I have a heightfield rendered from Houdini, stored as 32 bit floats in range -1 to 1, saved as a single channel EXR. Other applications read this exr file as 32-bit single channel.
Using the command line tools, I cannot get IM to read the Alpha channel as anything but 16 bit. Even with -depth 32, it will identify in verbose as "Depth: 32/16" with Alpha listed as 16-bit and Red/Green/Blue as 1-bit. I have also tried -define exr:color-type=A, also G, R, Y to no effect.
Version: ImageMagick 7.0.8-25 Q16 [HDRI] x64 2019-01-28 (dll)
I can perform crop on the image with exr:color-type=A, but the result being 16bit quantizes the heightfield. I have also tried -define quantum:format=floating-point with no luck, seems irrelevant if IM reads the original source image channel as 16bit.
ImageMagick only reads 32b EXR alpha as 16 bit?
- fmw42
- Posts: 25562
- Joined: 2007-07-02T17:14:51-07:00
- Authentication code: 1152
- Location: Sunnyvale, California, USA
Re: ImageMagick only reads 32b EXR alpha as 16 bit?
I believe that Imagemagick only handles currently 16-bit EXR files. I am not sure if it handles transparency or not.
Re: ImageMagick only reads 32b EXR alpha as 16 bit?
I may have misread the documentation says the HDRI version should read/write 32b and negative values. I could be wrong, my tests seem to identify RGB as 32b but never A, leads me to believe that RGB may support 32b whereas alpha was neglected. ![Sad :(](./images/smilies/icon_sad.gif)
If there is no support for 32 bit single-channel, is a fix on the horizon any time soon?
![Sad :(](./images/smilies/icon_sad.gif)
If there is no support for 32 bit single-channel, is a fix on the horizon any time soon?