1957 Simca Aronde Elysée

[Stereo]

How many months since I worked on this?

I guess 1 year makes sense. I'm off between exams again.

Front doors are now solid, and assigned generic models + script so I can open them. Still needs the lock, window crank, window latch, and interior panel detailing (mocked out in black) plus probably some tweaking of normals.

I made some fixes on the body as well, but it's much more of a pain to export so I didn't (merge 4 layers in blender, with about 90% of the objects needing mirror modifiers applied, and I haven't found a tool in blender to apply all modifiers to selected)

I also did most of the HUD art. The car doesn't have a tach so that's pretty much it, unless I decide to add oil/fuel needles. (skipped a couple warning lights cause I couldn't read them in any pictures)

 

[KS95]

Looking very very nice.

 

[Ruud]

The original beta download link is down; is this car still available for download somewhere? I can have a look at the scripts (and perhaps also use it is a testbed for an Onyx variant).

 

[Stereo]

http://nmurdoch.ca/simcaaronde90a.7z

Only tested on .38; includes a glass .cg in /cg that needs to be moved to the data renderer/shaders folder. (I don't know offhand how it differs from the ones included in Racer, not using alpha for reflectivity maybe?)

 

[Ruud]

http://nmurdoch.ca/simcaaronde90a.7z

Only tested on .38; includes a glass .cg in /cg that needs to be moved to the data renderer/shaders folder. (I don't know offhand how it differs from the ones included in Racer, not using alpha for reflectivity maybe?)

I haven't checked the Cg shader much yet, perhaps indeed alpha handling. Note that v0.8.40 also included it by default, so you can leave it out for future versions.

 

[boomer541]

Check out the latest version and had problems with the shader .cg file. only one was included the fragment or vertex was missing. Also the glass.tga texture image was missing.

I liked the grey color better. Great job Stereo, keep going with this wonderful car!

 

[Stereo]

The vertex shader is the regular dyn_standard_reflect_window_v.cg, so I didn't include it.

glass.tga was missing, I updated the archive and also the file itself is at glass.tga.

For the grey, that's something like

  ; silver metallic paint
  specular=.5375 .535 .515
  ambient=.32 .32 .35
  diffuse=.31 .31 .31
  reflect=0.4
  shininess=25

In the body shader. Just comment out the section described as red. Similar for switching to black or blue. For chrome, the fresnel section at the bottom of the shader also needs to be switched.

 

[boomer541]

Thanks Stereo!

 

[boomer541]

Stereo,
Keyboard key codes useful for scripting. List includes arrow keys, page up/down and virtually all keys on a standard windows computer.
Link:
http://www.webonweboff.com/tips/js/event_key_codes.aspx

 

[luthobu]

http://nmurdoch.ca/simcaaronde90a.7z

Only tested on .38; includes a glass .cg in /cg that needs to be moved to the data renderer/shaders folder. (I don't know offhand how it differs from the ones included in Racer, not using alpha for reflectivity maybe?)

So happy to see that you are still working on the Simca! We are two of
a kind, perfectionists Thnak you so much.

 

[Mr Whippy]

The vertex shader is the regular dyn_standard_reflect_window_v.cg, so I didn't include it.

glass.tga was missing, I updated the archive and also the file itself is at glass.tga.

For the grey, that's something like

  ; silver metallic paint
  specular=.5375 .535 .515
  ambient=.32 .32 .35
  diffuse=.31 .31 .31
  reflect=0.4
  shininess=25

In the body shader. Just comment out the section described as red. Similar for switching to black or blue. For chrome, the fresnel section at the bottom of the shader also needs to be switched.

I think you will get more realism if you turn reflect to 1 and then adjust the fresnel coefficients.
Technically any surface that is even partially smooth will give near 100% reflection at 90deg viewing angle. By using reflect=0.4 you are taking away 60% of the energy that might bounce off the car at a glancing angle.
It'd be better if you wanted a duller frontal and partial side on reflection to use a steeper power, say 4, so the strong reflection is limited right to the very steepest viewing angles, and also adjust the bias down to say 0.2 or 0.3, leaving the scale near almost 1.

reflect=1
fresnel
{
bias=0.03
scale=0.98
power=4
}

The end effect will be nearly the same, but your extreme viewing angles will look more natural in response to the bouncing light.
Just tested, yeah it's almost identical haha, but the very edge just gets a teeny bit brighter... but technically it's the right way to tweak I'd say. Reflect is best left at 1 imo, unless you are adding dust or dirt or some other effect via alpha control to the paint work, ie, something that impacts the energy of the incident light before it gets to the paint, once it hits the paint the response is then via fresnel factors.

Just my 2p of course, but we have pretty accurate light intensity values so it makes sense to use realistic lighting responses on our materials as much as possible

Thanks

Dave

 

[Mr Whippy]

Haha, this car is amazing.

Needs Mz work on the pacejka as my G25 was going mad, but it's really nice looking.

Once all the script stuff and mirrors are implemented etc, and the car.ini tweaked a bit, it'll be a good default car showing off all the features imo!

That being said I think Racer needs a few different default cars to show itself off more, ie a car with slicks for racing with, a modern 'normal' car and a good classic like this.
They will also be good for testing new betas as they will contain so much material problems will easily show up when something is wrong!

Cheers

Dave

 

[Stereo]

Yeah I have no FFB so I don't have a lot of control over that aspect, any ini advice would be happily taken in and updated. I think the pacejka consists of taking an existing car and tweaking the main grip numbers downward to match the narrow tires.

 

[Mr Whippy]

Yeah. Cosmo and I were just talking about tyres a bit and how the FF forces (Mz) and slip angle shapes might correlate a bit.

I'd always done it a certain way in the past but when I had the pleasure of driving my M3 on a Cruden simulator it was really odd in the way it drove. The g forces felt wrong next to the FF, so I had to re-adjust the perceptions of how it worked in my head but I think Cosmo and I are in good agreement (which is always reassuring) about how those curves should look a bit more these days.

I'm sure Cosmo will help out, and I am willing to have a look too. I'm more than happy to play with Pacejka though and get it feeling a bit nicer for now

Cheers

Dave

 

[Mr Whippy]

Uh, seems I was wrong on the fresnel curve parameters

bias is the 0deg angle, scale is the 90deg angle, power is the power of the curve plotted between the points.

However, scale can go to any value and the intersection of the curve with y=1 is clamped.

But the y value at 90deg is a sum of bias and scale.

So if you want a material to reflect 100% at 90deg then the bias and scale must = 1... feels a bit odd to do things this way vs 3D applications which has the values as discrete and plots the line between them.

Do most 3D applications do the maths this way then just remove the summing in the front-end inputs to make it intuitive?
Just seems odd to everything I've used before to do it this way around in Racer.
Thankfully since car lacquer is the way it is it's meant I've used values that nearly sum to 1 any way haha.


I still don't get why the reflect value is set near 0.4 though... that feels too low to me by a long way.
I'd still say if the car generally looks duller then power =4 or 8 would be more appropriate, and then up the reflect value. Otherwise we are basically saying the car doesn't reflect more than 40% of incident light at 90deg which seems a bit iffy... I guess it might be right for the paint on that car though but I'd have to see it and check with my cameras spot metering to check

Dave