Racer v0.8.39 released

Ruud

RACER Developer
Finally something new!

Get the latest at http://www.mediafire.com/file/03gcsac51pd955u/racer0.8.39.7z (52Mb)

The changelist:
- car.ini susp<n>.k_factor now influences regular springs. Also works with the system ('set car0.suspension0.k_factor 0.5')
- 'doc scriptfuncs' and 'doc scripttypes' now output HTML (more readable).
- TrackEd: added checkbox to turn on/off colorizing of object flags in Properties (F4) mode.
- TrackEd could display bad colorize flags
- TrackEd now saves LOD Z information in geometry.ini when saving (used to throw it away)
- TrackEd's model information now displays information on all LOD models.
- Dof_Fix could report continously optimized indices; this was due to every object being loaded to get tangents.
Now, tangents are only created if a shader is attached that requires tangents (tangents=1 in .shd files).
- Camera offsets fixed when moving the CG for from/to cameras and fixed-type cameras (see http://www.racer.nl/tutorial/car_cameras.htm)
- Fading of menu->race->results->menu restructured internally.
- Add Lidar section to TrackEd (not that you normally would have Lidar data).
- Increased the number of supported DOFs in a track from 5000 to 10000.
- Render engine supported number of blocks (geobs) in view increased increased from 3000 to 5000.
- 'debug <n>' command extended to support a subpage, i.e. 'debug 6 2' for the Onyx profile view.
- Images/textures can now be loaded up to 8192x8192 in size
- racer.ini's textures.max_wid/max_hgt increased from 2048 to 8192
- Ghost cars now load their shaders from the same car directory, only using car_ghost.shd instead of car.shd.
No longer do you need a separate data/cars/<car>_ghost directory.
- Added 'reload onyx'. Onyx is an relatively complex scripting language that is much like C++, and will be used
in the future as an interface for external physics, or other places where script overriding might be useful.
See http://www.racer.nl/tutorial/onyx.htm
- Added 'run <script.oxs>' console command to run an Onyx script.
- Added onyx_run.exe, which compiles and runs an Onyx script. Useful for debugging (a .oxa ASCII file is generated as well).
- Added 'show triggerlines' and 'hide triggerlines' commands.
- Modeler's Import ASE function speed improvement.
- Particles did not use Z-testing.
- Backfire particles were spawned in the wrong position when the car had velocity.
 
Thanks for the update ruud ! :)
I've just noticied two bugs, first, when i do a lap, then i've the ghost car, and when i turn on the lights of my car, it's done to the ghost car and not mine :(
Also mount autobahn don't work since few versions :S
Yep noticed the same.....
here is the QLog to the autobahn track...

Wed May 02 20:41:04 (FATAL): [racer/5136] Exception 0xC0000005, flags 0, Address 0x0022DD82
(this dialog text is stored in QLOG.txt)

OS-Version: 6.1.7601 (Service Pack 1) 0x100-0x1

0x0022DD82 [newton]: (filename not available): NewtonMeshGetIndirectVertexStreams
0x0D6001E0 [(module-name not available)]: (filename not available): (function-name not available)

...hope there is a fix possible...
 
Ruud,
The problem with FFB appears to be a caching glitch. It happens when the "reload_car" console command is used, it doesn't always do it though. This problem also has very strange effects with the suspension lengths when changing the CG height.
I suspect this is also why one of my systems will practically "white out" when reloading a car.
I'm sure there are others but these are what I've found so far.

Alex Forbin
 
I have a problem with shdows in tunnels. There aren't any!

Tried my Watkins track in ver 0.0834, good shadows FPS 138, in ver 0.0839 with all shaders "cast_dhadows=1" FPS 193.
Tried Carlswood ver 0.834 good shadow in tunnel FPS 127, in ver 0.0839 with all shaders as above FPS 185 and NO shadow in tunnel.

I have stated before that the shadows as in ver 0.834 s/b the way to go and let track creators decide what shadows need to be turned off!!!!! A track without shadows is rather dull and unrealistic!!!

Thanks Ruud for getting the FPS speed increase but Please go back to the old way with shadows or tracks will NOT be correct!!
 
