YafaRay Plug-in - reviewing for Yafaray 3.0

[david.bluecame]

Hello,

Yafaray website is having problems again so I cannot answer your questions there.

I've seen why HDRI are not shown and are not reflecting in Glass:

* Make sure the glass mirror color is set to 1.0,1.0,1.0. In my tests I got 0.0,0.0,0.0 and that's why HDRI was not reflecting.

* For HDRI not showing, these lines should be inserted in the integrator section:
<bg_transp bval="false"/>
<bg_transp_refract bval="false"/>

In any case there are *many* features from YafaRay v2.1.1 and the new v3.0.0 that are not available in the current Wings3D plugin. I will try to learn how the plugin works and will start sending pull requests with some suggestions.

---

This is the result of using (v3) yafaray-xml on the xml generated by Wings3D after manually applying the above changes to the xml:

   

[oort]

David,
Thanks!
The code for <bg_transp_refract bval="false"/> is already there as an option under Render Settings Background.

println(F, "<bg_transp_refract bval=\"~s\"/>",[format(BackgroundTranspRefract)]),

Simply turning that option off fixes the problem.

What does <bg_transp bval="false"/> do? I seem to be getting the same results with and without that line.

Yes, there are many things that you have added that need to be added...

Thanks,
oort

[micheus]

Thread created to group the discussion about the plugin changes needed to support the future Yafaray 3.0: New YafaRay v3.0.0-beta for Blender v2.77a released!

------
david.blucame, you did a excellent work on enabling to set a per-material raytracing depth. That was a interesting idea and the result couldn't be better.

In any case there are *many* features from YafaRay v2.1.1 and the new v3.0.0 that are not available in the current Wings3D plugin. I will try to learn how the plugin works and will start sending pull requests with some suggestions.

If work with Erlang can be hard for you, I can try to implement the changes for the plugin with your and oort's help.

Because of the way the UI is coded and the Yafaray is the most complete and complex of them it can look hard to play with it at first sight.
The source is here
Eventually, by using the dialog images to point changes and the place they need to be done can help us to better talk about each one. A set of plugins images can be found here.

Cheers

[david.bluecame]

Thank you, Micheus!

No worries, I got Wings3D compiled and working so I can tweak the Yafaray plugin and do tests, etc.

If possible I would like to have access to some wings3d scenes already set for YafaRay so I can make sure I don't mess up things (too much

I will also use this experience to try to improve things in YafaRay. For example I noticed that the color log from Yafa does not agree with Wings3D log window (the ANSI color codes appear weird) so I will implement a new option in v3.0.1 "--log-no-colors" so Wings3D log window is nice and tidy with Yafa

Thanks and best regards!

[oort]

David,
Glad to hear you are already able to compile Wings3D. I figured Wings3D would be a walk in the park for you compared to YafaRay.

There are several example files on the bottom of the Wings3D page of my website. There is a link in all of my posts to the site. Wings3D YafaRay files

I saw your post in the other thread that you can fix the "-pbp none" issue. That is good news.

Thanks,
oort

[david.bluecame]

Thank you, I will take a look.
I'm testing the upcoming v3.0.1-beta and it no longer requires -pbp none.
Also, I've added to yafaray-xml a new option "-ccd" or "--console-colors-disabled". With that option, now Wings3D log window is a lot cleaner, for example:

-------------------------------------------------------------------------
Progress: [#################################################### ] (89%)
Progress: [##################################################### ] (91%)
Progress: [###################################################### ] (93%)
Progress: [####################################################### ] (94%)
Progress: [######################################################## ] (96%)
Progress: [######################################################### ] (98%)
Progress: [##########################################################] (100%)
[22:23:49] INFO (+ 4s): DirectLight: Overall rendertime: 13.8s
[22:23:49] INFO: imageFilm: Flushing buffer (View number 0)...
[22:23:49] PARM: --------------------------------------------------------------------------------
[22:23:49] PARM: YafaRay (v3.0.0-beta-1-g521b3a7-dirty) Linux 64bit GCC | 400x400 | 3 passes | Render time: 14s | Total time: 14s | CPU threads=8
[22:23:49] PARM: Direct Light RayDepth=12 | photon maps: 0.0s [8 thread(s)]
[22:23:49] PARM: AA Settings (gauss): Tile size=32 passes=3 samples=3 inc_samples=3 resamp.floor=0
[22:23:49] PARM: sample.mul=1 light.sam.mul=1 ind.sam.mul=1
[22:23:49] PARM: col.noise=0 AA thr=0.02 var.edge=10 var.pix=0 clamp=0 ind.clamp=0
[22:23:49] PARM: --------------------------------------------------------------------------------
[22:23:49] INFO: Saving image files
[22:23:49] INFO: PNGHandler: Saving RGB file as "//yafaray/tests/test1.png"...

