belcour / alta Goto Github PK

Hi,
I am just making a new attempt to use alta for fitting models to measured BSDF. Unfortunately, I had no success building on my current MacOS Monterey 12.4. My guess is that this is mainly due to some python incompatibilies, but I cannot decipher the error messages. So some observations.

• I have scons and most dependencies installed via macports. For some reason, ceres is not found (macports installed libceres.a in /opt/local/lib).
• I have python 3.10 as default, and I run scons --cfg=configs/scons/config-macos-clang.py
• scons then gives me several warning and finally throws an error (I cut the previous checks):

Checking for C++ header file unsupported/Eigen/LevenbergMarquardt... yes
scons: warning: Calling missing SConscript without error is deprecated.
Transition by adding must_exist=False to SConscript calls.
Missing SConscript 'build/sources/plugins/rational_fitter_eigen/SConscript'
File "/Users/lars/git/alta/sources/plugins/SConscript", line 52, in build_plugin
scons: warning: Ignoring missing SConscript 'build/sources/plugins/rational_fitter_leastsquare/SConscript'
File "/Users/lars/git/alta/sources/plugins/SConscript", line 52, in build_plugin
scons: warning: Ignoring missing SConscript 'build/sources/plugins/rational_fitter_quadprog/SConscript'
File "/Users/lars/git/alta/sources/plugins/SConscript", line 52, in build_plugin
scons: warning: Ignoring missing SConscript 'build/sources/plugins/rational_fitter_parallel/SConscript'
File "/Users/lars/git/alta/sources/plugins/SConscript", line 52, in build_plugin
TypeError: 'NoneType' object is not iterable:
File "/Users/lars/git/alta/SConstruct", line 341:
plugins = alta_sconscript('sources/plugins/SConscript')
File "/Users/lars/git/alta/SConstruct", line 335:
return env.SConscript(script,
File "/opt/local/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/SCons/Script/SConscript.py", line 597:
return _SConscript(self.fs, *files, **subst_kw)
File "/opt/local/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/SCons/Script/SConscript.py", line 285:
exec(compile(scriptdata, scriptname, 'exec'), call_stack[-1].globals)
File "/Users/lars/git/alta/sources/plugins/SConscript", line 78:
a_flat_list_of_targets = list( itertools.chain(*(list(targets))))

I tried the tutorial ( http://alta.gforge.inria.fr/tutorial2.html ) and it was a success to generate '.func' file.

but when i tried this ( http://alta.gforge.inria.fr/tools.html )

brdf2data --input test.func --output test.binary --data data_merl

It failed.

ERROR failed to load '': No such file or directory.'

This command also has an error.

brdf2brdf --input test.func --output test.shader --export explorer

INFO parsed input parametrization STARK_2D from name "STARK_2D"
ERROR not implemented virtual void alta::rational_function_1d::save_body(std::ostream&, const alta::arguments&) const in file sources/core/rational_function.cpp:147
terminate called without an active exception
Aborted

Hello,

I meet some compilation issues.
Compiler:
Visual Studio 2013 - x64

I modify my config.py by adding:
MSVC_VERSION = '12.0'
TARGET_ARCH = 'x86_64'

And:
OPENEXR_INC = ['#external/build/include/OpenEXR']
OPENEXR_LIB = ['Half.lib', 'Iex-2_2.lib', 'IexMath-2_2.lib', 'IlmImf-2_2.lib', 'IlmThread-2_2.lib', 'IlmImfUtil-2_2.lib', 'Imath-2_2.lib']
OPENEXR_DIR = ['#external/build/lib']

OpenEXR was compiled by myself using the lastest from (same compiler etc...):
https://github.com/openexr/openexr (6f6d9ce)

Eigen 3 was installed and pass all the test but I still have:
Checking for eigen3 using pkg-config... (cached) no
Checking for C++ header file Eigen/Core... (cached) yes
Checking for C++ header file unsupported/Eigen/LevenbergMarquardt... (cached) yes

Why "eigen3 "no""?

For OpenEXR still ok:
Checking for C++ library Half.lib... (cached) yes
Checking for C++ header file ImfStdIO.h... (cached) yes

I want to use brdf_slide but I have linking issues:
scons: done reading SConscript files.
scons: Building targets ...
link /dll /out:build\plugins\data_brdf_slice.dll /implib:build\plugins\data_brdf_slice.lib /LIBPATH:external\b
uild\lib /LIBPATH:build\core Half.lib Iex-2_2.lib IexMath-2_2.lib IlmImf-2_2.lib IlmThread-2_2.lib IlmImfUtil-
2_2.lib Imath-2_2.lib Half.lib core.lib build\sources\plugins\data_io\slice.obj
Microsoft (R) Incremental Linker Version 12.00.31101.0
Copyright (C) Microsoft Corporation. All rights reserved.

Creating library build\plugins\data_brdf_slice.lib and object build\plugins\data_brdf_slice.exp
slice.obj : error LNK2001: unresolved external symbol "private: static union half::uif const * const half::_to
Float" (?_toFloat@half@@0QBTuif@1@B)
slice.obj : error LNK2001: unresolved external symbol "private: static unsigned short const * const half::_eLu
t" (?_eLut@half@@0QBGB)
build\plugins\data_brdf_slice.dll : fatal error LNK1120: 2 unresolved externals
scons: *** [build\plugins\data_brdf_slice.dll] Error 1120
scons: building terminated because of errors.

