Tweet
Hi,
Would it be difficult to add the possibility to merge multiple EXR images into one single EXR file with the vrimg2exr tool (and preferably via commandline)?
We "tile render" large still image renders (using the vray -region flag or Maya's -reg flag) on our render farm (Windows and Linux machines). As a post-render operation I use a bit of an obscure, precompiled and Windows-only OpenImageIO Python binding to read the buffers for each EXR image and then just combine them all (using "over", A+B, for all channels) before writing out the resulting buffer as a new, "merged" EXR. Since the canvas size of the fullsize image is recorded in each EXR tile, there's no need to specify coordinates for each tile. You just have to combine the buffers, so it seems like it should not have to involve too much development time.
The vrimg2exr is cross-platform, bundled with all variants of V-Ray and already has all the compression and datatype output options available for EXR output, so it seems like a very appropriate feature addition.
What do you think?
I would be happy to share my (quite simple) python code utilizing the OpenImageIO python binding we have here if it would somehow help.
Would it be difficult to add the possibility to merge multiple EXR images into one single EXR file with the vrimg2exr tool (and preferably via commandline)?
We "tile render" large still image renders (using the vray -region flag or Maya's -reg flag) on our render farm (Windows and Linux machines). As a post-render operation I use a bit of an obscure, precompiled and Windows-only OpenImageIO Python binding to read the buffers for each EXR image and then just combine them all (using "over", A+B, for all channels) before writing out the resulting buffer as a new, "merged" EXR. Since the canvas size of the fullsize image is recorded in each EXR tile, there's no need to specify coordinates for each tile. You just have to combine the buffers, so it seems like it should not have to involve too much development time.
The vrimg2exr is cross-platform, bundled with all variants of V-Ray and already has all the compression and datatype output options available for EXR output, so it seems like a very appropriate feature addition.
What do you think?
I would be happy to share my (quite simple) python code utilizing the OpenImageIO python binding we have here if it would somehow help.
Comment