-depth "max" please

[snibgo]

I work with Q16. I sometimes need to include "-depth 16" in a script to ensure I don't lose precision. I'd like my scripts to be portable, certainly in Q32 and possibly in Q8. So I don't want to specify a particular number of bits, but "give me all the bits you can".

I can do this in a script, of course, calling IM first to enquire the Q-number, then use that for the real work. But this is slow and awkward. I'd prefer a "-depthMax" option, or similar.

Thanks.

[glennrp]

I work with Q16. I sometimes need to include "-depth 16" in a script to ensure I don't lose precision. I'd like my scripts to be portable, certainly in Q32 and possibly in Q8. So I don't want to specify a particular number of bits, but "give me all the bits you can".

I believe "+depth" is the same as your proposed "-depthMax". See code in wand/mogrify.c beginning around line 1284.
This feature isn't mentioned in the commandline documentation, but that's easily fixed.

Oops, looks like that won't always be true. wand/convert.c, wand/montage.c, and others just ignore "+depth", as far as I can tell.

Glenn

[snibgo]

Ha! Yes, that seems to work. Thanks.

[glennrp]

Doesn't work the way I expected.

• glenn.rp> convert rose: +depth -compress none rose+dmax.ppm
  glenn.rp> ls -l rose+d.ppm
  -rw-rw-r-- 1 glennrp glennrp 58670 Jan 21 15:09 rose+dmax.ppm
  glenn.rp> head -4 rose+dmax.ppm
  P3
  70 46
  4294967295
  12336 12079 11565 12850 12336 11822 13878 12850 12079 14392 13107 11822

Notice the 4.29G depth value in the PPM file, that should have been 16k.
The pixel values do seem to have the correct range, though. Mogrify didn't work either:

• glenn.rp> convert rose: rose.png
  glenn.rp> mogrify -format ppm +depth -compress none rose.png
  glenn.rp> head -4 rose.ppm
  P3
  70 46
  4294967295
  12336 12079 11565 12850 12336 11822 13878 12850 12079 14392 13107 11822
  glenn.rp> convert | head -1
  Version: ImageMagick 6.8.8-2 Q16 x86_64 2014-01-12 http://www.imagemagick.org