If I check the Mangling name from the dll/lib:
?_eLut@half@@0QBGB
?_toFloat@half@@0QBTuif@1@B

As you can see it is different:
ATLA expect "?_toFloat@half@@0QBTuif@1@B" the compiler generate "?_toFloat@half@@0QBTuif@1@B"
ATLA expect "?_eLut@half@@0QBGB" the compiler generate "?_eLut@half@@0QBGB"

The website (http://alta.gforge.inria.fr/install.html) do not provide the version of OpenEXR for now I'm using 2.2.0.

Did you know how could I by pass those issues.

The rest compile correctly thanks.

Best Regards

Chkone

For me the Python module could not be compiled with SCons, until the PYTHON_LIB variable was added to config-windows-cl.py

I tried the tutorial on this page with the same material.

http://alta.gforge.inria.fr/tutorial2.html

However, my L2 error is bigger than what was provided in the tutorial page.

This is the header from ' http://alta.gforge.inria.fr/rational-example-01.brdf '.

#CMD --L2 0.0142601 --Linf 0.240069 --fitter rational_fitter_leastsquare --func rational_function_chebychev

This is the header following the tutorial.

#CMD --L2 3.25913 --Linf 231.828 --fitter rational_fitter_leastsquare --func rational_function_chebychev

My L2 error is bigger than what it should be.

Could it be any bug in the code ?

Hi,

Another suggestion will be to support OSL as an equivalent of "save_body" to make it more portable.

Best,
Stéphane

Bonjour ;-)

First of all, thank you very much, it is a great Swiss army knife for BRDFs !

I was playing with it, and I got an issue with BRDFGrid. This is what I did :

./softs/data2data --input ./softs/red-fabric.binary --in-data data_merl --output red-fabric.grid --out-data data_grid  

./softs/data2data --input ./red-fabric.grid --in-data data_grid --output red-fabric.exr --out-data data_brdf_slice --param RUSIN_TH_TD_PD --angle 90  

That seems normal ( based on the source code ) because you are calling abort() in plugins/data_interpolation/grid.cpp:376

Why are you doing this, is it safe to bypass it ?

The question is how to convert Alta's BRDF to something else ?

Many thanks,

Stéphane

I tried using 'scons' to build the project but eigen seems to be not working.
The auto download might be pointing to an empty URL.

After I built with scons, I tried testing with
scons tests

There was an error about downloading file.

error: downloaded file 'build/tests/gold-metallic-paint.binary' is inauthentic.
error: got sha256 hash 7df0af666610f38658b54d8a8b32a7cdabdb0aadef37c66c3af9d93e709a4a8f but expected 69c493d7f36fa5267d15948a616056f9bc84e352ec6db49b64afe6ff232e27d6
scons: *** [build/tests/gold-metallic-paint.binary] build/tests/gold-metallic-paint.binary: downloaded file is inauthentic
scons: building terminated because of errors.

When running the first simple tutorial with
build\softs\data2brdf --input .\Kirby2.dat --output .\Kirby2.brdf --fitter rational_fitter_eigen
(build with config-windows-cl.py with SCons) I got the following errors:

A .pyd is not created during build process. The solution was to rename alta.dll to alta.pyd in the built python folder.

Hi,

while having a simple gnuplot interface would be very useful, the use of brdf2gnuplotis not really documented. I got that far by now:

brdf2gnuplot --polar-plot --inc_angle 0.5 --samples 180 --data-file gold-metallic-paint.dat --input gold-metallic-paint_lafortune_diffuse_Eigen.func --output gold-metallic-paint_lafortune_diffuse_Eigen.gpl

This produces a file with six columns and 145800 rows. Unfortunately, there is no header explaining the field contents. The first column monotonously increases from 0 to approx 1.6. The second also contains values in that range, possibly interleaved with the first. The third is in the range 0 to 3.2. And the next three must be related to the BRDF.

Is there any documentation available on the format, or an example how to use it to produce a gnuplot figure?

Best, Lars.

Hi Laurent,
if I want to submit code to ALTA I can't do that via its gitlab repo at INRIA, but I could use this clone of yours, couldn't I?
Greets,
Marko

Hi,

Recently PBRT (v3) and Mitsuba started to support FourrierBSDF
(https://github.com/mmp/pbrt-v3/blob/master/src/materials/fourier.cpp)
and the creation pipeline is based on https://github.com/wjakob/layerlab.

So supporting this new format as a plugin for input and output will be really usefull.

Best Regards,

stephane-lb

Associated publication : "A Comprehensive Framework for Rendering Layered Materials' by
Wenzel Jakob, Eugene D'Eon, Otto Jakob and Steve Marschner"

P.S : Could do to do it soon.

import alta
d1 = alta.get_data('data_merl')
Traceback (most recent call last):
File "", line 1, in
TypeError: get_data(): incompatible function arguments. The following argument types are supported:
1. (arg0: str, arg1: int, arg2: alta.parameters) -> alta.data
2. (arg0: str, arg1: int, arg2: alta.parameters, arg3: alta::arguments) -> alta.data

Invoked with: 'data_merl'

I think this is a bug in python binding.