during recent discussion mentioned sense & simplicity of acr calibration , while no 1 explicitly asked, thought should explain mean. in worst case wastes little bit forum space, , in best case save year or length of learning curve reach conclusion: acr calibration in fact makes sense , can quite simple.
let me start emphasizing obvious detail: calibrate tab sliders labels suggest; adjust red, green , blue hue & saturation main weighting on these primary colors , decreasing influence on secondary colors, respectively. while computed behind scenes on cie xy basis, there enough similarities hsb trace changes via hsb readings in photoshop (after raw conversion). example, red-hue-slider changes hsb-hue of red colors, though there second order effects on hsb-saturation & brightness.
anyway, calibrate tab not include brightness sliders red, green , blue allow compensate relative brightness differences. brightness of color - determined tonal controls. given can calibrate hue & saturation, suggest desist attempt linearize grayscale , eliminate nice tone curve of camera raw.
my eye-opener portrait of person wearing red shirt. image reasonably white-balanced based on previous shot of graycard, , tonal controls set according taste. always, skin tones magenta dragged red-hue-slider right (in camera raw), until noticed red shirt getting green. finally, acceptable compromise found; skin tones red shirt.
further, there bright green grass in background green dragged green-hue-slider left add touch of red.
----
this hue calibration can accomplished means of captured colorchecker chart:
/> camera raw set 2.x shipping defaults (ps cs); means 3.x without auto-corrections, , curve tab set linear (ps cs2). more details: shadows 5, brightness 50, contrast 25, saturation 0 (should valid cameras), sharpening 0, smoothing depending on camera needs , calibrate tab sliders @ 0. further, acr set prophoto rgb @ 16 bit precision.
/> whitebal second gray, patch #20 means of eyedropper. adjust exposure rgb= 190 patch.
/> make first raw conversion in order compare hsb-hue readings reference numbers follows. numbers derived averaging simulated charts offered colorremedies babelcolor:
red patch # 15: h= 10°, s= 62%
# 2 light skin: h= 30°, s= 28%
green patch #14: h= 101°, s= 45%
# 11 yellow green: h= 75°, s= 56%
blue patch # 13: h= 248°, s= 60%
# 3 blue sky: h= 230°, s= 29%
/> open raw file of captured chart again , adjust hue-sliders of calibrate tab in intuitive way eliminate numerical difference compared above target numbers. convert , compare again.
/> should done after 2 or 3 rounds , raw conversions. main challenge consider 2 patches simultaneously, e.g. #15 red , #2 light skin while adjusting red hue, thus, reach best possible compromise (if necessary).
----
now saturation? no problem @ add further rounds of raw conversions in order adjust saturation. far can tell, not distort hue again. question just, if calibration of saturation makes sense.
in case, colorchecker wants me increase green saturation. definitively wrong way go image described above
let me start emphasizing obvious detail: calibrate tab sliders labels suggest; adjust red, green , blue hue & saturation main weighting on these primary colors , decreasing influence on secondary colors, respectively. while computed behind scenes on cie xy basis, there enough similarities hsb trace changes via hsb readings in photoshop (after raw conversion). example, red-hue-slider changes hsb-hue of red colors, though there second order effects on hsb-saturation & brightness.
anyway, calibrate tab not include brightness sliders red, green , blue allow compensate relative brightness differences. brightness of color - determined tonal controls. given can calibrate hue & saturation, suggest desist attempt linearize grayscale , eliminate nice tone curve of camera raw.
my eye-opener portrait of person wearing red shirt. image reasonably white-balanced based on previous shot of graycard, , tonal controls set according taste. always, skin tones magenta dragged red-hue-slider right (in camera raw), until noticed red shirt getting green. finally, acceptable compromise found; skin tones red shirt.
further, there bright green grass in background green dragged green-hue-slider left add touch of red.
----
this hue calibration can accomplished means of captured colorchecker chart:
/> camera raw set 2.x shipping defaults (ps cs); means 3.x without auto-corrections, , curve tab set linear (ps cs2). more details: shadows 5, brightness 50, contrast 25, saturation 0 (should valid cameras), sharpening 0, smoothing depending on camera needs , calibrate tab sliders @ 0. further, acr set prophoto rgb @ 16 bit precision.
/> whitebal second gray, patch #20 means of eyedropper. adjust exposure rgb= 190 patch.
/> make first raw conversion in order compare hsb-hue readings reference numbers follows. numbers derived averaging simulated charts offered colorremedies babelcolor:
red patch # 15: h= 10°, s= 62%
# 2 light skin: h= 30°, s= 28%
green patch #14: h= 101°, s= 45%
# 11 yellow green: h= 75°, s= 56%
blue patch # 13: h= 248°, s= 60%
# 3 blue sky: h= 230°, s= 29%
/> open raw file of captured chart again , adjust hue-sliders of calibrate tab in intuitive way eliminate numerical difference compared above target numbers. convert , compare again.
/> should done after 2 or 3 rounds , raw conversions. main challenge consider 2 patches simultaneously, e.g. #15 red , #2 light skin while adjusting red hue, thus, reach best possible compromise (if necessary).
