Page 2 of 3 FirstFirst 123 LastLast
Results 21 to 40 of 46
  1. #21
    aha Python.thanks for the link Secrop ,i have a look into this.

    i have a idea for a addon/python node.a node as database reader only.
    this could read/browse the database,and maybe a pre select for the wavelengths,and the outputs are the N and K's for all RGB.this would be handy for my project,but for other/future shaders to, what needs the n k values.

    maybe you are interested building one?would be very usefull.my python acknowlede is very basic,so for me it is not on my radar now.



  2. #22
    Member
    Join Date
    May 2013
    Posts
    1,013
    Oh, python. Guess I should have guessed
    I'm clueless about it, and if I even knew a little bit about it, materials isn't what I would do with it (project automation).
    Still, I really wish the material GUI - what to expose, reroutes (plug in texture to a node group, then feed it the vector transforms), tooltips, graphical gui support, new gui elements etc - was a lot better than what we currently have.



  3. #23
    Member Secrop's Avatar
    Join Date
    Dec 2011
    Location
    Berlin
    Posts
    1,894
    Originally Posted by pixelgrip View Post
    maybe you are interested building one?would be very usefull.my python acknowlede is very basic,so for me it is not on my radar now.
    Why? I already built one.



  4. #24
    @Secrop

    Do you mean to make a node set up group or not with API python commands
    and not custom nodes thing?

    I think there is supposed to be some new features for nodes and python in 2.8
    should be more powerful


    happy bl
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  5. #25
    Member Secrop's Avatar
    Join Date
    Dec 2011
    Location
    Berlin
    Posts
    1,894
    Originally Posted by RickyBlender View Post
    Do you mean to make a node set up group or not with API python commands
    and not custom nodes thing?

    I think there is supposed to be some new features for nodes and python in 2.8
    should be more powerful

    happy bl
    Not sure if I understand you correctly...
    I had plans to make some UI editor for making custom nodes... But as soon I started to write more complex nodes, I froze that idea.
    The main reason was that it's quite impossible to include every possible option into it without filling the UI with operators, settings, etc. Not to mention that some things really need to be coded (hands dirty on Python). For example, my 'Loop' node would be impossible to be made alone without going into code.
    I could improve a little bit the nodegroup converter that ships with the ShaderNodesExtra, but for now i'm writing a new complex Node/Paint system, and I won't have time for the next months to diverge my attentions. Anyway, the converter Operator is working fairly ok (with some problems in the generated code, but those can be fix). The source is online, so anyone with some time can also contribute. I can still give away some guidance, but as I said, I won't have time to be fully involved in it.



  6. #26
    for myself i'm putting an hold on python for now
    too many changes are coming in 2.8
    so waiting to see the new improvements coming in 2.8
    then I will re start to learn the new things again LOL

    happy bl
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  7. #27
    Member
    Join Date
    Sep 2012
    Posts
    2,768

    Helping hand

    @Secrop
    Since you haven't posted your link...

    For all who don't care to be bothered by wasting time on metals but still wish to achieve physically accurate results (or only to support the community and efforts of talented, hard working individuals) here is...

    ... Metal Bsdf for Cycles by Secrop on Blender Market ($15)




  8. #28
    Member Secrop's Avatar
    Join Date
    Dec 2011
    Location
    Berlin
    Posts
    1,894
    @burnin thanks for the publicity... I haven't post the link in this thread only because I didn't wanted to flood it (thought my addon is related)



  9. #29
    thinfilm shader and conductor shader ,both rgb outputs multiplyed and the sum as fresnelfac into the Principle shader

    maybe thats enough.i try to implement a new Fresnel equation.

    Click image for larger version. 

Name:	thinconduct1.jpg 
Views:	10 
Size:	704.0 KB 
ID:	504937
    Last edited by pixelgrip; 12-Nov-17 at 01:09.



  10. #30
    Member
    Join Date
    Sep 2012
    Posts
    2,768
    @Secrop
    don't mention it, was at least i could do & thank you too for everything you share here with us, you're a great source of knowledge & an inspiration


    @pixelgrip
    hear, hear... inspiring love results you're showing
    Would you kindly specify/describe/share materials that you're wanting to represent or do you intend to do it later (once satisfied - as you did with thin film/microroughness)?

