An implementation of the renderscore. It enables the comparison of MC rendering algorithms (using Mitsuba) as outlined in the corresponding technical report.

Python 2.7:

• numpy
• scipy

Additional:

• Mitsuba Renderer

1. Setup your Python installation. (for Anaconda users see .conda.yml)

2. Download or build Mitsuba (see documentation to ensure that Python will be able to find the Mitsuba core libraries)

3. Copy or install the this module.

1. Setup Mistuba XML files for the reference, e.g. "./data/veach_ajar/bdpt.xml".
2. Setup Mistuba XML files for the test / comparison, e.g. "./data/veach_ajar/path.xml" and "./data/veach_ajar/erpt.xml"
3. Make sure sampler parameters match for all of the used scenes.
4. Choose iteration count (e.g. 1024 for the reference and 32 for the test runs) and execute:

python -m renderscore 1024 ./data/veach_ajar/bdpt.xml 32 ./data/veach_ajar/path.xml ./data/veach_
ajar/erpt.xml 

For detailed command line parameters see help:

python -m renderscore -h

Christian Freude, freude (at) cg.tuwien.ac.at

• 1.0.0
  • Initial Release

This project is licensed under the GNU GPL LICENSE - see the LICENSE.md file for details

Funded by Austrian Science Fund (FWF): ORD 61