Rendering Job returned: 0
Export time: 0.404"
Render time: 14.339"
Total time: 14.743"
Rendering Job ready

Loading rendered image
-------------------------------------------------------------------------

I will include the -ccd option in the Wings3D plugin directly so the users don't need to add it manually.

[oort]

I tested all of my example files. Turning off "Transp Refraction" makes all of the background colors work. Does this option still do what it is meant to do?

I do not think that Alpha backgrounds are working.

I am seeing lots of graininess in the Cardioid Caustics, Diamond Dispersion, Mug Caustics, Prism Dispersion, and Transparency Map examples. Also, the caustics in the Refractive Caustics (Marbles) and Cardioid Caustics examples are not as sharp.

I do not understand what the "console-colors-disabled" option does?? I am glad you can add it. I just do not understand what the difference is...

Thanks,
oort

[david.bluecame]

I tested all of my example files. Turning off "Transp Refraction" makes all of the background colors work. Does this option still do what it is meant to do?

I do not think that Alpha backgrounds are working.

I am seeing lots of graininess in the Cardioid Caustics, Diamond Dispersion, Mug Caustics, Prism Dispersion, and Transparency Map examples. Also, the caustics in the Refractive Caustics (Marbles) and Cardioid Caustics examples are not as sharp.

I do not understand what the "console-colors-disabled" option does?? I am glad you can add it. I just do not understand what the difference is...

Thanks,
oort

Hi,

All these differences you noticed, respect to what YafaRay version are you comparing?

[oort]

I thought I had been testing your E builds but I guess I wasn't. I have version E2.01, E2.02, E2.1.1 on my system and they all have the same problem... never noticed... The most recent build that I have that works as expected is 993 (Yafaray 0.1.99-beta3 (2015-05-02) David Bluecame experimental exporter for Blender 2.74.5).

Maybe you changed the way this feature works when you went to the E builds. I may be missing some code you added at that time.

Edit: Here is a thread in the YafaRay forum about Transparent Refraction... Transparent Refraction

Thanks,
oort

[oort]

I had some time to investigate the problem with Alpha Background today.

The change that you made in your first E Release 1.0.0 is what caused a problem with Wings3D.

Now, if there are no parameters set in the XML for transparent background or transparent refracted background, their default values will be "false" and not "true" as until now. This will avoid confusion due to the new way "transparent refracted background" works, not rendering the background at all so volumes rendered against a transparent background do not carry "remains" of the background with them.

If we add the following to all of the lighting methods it fixes the Alpha Background problem for Wings3D. println(F, "<bg_transp bval=\"true\"/>"),

See in context below.

Even with this fix, the "Transp Refraction" option is still not working the way it did in 0.1.99 Beta 4. To see the difference, compare results between version 0.1.99 and any E version.

Maybe an option needs to be added for turning bg_transp on or off???

Thanks,
oort

case Lighting_Method of
directlighting ->
println(F," "),
println(F, "<type sval=\"~s\"/>",[Lighting_Method]),
println(F, "<bg_transp bval=\"true\"/>"),
println(F, "<bg_transp_refract bval=\"~s\"/>",[format(BackgroundTranspRefract)]),
println(F, "<raydepth ival=\"~w\"/>",[Raydepth]),
println(F, "<transpShad bval=\"~s\"/>",[format(TransparentShadows)]),
println(F, "<shadowDepth ival=\"~w\"/>",[ShadowDepth]),
println(F," ");