Why so significant difference in render results??

by **Walker** » Mon Apr 23, 2012 4:06 am

Hello all!

I have downloaded LuxRender v0.8 (no OpenCL) from http://www.luxrender.net/en_GB/standalone#windows and LuxRender weekly build from http://www.luxrender.net/forum/viewtopic.php?f=30&t=8099. Also i have a simple scene (*.lxs) with ruby. Rendering results (1 hour on 4 cores) using the above versions are very different. Here they are:

LuxRender v0.8

and

Weekly build

Is it normal such poor light in development build? And what may it cause?

by **Meelis** » Mon Apr 23, 2012 4:20 am

Maybe path depth is too low, not sure.
Does it render better with one of this rendering settings?

Code: Select all: # Main Scene File Renderer "sampler" Sampler "erpt" "integer pixelsamples" [16] "string pixelsampler" ["vegas"] "integer chainlength" [1024] "string basesampler" ["random"] Accelerator "qbvh" SurfaceIntegrator "bidirectional" "integer eyedepth" [128] "integer lightdepth" [128] "string strategy" ["all"] "float eyerrthreshold" [0.983] "float lightrrthreshold" [0.983] VolumeIntegrator "multi" "float stepsize" [0.010000000000000] PixelFilter "gaussian" "float xwidth" [3.000000000000000] "float ywidth" [3.000000000000000] "float alpha" [3.000000000000000]

Code: Select all: # Main Scene File Renderer "sampler" Sampler "random" "integer pixelsamples" [16] "string pixelsampler" ["vegas"] Accelerator "qbvh" SurfaceIntegrator "bidirectional" "integer eyedepth" [128] "integer lightdepth" [128] "string strategy" ["all"] "float eyerrthreshold" [0.983] "float lightrrthreshold" [0.983] VolumeIntegrator "multi" "float stepsize" [0.010000000000000] PixelFilter "gaussian" "float xwidth" [3.000000000000000] "float ywidth" [3.000000000000000] "float alpha" [3.000000000000000]

Can it be the first image has light source smaller (point like),
2'nt has larger light source? Or scene is identical?

by **Walker** » Mon Apr 23, 2012 4:44 am

Meelis,thank you for fast reply, but i have only one scene. And different render results on it.

by **Meelis** » Mon Apr 23, 2012 5:00 am

Give it a try without bidir?

by **Walker** » Mon Apr 23, 2012 5:33 am

Meelis wrote:Give it a try without bidir?

Images in the begining of the post were made using PATH.
I have already tried with different settings... Different surface integrators, volume integrators etc. But always luxrender v0.8 results significantly differs from weekly build results.

by **Meelis** » Mon Apr 23, 2012 6:32 am

Can you share the blend file?

by **A-man** » Mon Apr 23, 2012 6:34 am

I have the same problem while trying to port my Cornell box scene over to blender 2.6. It looks nothing at all like the original..

by **zeealpal** » Mon Apr 23, 2012 6:58 am

Just to double check, the normals are still the same in each??

by **SATtva** » Mon Apr 23, 2012 7:06 am

Walker, can you provide the scene file?

by **Walker** » Mon Apr 23, 2012 8:11 am

SATtva wrote:Walker, can you provide the scene file?

Here it is.

Ruby.7z: RubyScene; (3.5 KiB) Downloaded 17 times

Why so significant difference in render results??

Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Re: Why so significant difference in render results??

Who is online