# 090707: Different brighten commands:
#      ppmbrighten -v $(calc "${EXPCOMP}-100") -s 10 |
#      pamfunc -multiplier $(calc "${EXPCOMP}/100") |
# The original:
#      convert ppm:- -modulate ${EXPCOMP},110,100 ppm:- |



#####################################################################################################
# NOTE: ImageMagick configured with --enable-hdri --with-quantum-depth=32
# QUESTION: ImageMagick always compute with internal quantumdepth (Q32) irrespectible of -depth option?
# NOTE: 'ppmbrighten -v $EXPCOMP' introduce color distortion in magenta (at least before pnmgamma?)!
# NOTE: One theory says it's better to apply sharpening in linear mode, before gamma curve, unlike here!
# NOTE: ImageMagick defaults to use 16-bit depth.
# NOTE: bt709ramp (lower gamma so use 'convert' GAMMA=1.0) best with some color-profiles (CaptureOne), 
#   for others (Canon) srgbramp (higher gamma so use 'convert' GAMMA=0.82) is best.
# NOTE: Specifying other gamma in pnmgamma can create color artifacts.
# NOTE: Saturation boost via '-modulate' before color profiling can give banding effects.
# NOTE: Saturation boost via '-modulate' at any place with CaptureOne profile give banding effects.
# NOTE: tifficc use Perceptual intent (-t 0) as default. Black point compensation (-b) is useful only 
#         for Relative Colorimetric intent (-t 1)
# WARNING: pnmnorm sometimes perform bad (banding) if it is applied in linear mode (before pnmgamma).
#          Works better with imagemagicks contrast-stretch or by placing pnmnorm after pnmgamma!

# WARNING: Color profiling with 'convert' performs shit in grayscale images - don't use it! 
#          Use 'tifficc' instead. Check e.g with D070101_a1980.cr2...
# This is not true! Update: 'convert -type GrayScale' as it converts to 8-bit and makes color artifacts!

# NOTE: Have tested color profiling with 'convert' instead of 'tifficc' and -wpercent 0.5 to make extra 
#       stretch of image. Color artifacts can appear with profiling performed after extensive stretching.

# NOTE: (080324) Have tested new dcraw option -H 2 with gradual
# clipping. Works ok, but darkens image a bit, and if brighten the
# result is not really any better than default -H 0.

# Alternative noise reductions:
# greycstoration -visu 0 -restore imagein.ppm -o imageout.ppm
# Noise Ninja is best - amazing! Unfortunatly no command line interface...
#####################################################################################################


#####################################################################################################
# TODO:
#
# 081104: Add so that UNSHARP works on TIFF as well if option '-sharp 1'  is set.
#
# 070515: Fix control of whitebalance and tone-curves. 
#         Have tested "white wedding" tonecurve: -sigmoidal-contrast 4,20% +sigmoidal-contrast 3,35%
#         See gnuplot script 'sigmoidal_contrast_test.gnuplot' and D070501_a3655...
#         Decrease ${PNMNORM2_OPT} to 0.005...?
# 
#         Test also -sigmoidal-contrast 6,25% -sigmoidal-contrast 4,55% +sigmoidal-contrast 3,35%
#         to get more linear curve in shades. 
#         Try also -sigmoidal-contrast 5,20% -sigmoidal-contrast 3,55% +sigmoidal-contrast 3.5,35%
#
# 070513: Fix so that a low mean pixel value will increase -wpercent (PNMNORM2_OPT) proportionally, 
#         as a means of forcing a histogram stretch despite strong but small regions of white in image, 
#         like a lamp or sun, to make dull images brighter.
#
# 070512: Fix so that -red -green -blue can set any channel combination for gray conversion.
# 
# 070510: Implement lens vignetting, radial distortion and chromatic abberation correction.
#         Needs EXIF info in file...
#         fulla -p -r 0:0:0.0007:1.0002 -b 0:0:0.0005:1.0001 \
#               -c 1:0.20:-0.51:0.2 -t 2 -l "Canon EF 28-135mm f/3.5-5.6 IS USM" \
#               -o ${TIF_FILE}.tmp1 ${TIF_FILE}.tmp2

# 070509: Ideally I would like to have color profile applied (tifficc) right after pnmgamma but 
#         before pnmnorm?!
#####################################################################################################