We compare our gradient path tracing (G-PT) method using L1 and L2 reconstruction to ordinary path tracing (PT). We show renderings with twice as many base samples for PT compared to G-PT. This provides an approximately equal time comparison, since the overhead of G-PT per base path is about a factor of 2.5. In other words, at exact equal render time the PT images would be slightly less noisy. The reference images were rendered with bidirectional path tracing. We measure the error as relative mean square error (relMSE).

Instructions: Click on the thumbnails to show the images at their original resolution.

Bathroom 2048spp

	G-PT L1 relMSE: 0.001984 PT (4096spp) relMSE: 0.012299	G-PT L2 relMSE: 0.002107 Reference 40000spp

Bathroom 256spp

	G-PT L1 relMSE: 0.009924 PT (512spp) relMSE: 0.092130	G-PT L2 relMSE: 0.011004 Reference 40000spp

Bookshelf 2048spp

	G-PT L1 relMSE: 0.000582 PT (4096spp) relMSE: 0.003673	G-PT L2 relMSE: 0.001090 Reference 18000spp

Bookshelf 256spp

	G-PT L1 relMSE: 0.003740 PT (512spp) relMSE: 0.028595	G-PT L2 relMSE: 0.005819 Reference 18000spp

Kitchen 2048spp

	G-PT L1 relMSE: 0.001481 PT (4096spp) relMSE: 0.006899	G-PT L2 relMSE: 0.001708 Reference 32000spp

Kitchen 256spp

	G-PT L1, 256spp relMSE: 0.007571 PT, 512spp relMSE: 0.050493	G-PT L2, 256spp relMSE: 0.009003 Reference 32000spp

Sponza 64spp

	G-PT L1 relMSE: 0.000798 PT (128spp) relMSE: 0.011806	G-PT L2 relMSE: 0.000872 Reference 24000spp

Sponza 8spp

	G-PT L1 relMSE: 0.006781 PT (16spp) relMSE: 0.093989	G-PT L2 relMSE: 0.006588 Reference 24000spp

Veach-Door 2048

	G-PT L1 relMSE: 0.005791 PT (4096spp) relMSE: 0.026903	G-PT L2 relMSE: 0.006503 Reference 32000spp

Veach-Door 256

	G-PT L1 relMSE: 0.043763 PT (512spp) relMSE: 0.209632	G-PT L2 relMSE: 0.048758 Reference 32000spp

This webpage was built using the code found on: http://www.cs.huji.ac.il/~raananf/projects/lss_upscale/sup_images/index.html.