not sure if this qualifies as a bug,
however it is not the behaviour I would expect:
When reducing the resolution on a PointfileinTOP or Pointfileselect TOP
the resolution is e.g. halfed, but also the image is scaled in the resolution twice.
Same is true when using the global res multiplier.
While this may make sense with actual pointclouds (didnt test), it doesn’t make sense when using
the pointfileselect to extract different render passes from a multi-channel exr file.
Its normal, there is no relation between the spatial position of pixel in the Top and the spatial position of point in space. On some pointcloud, as with kinect, there is a relation between the top distribution and the space distribution but only in z. I made some algoritms with compute shader to reorder the pointcloud top but its like a projection in one axis.
Question is if it has to be like that, as pointfileselect is to my knowledge the only way to extract channels from exr, which is actually pretty cool and maybe the resolution could be reduced the same way it is reduced in other tops to make that workflow more usable…
Thanks for the comment. As was stated, the Point File In/Select workflow was designed for point data where the position inside the image doesn’t generally matter. However, you make a good point that it is currently the only way to really access additional channels in an EXR file.
While it would be possible to add options to the Point File In/Select to change how it resizes images, I think the best method might be to add better multi-channel support to the standard Movie File In workflow to match the EXR features we added to the Movie File Out.
I’ll make an interal todo item for this, but in the short term I guess you could add a resolution top afterwards to resize it in the traditional way.
