Jump to content
Sign in to follow this  
DrScarlett

Neon light and area light samples

Recommended Posts

  • Topic Author
  • Thank you for those kind words, Vizn.

     

    The clips are indeed a type of black plastic, same material as the neon tube 'hub caps' at their ends, actually. That's the way I wanted it to be, they are small details, but I wanted them to be clearly visible. I can get a bit finicky, if results are not perfect, or if solutions do not work in the way I would, hmmm, expect them to.

    The area lights definitely have some unexpected side effects, due to the way that feature is implemented. I have switched over to GI illumination textures and found settings and rendering settings that produce an acceptable result.

     

    Now just wait for another 7 days or so for me to share the method details and the rendered result with you ::):

    Hang on, incoming.

    Share this post


    Link to post
    Share on other sites

  • Topic Author
  • OK. It's done.

    This version is made like this:

     

    Scene:

    • Outer neon class tubes is are cylinders of 20 mm diameter, wrapped onto the letter shaped splines. On top of that they have a glass thickness of 2 mm. The material is a 90% transparent material with 1.25 refraction index and a reflectance channel.
    • The inner neon light tubes are cylinders of 10 mm diameter, wrapped onto the letter shaped splines. The material only has a luminance channel. The blue is 240% brightness, the green 350%. In both, in the Illumination channel, the generate GI is cranked up to 2500% (strength only, saturation remains 100%).
    • Th outer neon class tubes have a compositing tag that turns off the "seen by rays" and "seen by GI". If the "seen by rays" is not turned off, the illuminated surfaces of the inner neon light tubes are not seen (completely black) in the render.  If the "seen by GI" is not turned off, the glass blocks the GI illumination of the inner neon tubes onto the rest of the scene, so the surrounding objects are not illuminated by the neon illumination material.
    • I fixed the earlier issues I had with the reflections by properly balancing out the diffuse and specular reflectivity of the floor tile material.
    • The letters themselves are a grey (no color) plastic material with a bit of surface bump noise.

     

    Lighting:

    • There are a couple of stark red object driven area lights in the background to the left and right, kind of soft-box like. That is all the lighting in the scene. It would be interesting to see a render with the neon turned off ::):

     

    Rendering:

    • The rendered used is physical in default setting and has ambient occlusion in default setting.
    • The Global illumination is QMC for both primary and secondary method, diffuse depth 3, samples set to "high", accuracy 90%, glass/mirror optimization at 91%.
    • I experimented with the "polygon light' setting of the illumination material. I believe it can help you optimize render times if you only care about the quality and graininess of the surfaces directly illuminated by the material. You can use it to focus calculation rays on those surfaces and drop down the ray count of the rest of the scene and secondary hops. in my case I could not find a way to use it and still get the overall quality I wanted.

     

    And here is the result. Rendered at 4800 x 2700. Took 212 hours to render:

    MuGfufs.jpg

     

    Overall, the way the scene looks, I am happy with it. I will not try any further without new idea's or information.

    I am not so happy with the neon tubes themselves, especially when rendered up close. Due to the compositing tag needed on the glass, the glass is not lit directly by the illumination material running inside it, it is only reflecting light on the outside from other sources. The look this creates is not natural. Also caustic dispersion effects of the glass do not occur, but that is not so important. Studying close-up shots of neon lights, and testing both the area light and the GI methods mentioned in this thread in some detail, in the end, I am still struggling to produce realistic looking neon tubes.

     

    I hope you all enjoy this info or find it in some way useful. I would be happy about any additional tips, especially on improving the close up look of the tubes.

    Share this post


    Link to post
    Share on other sites

    Thank you for sharing your result with such excellent explanation for what you did and why...

     

    Very nice result, but 212 hours to render 1 frame ?! :compHate: 

    We must be able to do better than that, surely ?

     

    Did you try any 'lesser' GI methods other than QMC ?

     

    CBR

     

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • Well, I had a look, and I have seen there are ways to use compositing tags to set quality requirements (stochastic sampling ratio, record density ratio) to force improvements where you want them. But my initial low and medium test settings were such bad results that I abandoned those directions.

    If you have a suggestion on what settings to use, that would be awesome.

    Share this post


    Link to post
    Share on other sites

    I was primarily wondering what would happen if you swapped out QMC for secondary bounces with Light Mapping, which is a hell of a lot faster, but not sure if it is up to your quality standards !

     

    CBR

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • I'll try it again :)

    Share this post


    Link to post
    Share on other sites

    Oh, wow..212 hours?! :ohmy: 

     

    If you're looking to do this sort of thing as a contractor, I would suggest looking into a third-party renderer. I would recommend VRay (it's the only one I've ever used) once they get the Next version out for C4D, but it will depend on your budget.

     

    With 3D rendering, a big part of the 'art' of it is utilizing modeling and rendering techniques that are highly efficient in order to get your work 'close enough'. Modeling everything as realistically as possible is almost never the most efficient path to the 'perfect' rendering. This will apply to varying degrees depending on the objects and scene being rendered. Most people won't notice the small things that might bother you as an artiste. Yes, it will look amazing, but perfectionism is an obstacle to productivity. The finer the granular realism you try to get out of your work, the longer it will take.

     

    From a purely rendering perspective, 9 days for one frame is a failure. The irony of this is no one will argue that the results aren't beautiful. So, if this is solely for your own pleasure, then have at it. But if your intent is to work and earn money in this field, it will only benefit you to get your efficiency chops up. 

     

    That said, I look forward to seeing more of your work!

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • Haha, I enjoyed reading that, mostly because I am well aware ::):. Don't worry, this would not have been my course of action had I a client waiting in the background ::): Also the less perfect GI methods seem to be paying off. I seem to be getting a very comparable result (if I ignore some of the finer points) for under 7 hours of rendering.

    I'll show it to you later and prove your point ::):

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • Here's IR+LM (not QMC+LM as Cerbera suggested, so even a lower method than that).
    The direct illumination on the letters is a bit washed out, less nice details in the illumination, and some blotchiness of the green light further out on the tiles. But sort of acceptable.

    In just over 6 hours, lol.

    Jtx9HPu.jpg

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • Some more experimenting:

    I figured out the blotchiness comes from the IR primary, not from the LM secondary. I managed to get some nice LM settings that work, but figuring out how to smoothen out the IR over a large planar surface was beyond me. Anyone have any info or tuts I can follow on that? I think the key is in the detailed settings in the IR tab.

     

    Also I tried the QMC + LM. No significant speed increase over the 200 hour QMC+QMC render, and no significant quality difference. It clear that it is using QMC as a primary is what bites me here. However, it is the only primary GI method so far that managed to give me quality without artifacts. Again, learning how to optimize primary QMC settings could help me here.

     

    I currently have no budget for a third party renderer, unfortunately :(

     

    Finally, I did some lights-out renderings that are nice to look at and then switch to a lights on render. 

    GvtEZIt.jpg

    3TYkwh3.jpg

    MuGfufs.jpg

    Share this post


    Link to post
    Share on other sites

    Join the conversation

    You can post now and register later. If you have an account, sign in now to post with your account.

    Guest
    Reply to this topic...

    ×   Pasted as rich text.   Paste as plain text instead

      Only 75 emoji are allowed.

    ×   Your link has been automatically embedded.   Display as a link instead

    ×   Your previous content has been restored.   Clear editor

    ×   You cannot paste images directly. Upload or insert images from URL.

    Sign in to follow this  

    • Recently Browsing   0 members

      No registered users viewing this page.

    ×
    ×
    • Create New...