Resolution
Render time (hrs)
Samples
Faces
- We decided to keep our scene simple and derived inspiration from things we saw in
our neighbourhood. To maximize the features of our ray tracing engine, we experimented a lot and arrived at a
hopefully good scene.
- We used Blender to model our scene. Some of the objects in the scene were hand modelled as well.
- We highlight the major features used in the scene here.
- Disclaimer: It is a night scene and hence is better in night time!!
The Basics
Features
Post Processing
- The bloom effect produces fringes of light extending
from the bright area in an image. This effect creat an illusion of intense bright light happening
in the scene. To get this effect, we first extract the bright area of the rendered image. We apply
a 2D convolution to blur the extracted brightness image then add the blurred image (scaled by a
constant factor and pixel intensities) to the original image. We repeat the convolution and
addition several times to produce the final image that has the bloom effect.
- /rt/renderer.cpp
- As a post processing step, we do do a bit of contrast
adjustment and then gamma correction for our scene. To reproduce image textures (environment, book
texture images) exactly how we percieve them on our screens, we apply gamma when we read the
images. Works wonders when we correct gamma back, the result of which is the scene here!! Color
correction is really important!
- rt/renderer.cpp; core/image.cpp; core/color.cpp
- We would like to thank Prof. Slusallek, our tutor Philippe Weier and
the others too for helping us throughout the semester. And our friends for their valuable and creative inputs
and discussions when creating our scene. And lastly, past year submissions for their amazing renders. It was a
pleasure thinking about how a particular effect is achieved.
- The more complex models were downloaded and are linked here. Book,Vase,pen,penholder,orb,curtain. The entire scene contains approximately 100K triangles.
- The website was built using Nicepage.
This a joint project implemented by the members of group1 - Adarsh Djeacoumar and Anh Tuan Tran.