Deep G-Buffers for Stable Global Illumination Approximation

Proceedings of High Performance Graphics 2016

Michael Mara, NVIDIA and Stanford University
Morgan McGuire, NVIDIA and Williams College
Derek Nowrouzezahrai, University of Montreal
David Luebke, NVIDIA

Extended Version (15 pages, more details than the proceedings version):

Proceedings Version (12 pages, 47 MB PDF)
Demo (108 MB Windows EXE + Source)
Video results (YouTube)
Video results (94 MB MP4)


We introduce a new hardware-accelerated method for constructing Deep G-buffers that is 2x-8x faster than the previous depth peeling method and produces more stable results. We then build several high-performance shading algorithms atop our representation, including dynamic diffuse interreflection, ambient occlusion (AO), and mirror reflection effects.

Our construction method s order-independent, guarantees a minimum separation between layers, operates in a (small) bounded memory footprint, and does not require per-pixel sorting. Moreover, addressing the increasingly expensive cost of pre-rasterization, our approach requires only a single pass over the scene geometry. Our global illumination algorithms approach the speed of the fastest screen-space AO-only techniques while significantly exceeding their quality: we capture small-scale details and complex radiometric effects more robustly than screen-space techniques, and we implicitly handle dynamic illumination conditions. We include the pseudocode for our Deep G-buffer construction in the paper and the full source code of our technique in our supplemental document.


Selected Images


  author = {Michael Mara and Morgan McGuire and Derek Nowrouzezahrai and David Luebke},
  title = {Deep G-Buffers for Stable Global Illumination Approximation},
  month = {June},
  day = {24},
  year = {2016},
  numpages = {11},
  booktitle = {HPG},
  url = {}