Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: Saturation Clipping – how would you edit such image to sRGB ?  (Read 10047 times)
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« on: December 30, 2011, 03:27:07 AM »
ReplyReply

 
Image # 1:  base image in ACR without any adjustments except an Exposure correction.
Image # 2:  same as # 1, output space is set from ProPhoto RGB to sRGB.
Image # 3:  basic tonal adjustments are added. Output space is reset to ProPhoto RGB like with # 1.
Image # 4:  same as # 3, output space is set from ProPhoto RGB to sRGB.

It is obvious that the majority of out-of-sRGB colors were not in the raw data,
but were created upon basic tonal editing in "1.0 pRGB" in ACR.

Comments ?

Peter

--
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #1 on: December 30, 2011, 04:18:29 AM »
ReplyReply

Comments ?

You sure this image is worth arguing abut?

Just process the image in PPRGB and print–unless you have some other agenda scheduled.
Logged
Rainer Ots
Newbie
*
Offline Offline

Posts: 7


« Reply #2 on: December 30, 2011, 06:50:41 AM »
ReplyReply

Quote
how would you edit such image to sRGB ?
By not doing edits that cause the colors to go out of sRGB gamut  Roll Eyes

But seriously  - what exactly are you trying to achieve?
Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1231



WWW
« Reply #3 on: December 30, 2011, 12:32:40 PM »
ReplyReply

View the B&W preview of the blue channel detail in the flower petal in Photoshop and you'll note there's not a lot to be concerned about. As long as you see definition in the microfine veins and wrinkles in the flower you've got all the data you need to print from.

I'ld suggest you see what happens to the preview and gamut warning when you shift color temp toward blue from its current warmish 5600K and -4 tint toward magenta. The overall image looks a bit yellow/green, but I doubt it's going to make a dent in preventing clipping converting to sRGB.
« Last Edit: December 30, 2011, 12:36:56 PM by tlooknbill » Logged
FranciscoDisilvestro
Sr. Member
****
Offline Offline

Posts: 558


WWW
« Reply #4 on: December 30, 2011, 12:47:53 PM »
ReplyReply

The way that ACR handles output color spaces is similar to softproofing, so if your purpose is to edit an image to sRGB, just set that color space as in you images #2 & #4.

Now, about editing, I would think about two choices:

1.- Edit directly with color space configured to sRGB, so you will know if you are clipping a channel.

2.- If you like to edit in ProPhotoRGB to create a "Master file" and need an sRGB output, then edit in PPRGB, save your edits and then change color space at the end to sRGB (as in #4). If you have clipped channels, a basic approach is to reduce vibrance, which affects saturated colors first.

About your comment that "the majority of out-of-sRGB colors were not in the raw data", well, that is one thing normal in editing, just adding saturation could make colors go outside a small color space like sRGB.
Logged

Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1325



WWW
« Reply #5 on: December 30, 2011, 12:51:24 PM »
ReplyReply

You can develop in ProPhoto RGB, go to Photoshop, and do softproofing for sRGB there. Then process (desaturate) the conflictive areas using a mask layer so that the softproof clipping warning disappears.

This is an example:

(left: image after conversion ProPhoto RGB -> sRGB, centre: gamut clipping warning, right: histogram of resulting image after conversion)


Mask layer used for desaturation:



Now the funny question: if you ask 100 people, how many will prefer the mega strong yellow school bus with the ugly clipped histogram, and how many will pick the perfect histogram image with soft gradations? and which one will look better in print?

Regards
« Last Edit: December 30, 2011, 12:57:44 PM by Guillermo Luijk » Logged

Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1231



WWW
« Reply #6 on: December 30, 2011, 12:56:36 PM »
ReplyReply

I'ld like to share a little anecdote that offers some perspective on equating data captured and lost according to what the math says instead of the preview.

