OTOY Forums

i'm sure, most of you have already noticed that: http://www.lytro.com/ - not a totally new concept, but the first commercial product.

now i'm curious if this principle can somehow be adopted for a raytracer; i don't know enough about the mathematical basics, but to me it looks like a lightfield imager could be possible without adding overhead to the rendering process? even if dof control is a lot easier in a renderer than if taking real photos, it would of course be nice to control the dof in the final image...

t_3 wrote:i'm sure, most of you have already noticed that: http://www.lytro.com/ - not a totally new concept, but the first commercial product.

now i'm curious if this principle can somehow be adopted for a raytracer; i don't know enough about the mathematical basics, but to me it looks like a lightfield imager could be possible without adding overhead to the rendering process? even if dof control is a lot easier in a renderer than if taking real photos, it would of course be nice to control the dof in the final image...

the Lytro camera creates a zbuffer map calculating the distances with light rays, then apply a de-focus filter. This is a post effects in the chip or pc software.
When Octane will be able to save the zbuffer channel we will apply this effect too with Photoshop or After Effects.

gabrielefx wrote:

t_3 wrote:i'm sure, most of you have already noticed that: http://www.lytro.com/ - not a totally new concept, but the first commercial product.

now i'm curious if this principle can somehow be adopted for a raytracer; i don't know enough about the mathematical basics, but to me it looks like a lightfield imager could be possible without adding overhead to the rendering process? even if dof control is a lot easier in a renderer than if taking real photos, it would of course be nice to control the dof in the final image...

the Lytro camera creates a zbuffer map calculating the distances with light rays, then apply a de-focus filter. This is a post effects in the chip or pc software.
When Octane will be able to save the zbuffer channel we will apply this effect too with Photoshop or After Effects.

mh, yes and no
the lytro (or other light field cameras already built, i.e. by adobe) don't just apply post-blurring to the image, it is a bit more complex: http://www.lytro.com/renng-thesis.pdf (warning: a tough read )

also to apply blurring only by z-depth can't reproduce real world dof (like octane already - and perfectly - delivers), where background object details become "visible" even if fully covered by foreground objects. of course post-dof by z-depth is sufficient for many situations, but if you try to recreate real world photography it (imo) fails and tends to make images look unnatural...