Monte Carlo Path Tracing

Part of Brown University CSCI2240 assignment.

Description

This project contains code to perform Monte Carlo Path Tracing written in C++ and was built on top of the Stencil code provided to us by the instructors of CSCI2240.

Requirements

Qt ( > 5.9.0)
Qt Creator ( > 4.5.0)
OpenMP

Implementation

Four basic types of BRDFs ​✔

Diffuse [640 x 640] [2000 Samples]

Mirror [640 x 640] [2000 Samples]

Glossy [640 x 640] [3000 Samples]

Refraction [640 x 640] [1500 Samples]

Soft Shadows and Indirect Illumination ​✔

Soft Shadows and Colour Bleeding Caustics

Russian Roulette path termination ​✔

Tone Mapping ​✔

Event splitting ​✔

Extra Features

Attenuate refracted paths ​ ​✔

Reference : https://www.scratchapixel.com/lessons/3d-basic-rendering/global-illumination-path-tracing

Importance Sampling ​✔

With importance sampling, the path tracer converges within 2000 samples per pixel, while the naive uniform sampling technique still produces noise for the same samples per pixel.

Reference: https://www.tobias-franke.eu/log/2014/03/30/notes_on_importance_sampling.html

Depth of field ​✔

Other Scenes

Usage

Build :

• Using Qt Creator.
• Using CLI:

cd Path_Tracer
mkdir build
cd build
qmake -makefile ../path-stencil.pro
make -j

Run :

• Using Qt Creator Set the following arguments in Qt Creator.

# <path to xml file> <rendered image path> <number of samples> <image height> <image width>
../Path_Tracer_2240/example-scenes/CornellBox-Sphere.xml ./output.png 100 256 256

• Using CLI : ./path-stencil ../Path_Tracer_2240/example-scenes/CornellBox-Sphere.xml ./output.png 100 256 256