The last couple of weeks I've been restoring 100 year old yellowed medium format prints (maybe from a Brownie camera) one of which has a woman wearing a blouse clearly blown out by the sun at the time of capture and processing. Scanning it in Epson's Raw setting (no color adjusts), editing in ACR by converting to Grayscale in the HSL panel, I dialed back the yellow channel to reveal a polka dot pattern that wasn't even seen in the print or in the Epson 4870 scanner preview no matter how I edited in that lousy piece of software.

The family I'm doing this project for is going to learn something new about their mother's sense of fashion they may not have known before.

So I really doubt a modern digital camera is going to lose a lot of data by comparison.
Logged
FranciscoDisilvestro
Sr. Member
****
Offline Offline

Posts: 558


WWW
« Reply #7 on: December 30, 2011, 01:15:33 PM »
ReplyReply

Now the funny question: if you ask 100 people, how many will prefer the mega strong yellow school bus with the ugly clipped histogram, and how many will pick the perfect histogram image with soft gradations? and which one will look better in print?

Regards

Maybe over 90 people will prefer the "mega strong yellow" version. Now, which one is closer to the real thing? I mean, if you print both versions and hold them close to the real school bus, which will look perceptually closer?

In this case, it seems to me that the scene color is outside the sRGB gamut, so it will be natural to prefer the blown out version, also because there are not really interesting details lost in the saturated areas.
Logged

Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« Reply #8 on: December 31, 2011, 10:53:58 AM »
ReplyReply

Quote
You can develop in ProPhoto RGB, go to Photoshop, and do softproofing for sRGB there. Then process (desaturate) the conflictive areas using a mask layer so that the softproof clipping warning disappears.
Quote
About your comment that "the majority of out-of-sRGB colors were not in the raw data", well, that is one thing normal in editing, just adding saturation could make colors go outside a small color space like sRGB.

Well yes, it may be normal with a large-gamut working space such as hardwired in ACR.
Even some basic tonal adjustments can move colors out-of-tiny-sRGB, thus, leading to channel clipping and increasing the risk for degradation of image details. Then, there are certainly several means to tweak saturation of conflictive areas, to "perceptualize" conversion to sRGB, however, in the context here it can be seen as kind of "treating a problem which was homemade before":

/>  Left image:  processed with Brightness 50 & Contrast 25 in ACR to ProPhoto RGB, then converted to sRGB.
/>  Right image:  processed with Brightness zero & Contrast zero in ACR to ProPhoto RGB, converted to sRGB, then in Photoshop an S-shaped tone curve was applied (corresponding to the effect of Brightness 50 & Contrast 25 as far as tonality is concerned).

With the left image the Red channel is heavily blown.
Degradation of image details is subtle but visible upon close examination.

Peter

--
Logged
jrsforums
Sr. Member
****
Offline Offline

Posts: 750


« Reply #9 on: December 31, 2011, 12:53:17 PM »
ReplyReply

How did you build the mask for this?

Thanks....John

You can develop in ProPhoto RGB, go to Photoshop, and do softproofing for sRGB there. Then process (desaturate) the conflictive areas using a mask layer so that the softproof clipping warning disappears.

This is an example:

(left: image after conversion ProPhoto RGB -> sRGB, centre: gamut clipping warning, right: histogram of resulting image after conversion)


Mask layer used for desaturation:



Now the funny question: if you ask 100 people, how many will prefer the mega strong yellow school bus with the ugly clipped histogram, and how many will pick the perfect histogram image with soft gradations? and which one will look better in print?

Regards

« Last Edit: December 31, 2011, 03:23:07 PM by jrsforums » Logged

John
digitaldog
Sr. Member
****
Offline Offline

Posts: 9225



WWW
« Reply #10 on: December 31, 2011, 02:48:26 PM »
ReplyReply

Now the funny question: if you ask 100 people, how many will prefer the mega strong yellow school bus with the ugly clipped histogram, and how many will pick the perfect histogram image with soft gradations? and which one will look better in print?

