VW scene

[wahn]

I will add a new test scene to the GitLab repository (for test scenes):

This will probably spawn more issues to be created. For example we will have to implement a new material, FourierMaterial:

$ rg bsdf -B 1  
vw-van.pbrt
54-MakeNamedMaterial "Mat.1"  # white painted stuff
55:    "string type" "fourier" "string bsdffile" "bsdfs/ceramic.bsdf"

Which can read .bsdf files. I'm also not sure if the already existing MetalMaterial can read .spd files:

$ rg spd -B 3
vw-van.pbrt
49-MakeNamedMaterial "Metal - Chrome" 
50-        "float roughness" [ 0.01 ] 
51-        "string type" [ "metal" ] 
52:        "spectrum k" "spds/Al.k.spd"
53:        "spectrum eta" "spds/Al.eta.spd"

Goal of this issue is to document the progress until we can render the entire VW test scene without changes ...

[wahn]

After checking in the scene the rendering via the current version (v042) fails:

$ ~/git/github/rs_pbrt/target/release/examples/rs_pbrt -i vw-van.pbrt
pbrt version 0.4.2 [Detected 8 cores]
Copyright (c)2016-2018 Jan Douglas Bert Walter.
Rust code based on C++ code by Matt Pharr, Greg Humphreys, and Wenzel Jakob.
thread 'main' panicked at '  --> 52:9
   |
52 |         "spectrum k" "spds/Al.k.spd"
   |         ^---
   |
   = expected statement, parameter, or last_statement', examples/rs_pbrt.rs:468:80
note: Run with `RUST_BACKTRACE=1` for a backtrace.

[wahn]

There are a couple of parser related issues which will be fixed via the ticket #57 ...

[wahn]

The following named material needs attention:

$ grep 'MakeNamedMaterial "Metal' ~/git/gitlab/rs-pbrt-test-scenes/pbrt/pbrt_vw_van/vw-van.pbrt  -A 4
MakeNamedMaterial "Metal - Chrome" 
        "float roughness" [ 0.01 ] 
        "string type" [ "metal" ] 
        "spectrum k" "spds/Al.k.spd"
        "spectrum eta" "spds/Al.eta.spd"

The two lines for "spectrum k" and "spectrum eta" will result in C++ calls to pbrt::ParamSet::AddSampledSpectrumFiles:

void ParamSet::AddSampledSpectrumFiles(const std::string &name,
                                       const char **names, int nValues) {
...
}

We don't have a Rust counter part to that function (yet). Time to open another issue ...

[wahn]

See issue #59 ...

[wahn]

After commit 3f79674 we can at least render the scene:

The two files for "spectrum k" and "spectrum eta" are read (but not used yet?), the FourierMaterial isn't implemented yet, and we still do not handle the keywords ObjectBegin, ObjectEnd, and ObjectInstance (see issue #58) correctly ...

[wahn]

After commit e7ab8e5 the resulting image rendered by the Rust version (rs_pbrt) should look pretty similar to the first image shown (rendered by the C++ version), but something goes wrong regarding the car paint (for the diff image I rendered a version with less samples):

> imf_diff -d -f pbrt_rust.exr vw-van.exr diff.jpg
differing pixels:	 27.186% (434975 of 1600000)
average difference:	  4.986%
maximum difference:	 57.942%
Summary: Many pixels differ strongly.
== "pbrt_rust.exr" and "vw-van.exr" are different

The FourierMaterial should work though ...

[wahn]

This is the image rendered with the current Rust version:

[wahn]

After fixing SpecularReflection::pdf() (which returns always 0) we get a much better match (see commit 153b0ea):

> imf_diff -f -d vw-van.exr pbrt_rust.exr diff.jpg
differing pixels:	  0.278% (4442 of 1600000)
average difference:	 27.465%
maximum difference:	 80.857%
Summary: Some pixels differ strongly.
== "vw-van.exr" and "pbrt_rust.exr" are different

So both images (C++ vs. Rust) should match close enough.

Closing the task/issue ...