Low-discrepancy adaptive 10 to 64 SPP, 0.70 confidence, 2334ms

Low-discrepancy adaptive 8 to 64 SPP, 0.99 confidence, 2725ms

Monte-Carlo adaptive 8 to 64 SPP, confidence 0.99, 2632ms

Low-discrepancy adaptive 8 to 64 SPP, 0.80 confidence

Monte-Carlo adaptive 8 to 64 SPP, 0.80 confidence

Low-discrepancy adaptive 10 to 64 SPP, 0.70 confidence, 2334ms

Low-discrepancy adaptive 8 to 64 SPP, 0.99 confidence, 2725ms

Monte-Carlo adaptive 8 to 64 SPP, confidence 0.99, 2632ms

Low-discrepancy adaptive 8 to 64 SPP, 0.80 confidence, 2187ms

Monte-Carlo adaptive 8 to 64 SPP, 0.80 confidence, 2110ms

Low-discrepancy adaptive 4 to 16 SPP

Monte-Carlo adaptive 4 to 16 SPP, 0.99 confidence

Low-discrepancy adaptive 4 to 16 SPP, 0.99 confidence, 1355ms

Monte-Carlo adaptive 4 to 16 SPP, 0.99 confidence, 1292ms // all adaptive SPP configurations use δ = 1/256

Low-discrepancy 16SPP, 2450ms

Stratified 16SPP, 2429ms

Monte-Carlo 16SPP, 2470ms

Low-discrepancy 4SPP, 655ms // kth samples screen-space coordinates determined by a per pixel uniform random offset + kth number of the base 2 or 3 Halton sequence for each dimension. If the offset plus the Halton number is 1, 1 is subtracted from it.

Stratified 4SPP, 650ms

Monte-Carlo 4SPP, 648ms // all images above use 8 bounces, tile size 8, 1280x720, i5-12600K 16 threads, no texture filtering

teapot-textured.xml, low-discrepancy multisampling with Halton + offset // 128 spp, 8 bounces, tile size 8, 27230.3ms, 1280x720, i5-12600K 16 threads // no gamma correction, no tone mapping, nearest texel texture filtering