Legacy ImageMagick Discussions Archive

The MIFF documentation does not say how grayscale images are formatted, but IM can successfully write and read them. By experimentation, I see that, for various color types, the samples in a scanline have the following arrangements:
(This would be for an image 8 pixels wide.)

That is, for grayscale images, each scanline is padded with "0" samples to make it use as much space as it would if it were a color image.

Padding with 0s seems odd to me, so I'm wondering if this is by design, or what.

To test this, I'm using commands similar to the following, with IM 6.7.2-1.

You found a bug. We'll get a patch into ImageMagick 6.7.2-2 Beta within a day or two. Thanks.

Note that internally a gray image is actually in memory as a RGB image, because of this any RGB image that only contains gray colors will be converted to a gray MIFF image on save.

Also note that there was a recent change such that if the image being saved as MIFF is marked as "sRGB" (a IMv6 mark saying the image was sRGB and is currently converted to linear-RGB) then the image will remain sRGB even if the image only contains gray colors. That is because to convert it to gray will cause the image to loose its sRGB significance in the intermediate image causing problems further down the pipeline. IM does not have a "sGRAY" colorspace.

Aside, IMv7 will see "sRGB" to really mean a "sRGB" image, and not "sRGB converted to linear RGB"