Day 18 -- 30/6/2011 Thursday

We had lesson from Zack, intern in SESI. He was more into lighting/rendering so he was appointed to teach us. He taught us the fundamental of rendering in Houdini. Before he began, I thought to myself, hmm, "I know something about rendering, it should be easy". He went through the trouble of explaining each steps in detail. I learned quite a lot of stuff along the way.

Basic stuff like saving rendered image: $HIP/filename/imagename.png (whatever file extension)

Depending on what extension, each hold different types of channel. 

.TIFF has RGB channel + alpha channel

.PNG has alpha channel

.EXR has all channel

.RAT is default Houdini texture file format

When rendering passes, best to use open EXR or Houdini.pic format. 

Some basic information in saving image and what type of format to use. Next was how each rendering engine works and how to optimize each engine to get nice rendering result. There are few steps to take note when rendering a scene. The first step of rendering is geometry refinement. In a scene, the user need to decide which rendering engine is more suitable, MicroPolygon or Ray-tracing engine. The concept of Micropolygons was developed within the Reyes Algorithm, in which geometric primitives are tessellated at render time into a rectangular grid of tiny, four-sided polygons. For faster rendering time, use Micropolygon but it might produce less quality motion blur. Ray-tracing will take much longer but better image produced and less noisy motion blur.

To optimize MicroPolygon rendering
Mantra-Node>properties>sampling> Min ray/ Max ray
Mantra-Node>properties>dicing> Shading quality
Pixel Sample for the last resort.

To optimize Ray-tracing rendering
Mantra-Node>properties>dicing> Ray-tracing shading quality
Mantra-Node>properties>render> ray-tracing accelerator
Pixel Sample for the last resort.

To optimize MicroPolygon Physically Based Rendering
Mantra-Node>properties>sampling> Min ray/Max ray
Mantra-Node>properties>PBR> Color limit, lower number to remove white dots
Mantra-Node>properties>dicing> Shading quality
Volume step size for rendering volume like fluid/fog

Sampling quality in light sop varies to mantra pixel sample. In order to get more accurate sampling quality, adjust mantra pixel sample then the light sample quality. He also talk about how to render wireframe.

Wire Frame render> Create any geometry (polygon/polygon mesh), put down a facet sop with unique points checked, go to OUT in the operation level gadet and tab in a "wren" sop, equip with a camera and hit render with the "wren" sop.

I continued with my own stuff for the rest of the day.