Od[force] Shaders Scene

[Marc]

Hey all

We're releasing this to you early so you can possibly get started on something and then migrate to the final file once it's been finished. If you've been to other renderer's shader sites, then this should be familiar to you.

Basically the file has been built to provide a fairly neutral environment. This will allow for standard presentation of all shaders submitted using this method. We'll probably put up some rules to follow in the coming days. For now though, feel free to download the file and try it out.

Thanks

the odforce guys

--- the files follow here ---

odforce_shaderball_h8_v3.otl

[vmuriel]

Good initiative.

[Ratman]

Hey Marc, cool file, though i'm having issues with the otl.

Once i install it and open a new scene and apply my shaders, the textures stop working , it all works fine from your scene files provided, but it goes awry on me once i open the otl on another scene.

Other than that it's pretty cool

[Marc]

Hey Ratman

This is all expected to be used from within the example file. We'll be posting the rules of use in the next couple of days. But basically if you want to submit a shader then everything in the file should be identical apart from the shaders (and any cops/vops you need of course).

There will probably be some upgrades to the file in the near future anyway, so this is just for playing with .

[Jason]

I've made a small update to the shaderball scene.

*/ Basically, I've contained everything into the single OTL - all textures are embedded.

*/ I've added displacement shader support - no longer should it be necessary to Allow Editing to use displacement shaders

*/ Improved the shadow quality a bit

*/ The name of the shader is derived via expression from the actual shader name. You can change this behaviour if you want to.

I've updated the original post at the start of this thread with the new OTL.

[photex]

I've made a small update to the shaderball scene.

*/ Basically, I've contained everything into the single OTL - all textures are embedded.

*/ I've added displacement shader support - no longer should it be necessary to Allow Editing to use displacement shaders

*/ Improved the shadow quality a bit

*/ The name of the shader is derived via expression from the actual shader name. You can change this behaviour if you want to.

I've updated the original post at the start of this thread with the new OTL.

noice!

Are we going to end up with a version separate for H9?

[Jason]

noice!

Are we going to end up with a version separate for H9?

Yeah, I think we should - just a Material slot and we're done, (rubs hands!)

[crunch]

Yeah, I think we should - just a Material slot and we're done, (rubs hands!)

You probably also want to be able to select geometry: teapot, sphere, hair, volume, etc. But yeah, should be a wee bit simpler.

[michael]

different geometry is 'on the list'

we wanted to just get something started that would give everyone a solid start - so that people could make shaders and share them with the community in a standard way...

this scene will evolve as time goes on - so keep offering suggestions!

[Marc]

As for different geometry, we're aiming to build a single piece of geometry that can be used to represent 90% of the shaders that people will build. Any special shaders for things such as hair, fx shaders etc can have the geometry modified by the shader writer, but otherwise one of the rules that we're busy writing is that nothing must be changed in the scene other than what applies directly to the shader.

That way we get a consistent presentation of all shaders so people know what they're getting.

We're basically modeling ourselves after these guys and it seems to work for them :

http://mxmgallery.maxwellrender.com/

http://www.vray-materials.de/

[dee900]

When rendering, I get pink walls characteristic of a missing texture. I haven't looked into why the v_walls diffmap reference into COPs isn't working properly. (This is on Windows, v8.2.13)

[dee900]

We're basically modeling ourselves after these guys and it seems to work for them :

http://mxmgallery.maxwellrender.com/

The Maxwell people have put a lot of thought into their ball. Why not just use it (with attribution)?

[DaJuice]

The Maxwell people have put a lot of thought into their ball. Why not just use it (with attribution)?

Good idea, why re-invent the wheel?

[Jason]

Good idea, why re-invent the wheel?

Yeah, maybe we should...

Anyone have that scene anywhere?

It'll be a lot more fun to set up with H9 - with a PBR mode and a regular mode.

[Marc]

I just kinda assumed we wouldn't be allowed to.

To be perfectly honest though, the real problem with grabbing that scene is that it's not just the geometry... I could probably whip up that specific model in a couple of days pretty easily. However it's all the work they've put into the lighting measurements to make sure there's very specific colour response at various points in the scene that's a pain to copy.

Personally I think we can look at what they've done and make our own version that can be just as good. Either way we're going to have to put a fair amount of work into setting up the scene correctly from a lighting/rendering perspective... the modeling's the easy bit.

my 2cents

[DaJuice]

It also just occured to me that for a lot of things that maxwell model will be inadequate. I bet it's great for architectural shaders and so on, but not so much for something like tree bark.

One more thing... what do you think about changing the camera view? Maybe I'm the only one but I find it really claustrophobic. Something closer to www.cgsphere.com maybe... ?

[old school]

Yeah the cg sphere contest would be a lot of fun and gives the user some (crazy) freedom in that they can use any geometry in the scene that somewhat represents a sphere. Not as rigid as the Maxwell test scene or the vray material scene:

http://www.vray-materials.de/all_materials.php

Oh yeah there is the vray test scene. It has been used for tree bark and a whole bunch of other stuff.

[Andz]

Yeah the cg sphere contest would be a lot of fun and gives the user some (crazy) freedom in that they can use any geometry in the scene that somewhat represents a sphere. Not as rigid as the Maxwell test scene or the vray material scene:

http://www.vray-materials.de/all_materials.php

Oh yeah there is the vray test scene. It has been used for tree bark and a whole bunch of other stuff.

Fryrender (materials gallery) is using something like that, but they have few more shapes cloth and plane on that same ambient.

I remember that the most important aspects of Maxwell's ball while they were developing it was that it should have a way to show:

- Round areas;

- Flat;

- Transparency areas for sss, refractions, etc...

[old school]

All these render scenes are fine for static images but as we all know, we deal with objects in motion and with procedural shaders.

Maybe a second scene where the shader can be run through a standard animation to see if there are aliasing issues. A long grid to see how a procedural shader behaves, concave-convex surfaces, etc.

[Marc]

While the cg sphere thing is fun, there's nothing stopping you from entering something into their site now. There aren't any limitations on renderers are there?

I think rigidity has it's place in that there is consistency across the board for all shaders. I think the second scene with animation is a great idea, as well as having a couple more standard shapes (cloth, wall or whatever). First and foremost though, lets get this one scene down and then we can move onto the rest.

If there are enough submissions then we'll open up a shader gallery on od[force].

Cheers

Marc