Attached is a file that loads a radiance file that should be a smooth gradient. The toe file shows that the file is being quantized to 8bit in the movie in node. I have tried 32 bit tiff files as well, but those do not read correctly. Are there any other formats that might work? This one is a bit of a showstopper.
edit - I made both 16bit int and 32 bit float tif files, they both read correctly, but touch reports the format to be 8 bit. Radiance is the only format that shows up as a 32 bit float, but is is still truncated.
MD 32bit_moviein_issue.rar (16.1 KB)
I traced through the code and everything looks fine.
I then opened the file in HDRView [url]http://ict.debevec.org/~debevec/FiatLux/hdrview/[/url], and if you zoom in on the image you can see that the image itself has been quantized already (zoom in and drag your mouse sideways and you see that you get multiple pixels of the same color). So the issue here is just the .hdr image has already been quantized.
ok, I have narrowed it down. I cannot seem to write a .hdr file that is not quantized. I have a valid 32 bit tiff file, houdini’s mplay show’s the values changing every pixel. When I save to HDR format from mplay, after effects, photoshop, and directly from vray or mental ray, the hdr format ends up quantized. I am looking for some other tools now, perhaps hdrshop
MD
If I understand this article correctly, the radiance HDR format is not really equivalent to a 32 bit tiff file, in that is stores the data as rgbe, 8bit per component + a shared exponent, so it is really only for handling extreme darks and lights, rather than a 32 bit per component representation. So, if I am correct, I’ll make this an RFE for support of 32 bit tiff or EXR formats (EXR would be the best in my opinion)