've done has made the game crash again
Postscript: Is there an easy pacejka hack to make a "burnout machine"? Something with very little traction on tarmac I guess.
and before getting further into smoke, maybe it should factor in tire temp when spawning the particles. I'm not sure if Racer handles tire temps yet. But it has to be pretty hot to create the thick clouds of smoke.
edit: Figured out what was making it crash. I can't use "Select a car" to select a car, as soon as I click one it crashes. First one shows up ok. Back to using Quick Race I guess.
Intuitively it does seem wrong to me to ignore regular lighting completely because "particles are flat". It should at least use the normal atmosphere.cg routines so it's consistent. Since the smoke is billboarded maybe its normal direction can still be used to check diffuse, since you're seeing one side of the smoke cloud. Just increase the relative weight of ambient light since it's softer. But something I've done has made the game crash again
Postscript: Is there an easy pacejka hack to make a "burnout machine"? Something with very little traction on tarmac I guess.
and before getting further into smoke, maybe it should factor in tire temp when spawning the particles. I'm not sure if Racer handles tire temps yet. But it has to be pretty hot to create the thick clouds of smoke.
edit: Figured out what was making it crash. I can't use "Select a car" to select a car, as soon as I click one it crashes. First one shows up ok. Back to using Quick Race I guess.
Yeah the smoke particles are wrong anyway, and where does the "wind" come from? a track's special.ini? Is this a constant or an expression?
Smoke particles should use bump mapping too, you could get away with using half the particles and have it look even better.
Currently I just don't have the time to play around with stuff like I used to, real life is creeping in too much and I just can't devote hours to figuring out problems that pop up.
Racer has SO MUCH potential...I just wish I had the time to unleash it
Smoke particles should use bump mapping too, you could get away with using half the particles and have it look even better.
Currently I just don't have the time to play around with stuff like I used to, real life is creeping in too much and I just can't devote hours to figuring out problems that pop up.
Racer has SO MUCH potential...I just wish I had the time to unleash it
I must admit some of the graphic bugs I'm getting with 0.8.9 are entertaining.
Changed some racer.ini settings (renderer.bloom.enable=0) and now when I open up the throttle, it draws the car's gui at the origin and it is massive.
LMFAO
it reminds me when i tried to emulate ps2 in my pc, using gt4 rom..
Got trees and fences to work using alpha to coverage with standard_transparent v/f shaders, not tree v/f shader as in Carlswood, they look very good now.
Got a problem with the tower on my 2nd Creek track which was made with a single tga image. The rail casts a shadow if using standard v/f shader but it's not transparent, no holes. Using the s_t v/f shader I get the holes but no shadow, go figure!
Still having problems with skies on very large track. Using flag=65 it doesn't cover the whole track!
We need better info about making skies and gettin them to work right and then have Ruud leave them alone!
FPS on carlswood= about 176.
Got a problem with the tower on my 2nd Creek track which was made with a single tga image. The rail casts a shadow if using standard v/f shader but it's not transparent, no holes. Using the s_t v/f shader I get the holes but no shadow, go figure!
Still having problems with skies on very large track. Using flag=65 it doesn't cover the whole track!
We need better info about making skies and gettin them to work right and then have Ruud leave them alone!
FPS on carlswood= about 176.
Fiddled with the shader and stuff a bit, right now it's really a kludge but it doesn't look like a dust storm anymore. The smoke.influences.turbulence doesn't seem to have an effect on the dispersion of the smoke, unless I'm misinterpreting it. (I would consider turbulence a random change in velocity in this application)
What the smoke needs is a smooth fade to nothingness, or what I'd call "conservation of smoke":
As the particle gets larger, it should also get more transparent. Since each particle spawned is some physical amount of smoke, it should probably drop off proportional to the area covered by the particle. Or, growth^-2. Then it can naturally disappear.