The RLD format is very simple, but where do you get your LIDAR data? The data I'm used to is not exactly public domain.
And define 'conventional data'. ;-) Most LIDAR scans ultimately seem to be just big XYZ files (optionally with a color or brightness field per 'pixel') where some creativity is needed, depending on the contents.

I guess it's important to note that what Racer is using is actually just a very high quality and density driving surface mesh derived from LIDAR data, not actually raw LIDAR data?!


Surfacing LIDAR data is what takes the time/skill/tools, and thus money hehe... a 5 mile stretch of track in high point density would take up about 300gig in just data points. Rationalising out the track surface in this raw format would still leave you with about 10gig of data! Eeek!

You can see why generally it's not public domain. To start with collecting the data is time intensive/costly, then combining the scans takes more time. Then storing it is another mammoth task, certainly not something you share over your internet connection :D

Then you need some nice software to build from it efficiently, and if accuracy is important (ie, not just drawing over it in 3DS Max), then you need someone with cloud > surface rationalisation software that can do it to a certain accuracy and automatically otherwise you'd spend years doing it manually.



I personally prefer just getting the surface right to within about 10cm, then adding some 'physics' perlin noise haha :D


I think in the next 5 years photo based image modelling will take precedent over laser scanning for most of our purposes, and another 5 years after that will replace laser scanning absolutely.

Maybe by then Racer will support tracks made by us just going out and taking a video with a 360deg FOV video head and then leaving it processing for a few weeks ;) :D


Dave
 
Ruud,
The problem with FFB appears to be a caching glitch. It happens when the "reload_car" console command is used, it doesn't always do it though. This problem also has very strange effects with the suspension lengths when changing the CG height.
I suspect this is also why one of my systems will practically "white out" when reloading a car.
I'm sure there are others but these are what I've found so far.

Alex Forbin

The FF generally doesn't like doing things like reload_car...

I have some scripts that set ff forces but when you reload scripts or reload car, then the ff forces can go weird, usually seeming to double up on reload, then be 3x higher on a 2nd reload etc etc...

It's like you need to flush the ff system whenever you reload or something?




Not tried this yet but will do some time soon. Looks like some nice updates.

My main curiosity right now is those 8192 x 8192 texture maps... that's 256meg/texture on 32bit rgba :D

A bit huge hehe!



PS Ruud, is there any chance any time soon of being able to create hdr cube maps (perhaps have some sample locations in xyz in track coords to take them from at track load or TOD change), then be able to reference them for reflective objects?

Ie, if I add a grandstand to a track with lots of glass on it, it reflects what the car sees, and with live env-mapping on that looks really horrible as you see tarmac up close in windows on nearby buildings!

We have kinda taken a step back here because the old system let us define sphere maps and reflection maps as much as we wanted, but now we are limited to just one.
With the new envmap mipmapping we can do some fantastic materials on cars, but the track items are still limited by reflecting the car surroundings :(


Thanks

Dave
 
The FF generally doesn't like doing things like reload_car...

I have some scripts that set ff forces but when you reload scripts or reload car, then the ff forces can go weird, usually seeming to double up on reload, then be 3x higher on a 2nd reload etc etc...

It's like you need to flush the ff system whenever you reload or something?


Dave

I haven't had any trouble using it before and it also affects other things as well, so I thought it might be worth looking into.
You should now be able to put almost all of your textures for your track in one map. You wanted big textures, Ruud delivered. ;)

Alex Forbin
 
We have kinda taken a step back here because the old system let us define sphere maps and reflection maps as much as we wanted, but now we are limited to just one.
With the new envmap mipmapping we can do some fantastic materials on cars, but the track items are still limited by reflecting the car surroundings :(
Being able to load static cubemaps could be put to good use, too - with alpha transparency, you could for example render an 'interior' map with transparent windows, so that inside the car reflects other bits inside the car, while the transparency lets the regular cubemap through the windows. Or, if the car has specific chrome parts like mirrors, they could get their own little cubemap of nearby bits of the car, backed by the regular environment map or the mirror.
 
