| A | B | |
|---|---|---|
1 | Name | Description |
2 | gproject | type of a spherical projection used for sampling a triangular light by the G sampling technique: |
3 | "spheresimple" - uniformly sample a planar triangle defined by the original corners projected onto a sphere | |
4 | "sphereprecise" - uniformly sample true spherical projection of the triangle | |
5 | lproject | type of a projection used for sampling a triangular light by the L sampling technique: |
6 | "none" - uniform sampling of the light surface as is | |
7 | "parallelplane" - uniform sampling of the light projected on a plane parallel with the shaded surface in unit distance | |
8 | lsat | type of the A light selection technique |
9 | "P" - selection according to the total light power (constant over the scene) | |
10 | "S" - selecting the strongest light only (shading point dependent) | |
11 | "NS" - selection according to estimated light contribution, the strongest light is excluded (shading point dependent) | |
12 | lselect | light selection technique: |
13 | "A" - additional technique defined by the lSAT parameter | |
14 | "E" - selection according to estimated light contribution (shading point dependent) | |
15 | "U" - uniform selection (constant over the scene) | |
16 | "AE", "AU", "EU", "AEU" - combinations using weights defined by the misWeights parameter | |
17 | "ALL" - no random selection, one sample is taken from each light | |
18 | misweightsa | MIS weights used for combining samples when estimating the alphas: "balance", "power" |
19 | misweights | MIS weights used for combining the sampling techniques or the light selection techniques: "balance", "power", "optimal" |
20 | moredatalayers | if true, the resulting exr file will contain additional layers with auxiliary data (alpha and weight for each technique, average relative sample variance), false by default |
21 | optimalmode | how the optimal weights are computed and used |
22 | "alphaSum" - the result is the sum of estimated optimal alphas | |
23 | "alphaSumL2" - the result is the sum of estimated alphas corresponding to L2 minimization | |
24 | "alphaSumIndB" - the result is the sum of estimated optimal alphas, the A matrix and the b vector use independent sets of samples | |
25 | "fullweights" - the result is obtained by weighting sampler.samplesCount samples by the optimal weights with alphas estimated from a different set of trainSamples samples | |
26 | "progressive" - the result is obtained by weighting sampler.samplesCount samples by the optimal weights with alphas estimated from previous samples | |
27 | "alphasumprog" - the result is the sum of estimated optimal alphas, each pixel has its own random seed so renderings with increasing number of samples are better comparable | |
28 | seedoffset | offset added to each tile sampler seed |
29 | skipTech | index of a technique skipped from estimation in the method of Fan et al. (-1 = our method, 0 = Fan's method skipping the first technique, ..., N+1 = Fan's method skipping the intercept term) |
30 | targettime | if positive, number of seconds the rendering should run for instead of the number of samples, -1 by default |
31 | technique | sampling technique (for sampling a direction on a hemisphere): |
32 | "B" - brdf sampling | |
33 | "G" - spherical light projection sampling (projection defined by the gProject parameter) | |
34 | "L" - light surface sampling (projection defined by the lProject parameter) | |
35 | "BG", "BL", "GL", "BGL" - combinations using weights defined by the misWeights parameter | |
36 | trainsamples | number of samples taken from each technique in preprocess to etimata the alphas (int) |
37 | updatestep | number of iterations perfomed between two updates of the technique matrix and the contribution vector (1 = update after every iteration, 2 = update after every two iterations, ...) |