    Am also very interested in comparing results/techniques coming from Cycles to those rendered with other engines as well.


    cheers & stay well



  11. #31
    burnin,i post the conductor build soon.

    i am not sure atm ,what is the best method, to reproduce the reflection intensity with cycles.
    since this build and the thinfilm build gives the calculated Fresnel (R) reflection as output result.

    as you maybe know ,cycles has no reflectivity input for this.so to bring this colors to the shader, there a the standard methods.like mixing the result with other colors, or use as color directly.or use the result as fac to mix the colorchannels,like you would to with maybe a common diffuse/glossy fresnel-fac mix.but all gives different results in saturation and intensity.

    in this render i have used the nk values from titanium.btw i have made 14 "metal presets" .all metal values for 3d artist from refractive index site,and a few more.

    the brighter colors are the result into the colorinput from the glossyshader.the darker more saturated colors are seperate r g b result into a layer weight blend input and the Fresnel outputs combined into the glossy colorinput.on the left are google search result of titanium.the funny is both colors are present (this brighter and this brownish).

    maybe i have to build another equation.hmm,but the problem how to handle the Fresnel reflectivity still exist to me.maybe someone has an idea?

    Click image for larger version. 

Name:	titanium1.JPG 
Views:	19 
Size:	276.5 KB 
ID:	505093



  12. #32
    i have rebuilded another Fresnel equation code.it gives me exactly the same results,with the preset values,like the first build.so on this side, it seems the calculations are correct.i make then a minimalistic blend file,and post this soon,so you can play with.the new rebuild has F0 reflection output and edge reflection output additional to the R reflection.i give them seperate r g b outputs like the first build to.i think thats enough outputs to play and experiment with.



  13. #33
    Here you can download the conductor shader v2

    conductor-shaderv2.blend

    v2_1 same build with quick 2 point preset connection

    conductor-shaderv2_1.blend


    the shader is based on this code,for reference

    http://jcgt.org/published/0003/04/03/

    Click image for larger version. 

Name:	worm1.jpg 
Views:	5 
Size:	575.0 KB 
ID:	505229
    Last edited by pixelgrip; 15-Nov-17 at 19:48.



  14. #34
    Pixelgrip, this is looking really cool! Thanks for your work on redoing the math.

    When you combined this new metal node with thinfilm, did you literally Color: multiply the color output of each node and plug that into the glossy shader? What values did you use for the thinfilm IOR, and more importantly the substrate, since in this case the substrate is the RGB n,K metal? In my tests using color multiplication the thin film color is very dependent on the substrate IoR and many of the metals have a strong difference between r,g,b IoR so it's hard to "guess" at this value.

    Something you might try as a usability improvement for the node group is lumping the n,K values into vectors so you make 2 links instead of 6 when changing presets.



  15. #35
    thanks spiderbrigade.
    this render was just a quick connect and try out.the settings from the thinfilm was with a filmIOR around 2.4-2.7 that i have used,for a cadmium render before in another thread.the substrate was IOR1.67 for a zinc mat.so this was matching abit the gold color from the conductor shader.
    the reason for multiply was just experimenting,to find out a good method to mix the two shader,without getting to dark or overblending.i get the idea to multiply it ,as i remember that the thinfilm layer in the equation are multiplyed to. maybe you came up with a better idea to mix it?
    since this conducor shader is new to me to,i can learn from users who came up with ideas to.
    would be nice to see some testrender.and would be nice to hear ,if someone knows how accurate and usefull the shader output is.
    good idea with the vectors,maybe i try the idea later,if its going i update then the version.

    edit v2_1 is uploaded with 2 point connection

    here the settings from the render

    Click image for larger version. 

Name:	multiset1.JPG 
Views:	6 
Size:	194.6 KB 
ID:	505365

    edit,if you want seperate rgb outputs from the thinfilm layer,then use the converter-seperate rgb.
    btw the k value in the thinfilm shader is almost useless,and should set at 0.i simply forget to delete it.if i remember it is just a beer lambert absorption equation that i tryed to simply add to the n value from the substrate.that was before i know, that for metals another calculation for the extinction coeffcient is needed.
    Last edited by pixelgrip; 15-Nov-17 at 19:50.



