New Mental Ray shaders development

[Maximus]

I'm using portal light and physical light only. I've spent hours yesterday tryin to figure out this.

In that render i posted is quite clear. Mila material is wrong in terms of intensity/shading color of the texture, to get the same texture color (correct one) of an architectural material i have to multiply the mila weight intensity by 3.14.
So either i got this totally wrong, either there is something wrong in softimage

[Kzin]

I'm using portal light and physical light only. I've spent hours yesterday tryin to figure out this.

In that render i posted is quite clear. Mila material is wrong in terms of intensity/shading color of the texture, to get the same texture color (correct one) of an architectural material i have to multiply the mila weight intensity by 3.14.
So either i got this totally wrong, either there is something wrong in softimage

its a different shading convention so it cant be the same. its like the ibl intensity, you have to divide the mia diffuse weight by pi, otherwise the rendering would be way to bright. its exactly the same here with mia and mila. mila follows the same convention like the ibl. so dont compare mila and mia without to change mia (if you are using booth in the same rendering). its also the same with mia and bsdf. bsdf is correct while you have to divide by pi for mia when mixing mia and bsdf in the same rendering.

its not mila that is false, its the mia shader which is wrong in the mila/ibl/bsdf universe. ;)

[Maximus]

I'm using portal light and physical light only. I've spent hours yesterday tryin to figure out this.

In that render i posted is quite clear. Mila material is wrong in terms of intensity/shading color of the texture, to get the same texture color (correct one) of an architectural material i have to multiply the mila weight intensity by 3.14.
So either i got this totally wrong, either there is something wrong in softimage

its a different shading convention so it cant be the same. its like the ibl intensity, you have to divide the mia diffuse weight by pi, otherwise the rendering would be way to bright. its exactly the same here with mia and mila. mila follows the same convention like the ibl. so dont compare mila and mia without to change mia (if you are using booth in the same rendering). its also the same with mia and bsdf. bsdf is correct while you have to divide by pi for mia when mixing mia and bsdf in the same rendering.

its not mila that is false, its the mia shader which is wrong in the mila/ibl/bsdf universe. ;)

And I might have understood this but now for a moment lets take out mia.
Create a scene and use mila only. Sphere on a plane, texture it. Render, the texture renders way darker than the correct color/intensity of the texture itself opened in photoshop.
So how you render correct texturing with mila? I was using mia + mila to compare the texture shading. of course if i apply a /pi on mia it renders darker as mila, but that is wrong in terms of shading. I cannot reproduce the same amount of color intensity of a texture opened in photoshop with Mila, if not by mulitplying the diffuse weight by 3.14.

[Kzin]

I'm using portal light and physical light only. I've spent hours yesterday tryin to figure out this.

In that render i posted is quite clear. Mila material is wrong in terms of intensity/shading color of the texture, to get the same texture color (correct one) of an architectural material i have to multiply the mila weight intensity by 3.14.
So either i got this totally wrong, either there is something wrong in softimage

its a different shading convention so it cant be the same. its like the ibl intensity, you have to divide the mia diffuse weight by pi, otherwise the rendering would be way to bright. its exactly the same here with mia and mila. mila follows the same convention like the ibl. so dont compare mila and mia without to change mia (if you are using booth in the same rendering). its also the same with mia and bsdf. bsdf is correct while you have to divide by pi for mia when mixing mia and bsdf in the same rendering.

its not mila that is false, its the mia shader which is wrong in the mila/ibl/bsdf universe. ;)

And I might have understood this but now for a moment lets take out mia.
Create a scene and use mila only. Sphere on a plane, texture it. Render, the texture renders way darker than the correct color/intensity of the texture itself opened in photoshop.
So how you render correct texturing with mila? I was using mia + mila to compare the texture shading. of course if i apply a /pi on mia it renders darker as mila, but that is wrong in terms of shading. I cannot reproduce the same amount of color intensity of a texture opened in photoshop with Mila, if not by mulitplying the diffuse weight by 3.14.

have you increased the lighting intensity?

so you say you cant render a texture with mila that looks like the texture in ps? (completely take away mia and its rendering)
i did a fast test and it looks good to me, texture looks like expected and not to dark.

but i know what you mean from my test. mia renders slightly brighter. ok, so conlcusion would be not to use mia and mila in one render. a diffuse weight of 0.26 gives a rendering that is nearly the same, visually of course.

