[Bf-committers] 2.28 Ahoy 2!

Ton Roosendaal bf-committers@blender.org
Thu, 24 Jul 2003 10:49:33 +0200

Hi Nathan,

No panic! Here's my reasons why I've implemented it this way:

- the commits you did for this system were so many, that I could not  
locate all changes for it. So I've loaded tuhopuu.exe, checked what I  
liked from it, and started over from scratch.

- I've only copied the shader algorithm code you provided. Even that  
needed cleanup to match coding style in Blender better :)

- Yep, I completely forgot about the texture channels for it. A nice  
addition for a 2.29 release!

- The buttons layout you did in Tuhopuu really was messy, with buttons  
crammed together below a minimum acceptible size. I know it's a pity I  
couldnt find space to make shader buttons a slider, but the current  
layout at least groups most functions together that relate;
   - the 'vpaint col' etc options with the RGB sliders
   - all shader options (and spec and ref sliders) together
It was the closest I could think of to remain in the current layout...  
we now really need a method to expand buttons easier. Proposals for  
that are being discussed.

- I found some tuhopuu code that related to shader plugins, but could  
not find clues how it was supposed to work. Apart from that, I don't  
even *know* if we want it in bf-blender. Plugins in the past (texture,  
sequence) only have given us troubles. My preference is to look first  
if we can do this with Python. The python coders already investigate  
feasibility. Then at least we have something stable & cross platform.
I am very confident the current implementation can easily be tweaked  
for such plugins. Let's first discuss & research this topic!

- Blender already used the 'hardness' value to determine specular size.  
I decided to copy that for Blinn, since it works almost the same. I  
transformed the hardness integer value to a float, with a formula  
that's close to what Phong and CookTorrance offer. The 'roughness' you  
provided in Tuhopuu was impossible to work with... with a slider which  
offered the most interesting results in the lower 1% of the values.
I realize what i did is not 'scientifically correct'... it was just a  
quick decision, with the aim to make something that worked. I also  
didn't know the Blinn roughness was tied to Oren's roughness. In all  
articles I read about it, I couldn't find a reference for it.
Again, no harm done if we decide later on to change it. Blender  
filereader can convert it.

- The 'Toon diffuse' size value can of course easily be allowed to go  
up to 1.7.

Cessen, I really liked your idea to improve the shading in Blender, and  
the code for the math was extremely welcome. I will be glad to further  
work with you on this, and come together with improvements for further  


On Thursday, Jul 24, 2003, at 09:02 Europe/Amsterdam, Nathan Vegdahl  

> I keep on realizing more and more things wrong with the material  
> system.  I
> really should just wait until I've figured out everything...
>    Based on empirical evidence, it appears that much of the underlying
> structure I had coded for the shading system was not carried over  
> during the
> port.  That structure was *very* important for the planned  
> plugin-system.  If
> it is left as is, it will be very difficult to impliment the shader  
> plugin
> system without breaking compatability.  And, even in terms of the  
> short-term,
> the structure put in place made it very easy to add additional shaders  
> (it was
> set-up such that no further modifications needed to be made to SDNA to  
> add
> additional shaders, regardless of the number of shader parameters it  
> would
> entail).
>    On a more minor issue: another problem I noticed with the Blinn  
> shader in
> 2.28 is that even at it's highest hardness setting (255), the specular
> highlight still doesn't get nearly as sharp as the underlying shader  
> code
> allows.  And, on the flip side, even at a setting of 1, it doesn't get  
> as broad
> as the underlying shader allows.
>    I hate to ask this, but is it possible to roll-back just the shader
> up-dates, and postpone it till a later release (so that we can  
> hopefully get
> the shader system ported in it's entirety)?  I feel really bad about
> criticising Ton's work so much, because he really *did* do a great job  
> of
> porting it for not knowing the reasons why everything was the way it  
> was.
>    Again, I apologize for all of the trouble.  I've made a real mess  
> out of
> things.  I feel terrible.  :-(
> --Nathan Vegdahl
> --- Ton Roosendaal <ton@blender.org> wrote:
>> Hi,
>> Committed since my last mail:
>> - stefang found & fixed a 2.27 bug (SHIFT+O), great!
>> - matt provided a more crispy splash
>> - changed wording of insert key menu.
>> I did a test with all my testing files, Blender still behaves nice! No
>> reports in yet which block final compilations...
>> Let's go! Aim is to put it all up this evening (euro time). I'm also
>> available thursday morning for some chores, if needed.
>> I'll put files at download.blender.org immediately, to wait for first
>> reports. If it all goes nicely, we can post it all officially  
>> thursday.
>> Now I'll write the changelog & blender.org article...
>> -Ton-
>> ---------------------------------------------------------------------- 
>> --
>> --
>> Ton Roosendaal  Blender Foundation ton@blender.org
>> http://www.blender.org
>> _______________________________________________
>> Bf-committers mailing list
>> Bf-committers@blender.org
>> http://www.blender.org/mailman/listinfo/bf-committers
> __________________________________
> Do you Yahoo!?
> Yahoo! SiteBuilder - Free, easy-to-use web site design software
> http://sitebuilder.yahoo.com
> _______________________________________________
> Bf-committers mailing list
> Bf-committers@blender.org
> http://www.blender.org/mailman/listinfo/bf-committers
Ton Roosendaal  Blender Foundation ton@blender.org