  16. #36
    I'll see if I can work up some more test renders for you this afternoon, mostly focused on the combination of air-conductor with thin-film. I think the air-conductor by itself is so close to correct that human eyes aren't going to be able to spot problems. The thin-film interaction is where the color shifts are dramatic and very noticeable when they are incorrect to real-world examples. For instance, you already pointed out that photos of titanium have a silvery and a brownish color - the brownish color is caused by films of Titanium dioxide so the pure air-conductor node shouldn't account for it.

    Here's a test with your nodes (top) compared to the older OSL thin-film (bottom). Both have titanium as the substrate with a film of TiO2 varying from 0-100 across the x axis. For your nodes, I actually used three copies of the thin-film group, with the TiO2 n values for r,g,b. Each output had the respective component multiplied with the substrate component.

    As you see, compared to the included real-world image of Ti/TiO2 yours is closer to accurate, but MUCH less saturated than it should be, to my eye.

    I started with TiO2 because it's a fairly common film for metallic/pearlescent paints
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	FilmTest001.png 
Views:	2 
Size:	203.5 KB 
ID:	505449  

    Attached Images Attached Images  



  17. #37
    I think you should be plugging the color into a normal glossy shader (same with metal shaders that take n and k values), because if I understand it correctly, thin film calculations are Fresnel equations themselves.

    I have actually been working on a thin film shader that takes both n and k values, too. The equations I've been using follow Pruster's code for thin films in one of the links earlier in this thread. Though, it seems there is a bug in my nodes I haven't been able to solve for the past few weeks (month). I would post the blend file (which also includes the conductor shader group I've created), but I'm still under the 10 post requirement for that, haha (trying to work on that so I can be more involved in the community!). Basically, the node group defines various complex number arithmetic to produce the fresnel (and thin film) equations.

    Multiplying the values (as you show in #35) does not produce correct results; the thin film calculations should already give the correct/close to correct values. You would have to instead adapt the thin film equation you are using to be able to take the k values.

    Good luck!



  18. #38
    thanks spiderbrigade,thats very usefull test,and thats what i am was asking for.

    i have tryed to render the same materials.thin film shader with 1 for air,then conductor n output for thinfilm ior (titaniumdioxide) and substrate titanium ior from a second conductor shader.

    the get n and k2 are functions from the conductor code, that i made a additional output for.i was testing this,if this gives the same visual output ,in the film shader vs the conductor.and i am not sure ,if this functions are the "real" n and k results, or if they are only some simple helperfunctions in the calculation.

    this is the render result

    Click image for larger version. 

Name:	tidioxide1.JPG 
Views:	14 
Size:	163.2 KB 
ID:	505467



  19. #39
    What are you using to vary the thickness in that render? I find it's helpful to use a linear gradient of some kind so you can mentally map the color range to the thickness range more easily when doing comparisons.

    Edit: it also looks like you might have the Titanium substrate k going into the IoR for the film group.

    Here is a test of MgF2 over Silicon, another common pearl-paint colorant combination. This is using my method of combination (three separate thin-film groups for r,g,b). Thickness is 400 (What's your node group thickness unit?). I've artificially increased saturation by 1.5.

    I'd be interested to see if your method posted above returns similar results. For values I'm using:
    Silicon 3.84,4.14,4.68 n; 0.015,0.030,0.091 k
    MgF2 1.377,1.38,1.382 n
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	MgF2.png 
Views:	2 
Size:	1.44 MB 
ID:	505471  

    Last edited by spiderbrigade; 16-Nov-17 at 15:03.



  20. #40
    Based on some info I gathered when testing the OSL thin-film, here are a few tests with Fe2O3 over Silicon, which is supposed to be the combination used in Merck Colorstream pigments. These are 390 and 510 thickness. Compare to the second and last material samples (in the next post down). I haven't been able to duplicate the others but Colorstream also uses TiO2 so there may be some combinations necessary.

    Unforunately many of these pigments use more complex sandwiches rather than just substrate/film, so they might not be doable via this node group. Possible to stack these somehow?
    Attached Images Attached Images    
    Last edited by spiderbrigade; 16-Nov-17 at 16:33.



Page 2 of 3 FirstFirst 123 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •