Does anyone know of a tech for unbiased GI with caching?
Posted: April 20, 2010
Post subject: Does anyone know of a tech for unbiased GI with caching?
Post subject: Does anyone know of a tech for unbiased GI with caching?
Hey,
I have done a brief search and I can't locate any global illumination techniques that solve the rendering equation in an unbiased way for a full range of BSDF models while also using a large cache of previous radiance calculations.
I have found some techniques that solve the rendering equation in an unbiased manner: Path-tracing, Bi-directional Path-tracing, Metropolis Light Transport, etc. But none of them reuse much data, and none of them have a cache of any sort.
I have also found some techniques that use a cache, but they seem to use limited reflection models (diffuse only) or other approximations that add bias to the solutions they produce: Radiosity (diffuse only, generalisation of irradiance across patches), Photon Mapping (interpolation between photon samples), Irradiance Cache (diffuse only), etc.
Are there any techniques that use a cache, but ALSO provide an unbiased solution to the rendering equation AND allow a full range of BSDF models?
I have done a brief search and I can't locate any global illumination techniques that solve the rendering equation in an unbiased way for a full range of BSDF models while also using a large cache of previous radiance calculations.
I have found some techniques that solve the rendering equation in an unbiased manner: Path-tracing, Bi-directional Path-tracing, Metropolis Light Transport, etc. But none of them reuse much data, and none of them have a cache of any sort.
I have also found some techniques that use a cache, but they seem to use limited reflection models (diffuse only) or other approximations that add bias to the solutions they produce: Radiosity (diffuse only, generalisation of irradiance across patches), Photon Mapping (interpolation between photon samples), Irradiance Cache (diffuse only), etc.
Are there any techniques that use a cache, but ALSO provide an unbiased solution to the rendering equation AND allow a full range of BSDF models?
Thanks Sam! Posted.
I'm not sure if I have already responded to this... It seems I'm not human enough to read those numbers
I'm not sure if I have already responded to this... It seems I'm not human enough to read those numbers