If the initial "quantity" of smoke is determined by the sliding (more slip/hotter tire = more smoke per particle, ie spawning with higher alpha) then for the regular growth rate it can fade naturally into invisibility without needing an explicit life span.
Also have a question about the LightingSun function:
It takes lightColor as an input and outputs diffuse.
Why is the baseCol then multiplied by both? (baseCol*diffuse*lightColor)
Am I misinterpreting what diffuse means, or should it already be taking into account the color of the light source, and not need to be multiplied again.
The smoke.influences.turbulence doesn't seem to have an effect on the dispersion of the smoke, unless I'm misinterpreting it. (I would consider turbulence a random change in velocity in this application)What the smoke needs is a smooth fade to nothingness, or what I'd call "conservation of smoke":
As the particle gets larger, it should also get more transparent. Since each particle spawned is some physical amount of smoke, it should probably drop off proportional to the area covered by the particle. Or, growth^-2. Then it can naturally disappear.
If the initial "quantity" of smoke is determined by the sliding (more slip/hotter tire = more smoke per particle, ie spawning with higher alpha) then for the regular growth rate it can fade naturally into invisibility without needing an explicit life span.
Also have a question about the LightingSun function:
It takes lightColor as an input and outputs diffuse.
Why is the baseCol then multiplied by both? (baseCol*diffuse*lightColor)
Am I misinterpreting what diffuse means, or should it already be taking into account the color of the light source, and not need to be multiplied again.
Attachments
Skies have become a little tatty...
The definitions are also misleading. Diffuse and ambient settings are really sun intensity and sky intensity, for all intents and purposes. They should be seperated.
I'd prefer to see things defined like...
sun
{}
sky
{}
The sun has an on/off switch and when switching it removes all the sun influences, same with the sky...
A fully procedural sky would be cool, even for now if it's just an option to turn on and it just creates a nice gradient sky (cloudless)... we can then have some presets for mie/rayleigh curves vs TOD, ie, foggy climate, clear climate etc etc to simulate deserts or rain forest environments etc.
Again, Maxwell render uses a system like that and is really nice, generating HDR's if you like... could be useful if anyone wants one... give me lat/long/TOD and I'll render a few out for you
Dave
You might have anti-aliasing turned off? For shadowmapping (HDR/CSM) you need for example in motion_blur:
samples=2
coverage_samples=8
Without it, alpha_to_coverage can't dither its pixels (no internal higher-resolution buffer) so the alpha 'blending' (really coverage) can't work.
I posted a download for my Surfaces and Sounds track, just the .ini files and it works quite well. Can't get the tree to sway in the breeze like it did in ver. 0.8.8.1. It needs to combine the animate and standard_transparent cg's, imho!
Most setting for the sun are close, just remember the sun is 93 million miles away not 93 meteres.
Most setting for the sun are close, just remember the sun is 93 million miles away not 93 meteres.
You can't really remove Cg all the way from the standard racer.exe. I did manage to get the non-Cg compilation working again, try: http://racer.nl/download/racer089_nocg.zip (just racer.exe).
That screws up the graphics but allows you to work on physics and such.
It's worse than normal "non-CG" 0.6.x . Ruud can you use "engine graphic" from 0.8.8.1 and others from 0.8.9? Additionally version for old gpu "graphic" from v0.6.5.5 and other "news" from 0.8.9 ?
Graphics... , why "time" don't work on replays?
Special "validation code" for the best lap would be very useful for check honesty during many competition.
Fuel consumption don't count clutch.
Hmmm,
One problem I have found is that the sun 'bloom' is visible on objects that are in the shade but would reflect the sun if they were not in the shade.
This is only apparent on objects with the envmap, where the bright reflection of the sun spot is rendered in the shaded area, and then a bloom is applied.
I can see why this is happening, and I can't think of an elegant way to remove it... I do think removing the bloom could be possible and reduce the impact of these 'bright' areas of reflection in shaded zones though.
Would there be any way to make the sun-spot come later in the pipeline, so shade is considered when it is rendered to the envmap?!