[Maximus]

Here is what i mean. The scene is lit with 1 hdri (alex apartment from sIBL collection) userIBL env on environment node, userIBL env on a spotlight, userIBL rect on another spotlight.
Gamma correction on.

I'm using mila shaders only in this scene.



And here is what i meant, i had to multiply the weight by PI to get the correct texture result.



Am I going nuts or what?

[Kzin]

Here is what i mean. The scene is lit with 1 hdri (alex apartment from sIBL collection) userIBL env on environment node, userIBL env on a spotlight, userIBL rect on another spotlight.
Gamma correction on.

I'm using mila shaders only in this scene.



And here is what i meant, i had to multiply the weight by PI to get the correct texture result.



Am I going nuts or what?

my best advice, dont compare against mia. use mila and light/shade your scene like you want.
its a bit like ambient environments and fg, you would not render these two in one scene in one pass.

[Maximus]

I'm not comparing with mia, i'm comparing to real texture color in photoshop.
Basically if i dont multiply the weight by PI there is no way i can match the texture. Even increasing the light intensity it blowns out the whole render and the shading is still weird and darker.
Its incredibily off to shade a scene with those shaders if you dont multiply the weight cause everything is dark. I mean they dont match the real textures..

[Kzin]

I'm not comparing with mia, i'm comparing to real texture color in photoshop.
Basically if i dont multiply the weight by PI there is no way i can match the texture. Even increasing the light intensity it blowns out the whole render and the shading is still weird and darker.
Its incredibily off to shade a scene with those shaders if you dont multiply the weight cause everything is dark. I mean they dont match the real textures..

i have not that problem. i increase my light until the texture looks like it should. like i also did with mia or phong or lambert.

[Kzin]

ok, did another test with 2 different textures and 3 colors (100%white; 50% white and 10% white) with a white physical area light and it renders like expected.
can you run a test like this max?

edit: the image

[Maximus]

i'll do in a bit, i'm running a 1600px interior rendering with those mila, gonna let it finish and post it
also what intensity you used on physical light?

[Kzin]

think i should set this on hold actually. i try to carry some more informations.

max is right with this problem.
finish it in your way with the scale and show some renders and also shadersettings. meanwhile, i will do some more tests.

[Maximus]

Well this was a render I left running overnight with all the diffuse weight multiplied by 3.14. What i was tryin to do now is reshade this without the multiply as you said but i got frustrated on how much changes in light intensity and everything in the scene i have to do to achieve a correct texture result so i stopped it. (beside wtf final gather is a pain in the ass to get rid of splotches in interiors!)

Here textures match the real ps texture, i'm using this multiply thing using a scalar basic (set to 1, 3.14, and multiply as operation) plugged into the weight ports of the shaders.



There are some issues in sampling and scene is not finished, also there are some splotches from final gathering in the outside geometry wall.
Scene is rendered using portal lights with custom color on ceiling things, 2 portal lights outside with environment node and HDRI, and a spot light with userIBL env with the same hdri.

[Maximus]

Here you can find a screen of the column wood texture and the setup i used

http://img805.imageshack.us/img805/7927/grabi.jpg

edit: I've tried your setup with those spheres and texture/black white, i always have the same issue, it doesnt match the texture color values, i have to multiply the weight by PI.

[Kzin]

Here you can find a screen of the column wood texture and the setup i used

http://img805.imageshack.us/img805/7927/grabi.jpg

edit: I've tried your setup with those spheres and texture/black white, i always have the same issue, it doesnt match the texture color values, i have to multiply the weight by PI.

problem with that weighting by pi is that you cant layer correctly anymore. every diffuse layer would kill all other layers above him because the weight is way over one. thats the thing i look at the moment.

[Maximus]

I've posted on mental forum and apparently the solution i've found is quite right in terms of texturing diffuse, the problem with the layering .... well you are right, but for the moment i'm multiplying only the diffuse/texture part of the shader, not the reflective/refractive, and its working, at least it seems to.

Sure its a total mess since we cant just use a damn string option cause Softimage doesnt work like maya. Yay autodesk again

thats what i got from mental forum "Your current solution might be the only workaround I know of unless a developer has an idea."

[Daniel Brassard]

Max and KZin,

Silly question, are you working linearly or with gamma correction on?

Dan