Being able to load static cubemaps could be put to good use, too - with alpha transparency, you could for example render an 'interior' map with transparent windows, so that inside the car reflects other bits inside the car, while the transparency lets the regular cubemap through the windows. Or, if the car has specific chrome parts like mirrors, they could get their own little cubemap of nearby bits of the car, backed by the regular environment map or the mirror.

Yeah, I've wanted that feature for a long time hehe. Seeing the sky reflected in the car interior isn't so nice when it's not a convertible :D

That said, GT5 does it haha... such a simple fix though as you point out.


Alex, I guess so... but my main issue is that since material definitions and such are done per 'texture' too, then we are stuck with that whole texture just offering one material type.

I certainly agree that we need to atlas and optimise etc etc, but I can't see how one texture THAT big can be helpful generally.



However, I think for high end gfx card users, providing a sky texture that big will be really nice :D

However I also think that skies need to be able to use a cube map projected onto a dome (for the vert colouring), because the cube map gets you more data per area of sky per texture space than the other varieties we employ, especially that top-down map that Ruud uses on Carlswood which stretches the horizon data a lot (ie, the data you see the most)

Cube skies ftw!111!!!!

Hmmmm

Dave
 
There's something dodgy going on with colourgrading.

In photoshop if I adjust the curves so that the image is tinted blue then in racer it's actually tinted green.
Photoshop yellow - Racer magenta
PS Green - Racer Blue
PS Blue - Racer Green

Also, not sure if it's something I'm doing or what...but I had to disable cubemap mipmapping to disable the liveenvmap from always showing up instead of a custom cubemap i'm trying out. Which also doesn't work.
 
Made a tunnel for a track but when running my cyclops car the hill it went thru flikkered. Tried an identical copy and it was ok.

Took me two days to figure out what was wrong. I had two objects with the same texture used on both. One object had vertex colors and the hii didn't. I had exported the hill as a non-vertex colored object. When exported as vertex colored it was ok and no more flikkering. I use Blender, so it's important that objects that use textures that have vertex colors are all exported using the racer vc exporter and not the regular one.

Doing all this to demonstrate tunnel shadow problems.

Why is the damage inage in a fixed position? It would be better if it was in the views.ini file and the position changed.
 
Sat May 12 22:47:26 (INFO): [racer/3752] --- application start ---
Sat May 12 22:47:26 (INFO): [racer/3752] Racer version: 0.8.39 (Apr 27 2012/15:30:39) - customer: Internet
Sat May 12 22:47:27 (INFO): [racer/3752] DGPUShaderManager: render engine using Cg (3.30 NVIDIA via Cg compiler)
Sat May 12 22:47:27 (WARN): [racer/3752] DGPUShaderManager:MakeObject(data/renderer/fullscreen_shaders_hdr/shadows_f.cg): can't create CG fragment shader program
Sat May 12 22:47:27 (WARN): [racer/3752] DGPUShader::LoadAndCreateFromFile[data/renderer/fullscreen_shaders_hdr/shadows_f.cg]: The compile returned an error.
Sat May 12 22:47:27 (WARN): [racer/3752] data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(242) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(273) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(311) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(316) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(425) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/shadows_f.cg(61) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/shadows_f.cg(61) : error C1035: assignment of incompatible types
Sat May 12 22:47:28 (FATAL): [racer/3752] DGPUShaderManager:MakeObject(data/renderer/fullscreen_shaders_hdr/shadows_f.cg): can't create CG fragment shader program
The compile returned an error.
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(242) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(273) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(311) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(316) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/../common/hdr.cg(425) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/shadows_f.cg(61) : warning C7011: implicit cast from "float4" to "float3"
data/renderer/fullscreen_shaders_hdr/shadows_f.cg(61) : error C1035: assignment of incompatible types


Hey guys. Got this error while starting, maybe you got some suggestions?
 
I have updated my Surfaces and Sounds track woth a tunnel in the hill to demonstrate shadows and three kinds of water. There is a small problem when driving into the pond by the big cube. The car drops into the water but doesn't drop to the bottom, it gets stuck. It didn't do that before compressing and uploading to mediafire but does it when downloaded.

I'll have to work on it before I can release it to the community.
 

Latest News

Are you buying car setups?

  • Yes

  • No


Results are only viewable after voting.
Back
Top