Bag of worms!
/runs for the hills
Dave
One problem I have found is that the sun 'bloom' is visible on objects that are in the shade but would reflect the sun if they were not in the shade.
This is only apparent on objects with the envmap, where the bright reflection of the sun spot is rendered in the shaded area, and then a bloom is applied.
I can see why this is happening, and I can't think of an elegant way to remove it... I do think removing the bloom could be possible and reduce the impact of these 'bright' areas of reflection in shaded zones though.
Would there be any way to make the sun-spot come later in the pipeline, so shade is considered when it is rendered to the envmap?!
Bag of worms!
/runs for the hills
Dave
Thanks alot Ruud, I really appreciate that you made an effort to provide this! I did some test-runs on my computer at work which also has an non-cg Intel GPU, and it ran just fine. The graphics may be "old-fashion" by the newer Racer-standards, but since I can't use cg it looks as it always has..
Attached is an image of the test-run on Win7-64bit with a 800x600 screen (I am at work, can't start running car-sims at fullscreen
) The image is just to show that it runs with the proper ini-settings.Again, thanks Ruud, I think more users than myself will enjoy this.
Edit: Confirmed working on WinXP-32bit as well.
Attachments
So just from an authoring point of view, if you turn off Cg, does it just revert to the old shader system?
Ie, if we provide a cg car.sh, and a non-cg car.shd, in theory it should all work ok from a downscaling point of view?
Just trying to consider those using older versions and supporting them.
I totally agree that supporting Cg in a scaled fashion is a bad idea, FAR too much mess, but supporting Cg fully, or reverting right back to the old system if not, seems a good compromise. It still looks good with good content
Dave
Ie, if we provide a cg car.sh, and a non-cg car.shd, in theory it should all work ok from a downscaling point of view?
Just trying to consider those using older versions and supporting them.
I totally agree that supporting Cg in a scaled fashion is a bad idea, FAR too much mess, but supporting Cg fully, or reverting right back to the old system if not, seems a good compromise. It still looks good with good content
Dave
I think so, yes. I did these days one car without cg (i dont use cg here, i dont like nfs graphics too much, bloor, brightness. nah). And raphael did the cg shader to it, and worked fine too, just changing some textures and the car.shd.
changing the focus, were the needles bug fixed in 089?
Latest News
-
How The BTCC and Motorsport Games ReunitedIn a surprise move last week, official BTCC content will once again be present within rFactor 2...- Thomas Harrison-Lord
- Updated:
- 5 min read
-
Sponsored MOZA Racing & Lamborghini Redefine Racing Boundaries with The Real Race Super Trofeo 2024MOZA Racing proudly announces its collaboration with Lamborghini for the launch of The Real Race...- OverTake.gg
- Updated:
- 6 min read
-
Forza Horizon 5 Apex Allstars Update Adds Five New Race Cars, Plus Yet More DLCA new Apex Allstars Festival Playlist kicks off this Thursday in Forza Horizon 5, joined by a...
- Thomas Harrison-Lord
- Updated:
- 3 min read
-
F1 24's Updated Car Performance and Handling ExplainedWhile we haven’t tested it yet, on paper it sounds like significant changes have been made to...
- Thomas Harrison-Lord
- Updated:
- 3 min read
-
Tim Jarschel Wins 2024 DTM Esports Pro championshipCombining searing pace with consistency, Jarschel took home the title despite a charge from...
- Thomas Harrison-Lord
- Updated:
- 3 min read
-
Embracer Group Splits Into Three – Wreckfest, Milestone, Set For New HomesTwo racing game publishers are now forging their own paths as Embracer splinters into three...
- Thomas Harrison-Lord
- Updated:
- 3 min read
-
Recreating The Goodwood Members Meeting In Assetto CorsaThe Goodwood Members Meeting is one of the highlights of the British motorsport calendar. A...- Connor Minniss
- Updated:
- 5 min read