Specify 16 bit int per channel on save

[jules43]

When saving the results of an image magic operation to a tiff or png, is there a way to specify 16 bit int per channel output?

Thanks

Jules

[fmw42]

You must have compiled your IM version as Q16 or Q32. Then just add -depth 16 before writing the output. But if you input image was 8-bits, you will still have 8-bits at the end but spreach of 16 -bit range. Perhaps you can explain what processing you are doing. If it resamples the image, then you would get more values.

[jules43]

Thanks, I hadn't been certain that IM would preserve the bit depth on save. I should have provided more detail.

I am reading a 16 bit per pixel loss-less PSD and after processing (if any) I need to save it as as 16 bits per pixel loss-less TIFF or PNG.

I am using the Q16 version of ImageMagick 7 - I may look at moving to Q32 bit if I hit any precision problems.

Having seen the -depth option it looks like I don't need to specify it as it will be preserved by default.

Reading the docs on -depth then to led me to the -define options for controlling specific image format and the -quality, -type and -compress options. The combination of all of these looks like it gives me all the control I need.

I also found the section on common formats which clarified a few remaining questions I had.

Thanks again for pointing me in the right direction.

[fmw42]

IM 7 in Q16 should properly preserver 16-bit images. Yes you will need the proper -defines and other optios for saving to TIFF and to PNG

Results will be non-negative 16-bit integers. If you need negatives and fractional values, then you need Q16 HDRI. But I don't think from what you said you need that accuracy. Q32 would also be integers only without HDRI.