----
now saturation? no problem @ add further rounds of raw conversions in order adjust saturation. far can tell, not distort hue again. question just, if calibration of saturation makes sense.
in case, colorchecker wants me increase green saturation. definitively wrong way go image described above
oops ...
in case, colorchecker wants me increase green saturation. definitively wrong way go image described above. green grass in background quite glooming anyway; increase saturation have been contradictive pleasing rendition. and, theres explanation finding.
by getting color saturation right, de facto tonecurve set forth shadows, brightness , contrast silently turned hsb brightness curve. , here meet known downside of hsb color model: different hues numerically treated same. said de facto tone curve elegant brightening s-curve leads situation cyan-green-yellow hues gaining visual dominance , lab-lightness.
human eyes (at least mine) call hue-accuracy, idea of accurate saturation lets run difference between accurate vs pleasing rendition.
theres more 1 way deal situation. purposes found useful implement calibration accurate red-saturation , under-saturated green. starting point tweak saturation-sliders of calibrate tab depending on image-specific needs.
----
some final remarks:
i found procedure suggested above leads quite reproducible results comparing different captures of colorchecker target. shots done @ full spectrum lighting via direct sunlight should minimize metamerism. however, cannot comment on influence of in-door light sources , studio conditions. further, not find relevant influence of different exposures +/- ev.
whereas main tonal controls: shadows, brightness , contrast have significant , complex side effect on color saturation know rgb curves. interesting chapter math perspective, not without problems in practice. inevitably, settings shadows, brightness , contrast part of calibrated state. changing away it, more or less changes color saturation. 1 trick reduce such effects maintain ratio of brightness : contrast = 50 : 25 = 40 : 20 = 30 : 15, etc. if acceptable depending on image content.
like acr calibration procedure, final success depends on bayer r/g/b filters , spectral response. in ideal case colors snap place upon getting 3 primary patches #13-15 , 3 cie xy primaries right. real-world cameras better consider more color patches best possible average. above suggested memory colors: light skin, yellow-green , blue sky of course reflect personal preferences.
with best regards, peter
--
some references:
http://www.creativepro.com/printerfriendly/story/21351.html
http://home.comcast.net/%7eajpacheco/acrcalibration.zip
http://www.xs4all.nl/~tindeman/raw/color_reproduction.html
http://www.babelcolor.com/main_level/colorchecker.htm
http://www.colorremedies.com/realworldcolor/downloads.html
http://en.wikipedia.org/wiki/hsv_color_space
http://www.c-f-systems.com/photomathdocs.html
http://www.color.org/icc_white_paper_20_digital_photography_color_management_basics.pdf
in case, colorchecker wants me increase green saturation. definitively wrong way go image described above. green grass in background quite glooming anyway; increase saturation have been contradictive pleasing rendition. and, theres explanation finding.
by getting color saturation right, de facto tonecurve set forth shadows, brightness , contrast silently turned hsb brightness curve. , here meet known downside of hsb color model: different hues numerically treated same. said de facto tone curve elegant brightening s-curve leads situation cyan-green-yellow hues gaining visual dominance , lab-lightness.
human eyes (at least mine) call hue-accuracy, idea of accurate saturation lets run difference between accurate vs pleasing rendition.
theres more 1 way deal situation. purposes found useful implement calibration accurate red-saturation , under-saturated green. starting point tweak saturation-sliders of calibrate tab depending on image-specific needs.
----
some final remarks:
i found procedure suggested above leads quite reproducible results comparing different captures of colorchecker target. shots done @ full spectrum lighting via direct sunlight should minimize metamerism. however, cannot comment on influence of in-door light sources , studio conditions. further, not find relevant influence of different exposures +/- ev.
whereas main tonal controls: shadows, brightness , contrast have significant , complex side effect on color saturation know rgb curves. interesting chapter math perspective, not without problems in practice. inevitably, settings shadows, brightness , contrast part of calibrated state. changing away it, more or less changes color saturation. 1 trick reduce such effects maintain ratio of brightness : contrast = 50 : 25 = 40 : 20 = 30 : 15, etc. if acceptable depending on image content.
like acr calibration procedure, final success depends on bayer r/g/b filters , spectral response. in ideal case colors snap place upon getting 3 primary patches #13-15 , 3 cie xy primaries right. real-world cameras better consider more color patches best possible average. above suggested memory colors: light skin, yellow-green , blue sky of course reflect personal preferences.
with best regards, peter
--
some references:
http://www.creativepro.com/printerfriendly/story/21351.html
http://home.comcast.net/%7eajpacheco/acrcalibration.zip
http://www.xs4all.nl/~tindeman/raw/color_reproduction.html
http://www.babelcolor.com/main_level/colorchecker.htm
http://www.colorremedies.com/realworldcolor/downloads.html
http://en.wikipedia.org/wiki/hsv_color_space
http://www.c-f-systems.com/photomathdocs.html
http://www.color.org/icc_white_paper_20_digital_photography_color_management_basics.pdf
More discussions in Adobe Camera Raw
adobe
Comments
Post a Comment