Dear all,
I currently render a stereoscopic version of a camera walk through. The irradiance maps and lightcaches are already calculated for the entire scene for the center camera in flythrough mode / multiframe incremental, since the monoscopic version of the movie was rendered in fall last year. Now the customer wants a stereoscopic version of the movie
here is the workflow (nothing special):
1. Switch vrayStereoscopicHelper to create shademaps, store them on a network drive where all render servers can access files, use base file name: shademap_.vrst. Adjust the irradiance map and lightcache to load from precalculated file. The stereoscopic helper has also the following settings:
- eye distance about 30mm
- specify focus on
- shear, no rotation
- shift both
- view both
- camera view is the center camera
2. Render the scene with just the center camera, store also an rgb-image for comparison.
3. Switch vrayStereoscopicHelper to use shademap, standard setting reuse threshold 1.0. Locate the same folder and use the base shademap filename, for example shademap_.vrst. Render now the stereoscopic view with left and right image (3840 x 1080).
I don't see any performance gain as the rendertime for the stereoscopic view is twice as high as rendering the center view for the shademap, so generating and using the shademap takes even longer than just rendering the stereoscopic view without shademap.
But I made a test before with the same irradiance map and lightcache, but in a very small rendering resolution (320 x 180 instead of Full HD), and there I had a huge performance gain. But this was not working with the full HD image.
Is any of the following settings the issue?
- a scene with a lot of displacement maps and reflections
- Geometry is always referenced as xref-scenes
-lightcache and irradiance map are used from stored files in both steps (creation and reuse of shademap).
-Antialiasing is switched on for shademap creation, today I read in the forum that this is not recommended for the shademap generation?
- When looking at the frame buffer, I just see some geometry (5% of the image) that is visible in the loaded shademap pass and looks like the rendered image, all other areas are black.
- i use vray advanced 2.1
- irradiance map and lightcache were calculated for 1280 x 720 pixels, now i use them for FullHD
Would be great if anybody could help.
Thanks
Robert
I currently render a stereoscopic version of a camera walk through. The irradiance maps and lightcaches are already calculated for the entire scene for the center camera in flythrough mode / multiframe incremental, since the monoscopic version of the movie was rendered in fall last year. Now the customer wants a stereoscopic version of the movie
here is the workflow (nothing special):
1. Switch vrayStereoscopicHelper to create shademaps, store them on a network drive where all render servers can access files, use base file name: shademap_.vrst. Adjust the irradiance map and lightcache to load from precalculated file. The stereoscopic helper has also the following settings:
- eye distance about 30mm
- specify focus on
- shear, no rotation
- shift both
- view both
- camera view is the center camera
2. Render the scene with just the center camera, store also an rgb-image for comparison.
3. Switch vrayStereoscopicHelper to use shademap, standard setting reuse threshold 1.0. Locate the same folder and use the base shademap filename, for example shademap_.vrst. Render now the stereoscopic view with left and right image (3840 x 1080).
I don't see any performance gain as the rendertime for the stereoscopic view is twice as high as rendering the center view for the shademap, so generating and using the shademap takes even longer than just rendering the stereoscopic view without shademap.
But I made a test before with the same irradiance map and lightcache, but in a very small rendering resolution (320 x 180 instead of Full HD), and there I had a huge performance gain. But this was not working with the full HD image.
Is any of the following settings the issue?
- a scene with a lot of displacement maps and reflections
- Geometry is always referenced as xref-scenes
-lightcache and irradiance map are used from stored files in both steps (creation and reuse of shademap).
-Antialiasing is switched on for shademap creation, today I read in the forum that this is not recommended for the shademap generation?
- When looking at the frame buffer, I just see some geometry (5% of the image) that is visible in the loaded shademap pass and looks like the rendered image, all other areas are black.
- i use vray advanced 2.1
- irradiance map and lightcache were calculated for 1280 x 720 pixels, now i use them for FullHD
Would be great if anybody could help.
Thanks
Robert
Comment