Exactly, and the output to print will undergo yet another, different gamut clipping exercise. Point is, going from a larger to smaller gamut space is a fact of life. We’re so caught up in what happens going from ProPhoto to sRGB, imagine having to output images to newspaper press for a wake up call. It is what it is. Guillermo has provided the best answer in terms of seeing what is going to happen and if so desired, trying to aid the process as best as can be expected but colors are going to clip! If you don’t introduce color shifts (which shouldn’t really occur with good profiles), you live with the output you end up with and move on.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1325



WWW
« Reply #11 on: December 31, 2011, 10:21:57 PM »
ReplyReply

How did you build the mask for this?
Selection -> Colour gamut... -> Select: Out of gamut
(sorry if the translation is not precise since I don't have an English PS).

This option seems to provide an out of gamut selection for the CMYK work profile, which in this case matched quite well the sRGB gamut clipping (if you look at the sky there are out of gamut traces that were actually not a problem for sRGB). Too bad PS doesn't allow to build an out of gamut mask for any output profile, or at least I don't know how to do it without performing semi-surrealistic maneuvers.
« Last Edit: December 31, 2011, 10:30:59 PM by Guillermo Luijk » Logged

BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3874


« Reply #12 on: December 31, 2011, 10:47:40 PM »
ReplyReply

Too bad PS doesn't allow to build an out of gamut mask for any output profile, or at least I don't know how to do it without performing semi-surrealistic maneuvers.

+1

Cheers,
Bart
Logged
AlanPezzulich
Newbie
*
Offline Offline

Posts: 19


« Reply #13 on: January 31, 2012, 07:20:53 AM »
ReplyReply

When you use 'convert to profile' photoshop will bring all colors into gamut. 'Rendering intent' will allow the choice of how Photoshop handles it.

Perceptive - shift all colors to maintain relationship of colors - large color shifts - looks natural
Relatice Colormetric - Shifts in gamut colors less than perceptive
Absolute Colormetric - Clips out of gamut colors - does not shift in gamut colors.

No need to selectively change out of gamut colors. Choose rendering intent that looks best.

Alan
Logged
jbrembat
Full Member
***
Offline Offline

Posts: 177


« Reply #14 on: February 01, 2012, 07:22:52 AM »
ReplyReply

Quote
When you use 'convert to profile' photoshop will bring all colors into gamut. 'Rendering intent' will allow the choice of how Photoshop handles it.

Perceptive - shift all colors to maintain relationship of colors - large color shifts - looks natural
Relatice Colormetric - Shifts in gamut colors less than perceptive
Absolute Colormetric - Clips out of gamut colors - does not shift in gamut colors.

No need to selectively change out of gamut colors. Choose rendering intent that looks best.
1 - It's not a specific PhotoShop feature. Any colour-managed application can do.
2 - Perceptual: try to preserve color relathionship
3 - Relative: clips out of gamut colors to destination gamut boundary
4 - Absolut: same as relative without  aligning the neutral axes
5 - Perceptual doesn't exist for matrix profiles. So if you go from ProPhoto to sRGB asking for perceptual you get relative.

Jacopo
« Last Edit: February 01, 2012, 07:45:23 AM by jbrembat » Logged
rambler44
Full Member
***
Offline Offline

Posts: 103


« Reply #15 on: February 19, 2012, 09:47:56 AM »
ReplyReply

I just have to jump in here and say that I believe this is a terrific thread.  Although the question and content that follows is beyond my beginning skills in Photoshop, I find that Peter's questions are an example of how a forum is best served.  He has outlined specific questions with specific examples of what his issues are.  The majority of responses have avoided being judgmental or opinionated about the importance of the questions and have tried specifically to answer what poster asked.  I have been impressed by the thoughtfulness of the responses which also reveal the skills and expertise of the members who participated. Keep up the good work. 
Logged
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad