# April 21th progress

- [x] cube map completed
- [x] make images black, and create a line exteremly bright
- [x] compare environment map with default pbrt environment map
- [x] buddha scene
- [ ] vol path single scattering for point-line mis

## verify cube map

* I created these images to verify the cube map is exactly the same as pbrt default environment map by looking at six faces respectively.

More detail in images.zip and image2.zip

Images name with “test” are rendered using line sampling using cubemap and “point” refers to point sampling using latitude and longitude environ map

In images_revise.zip and image2_revised.zip, there are results with roughly equal noise compaison.

## buddha scene (100k primitives)

* Based on the assumption that the cube map is correct from the above experiment, I used buddha as a more complex example to test.

buddha scene with point sampling(32 samples)

buddha scene with line sampling(32 samples)

#### After raising exposure
* the noise does not match exactly but it already taken a lot time(> 1hr) to render the buddha scene with line sampling using 64 samples.

buddha scene with point sampling(1024 samples)

buddha scene with point sampling(2048 samples)

buddha scene with line sampling(64 samples)

## create a line exteremly bright

* I think this is just an example illustrating that line sampling is powerful in this extreme cases.

* I think the results are also interesting since the plane part, line sampleing performs better even the buddha itself it is not as good as point.

I create an environment map like the following using PS

rendering it with the cube scene

point sampling (64 samples)

line sampling (64 samples)

#### After raising exposure

point sampling(4096 samples)

line sampling (64 samples)