Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Editing in AdobeRGB, converting to sRGB  (Read 10379 times)
Muizen
Newbie
*
Offline Offline

Posts: 15


« on: June 29, 2006, 01:28:17 PM »
ReplyReply

When editing in Raw and in PS it is recommended to bring in as much data as possible by a.o. using a wide color space, usually AdobeRGB or ProPhotoRGB and start with 16-bit depth.
A large color space also reduces possible clippings as shown in the RAW histogram.

I wonder what happens thereafter to the image quality obtained when for printing purposes, the image has to be converted to sRGB and 8 bit?

Does one lose the advantages of editing in large color space 16 bit after converting to sRGB?

Thanks for reply!

Harry Briels
Mechelen, Belgium
Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #1 on: June 29, 2006, 06:51:21 PM »
ReplyReply

Quote
When editing in Raw and in PS it is recommended to bring in as much data as possible by a.o. using a wide color space, usually AdobeRGB or ProPhotoRGB and start with 16-bit depth.
No, not really. A large gamut per se helps nothing. Mor important is, how the colors are treated, are out-of-gamut (OOG) colors clipped or mapped into the working gamut. And if you do not want to do any post processing, you don't need the 16bit color depth. If you can do all corrections within your Raw conversion software, it would be ideal to convert it directly to the output profile. That's unfortunately not possible with Photoshop, so there you have to convert the image into an intermediate color space first. Due to the architecture of the RGB profiles in PS, there is no use in ProPhotoRGB as an intermediate color space, if you want to convert it to another RGB space afterwords. It only makes sense, if you convert directly from ProPhotoRGB to you printer profile and this profile (and your image!) contains a considerable amount of colors outside AdobeRGB. If this is not the case, you loose nothing converting directly to AdobeRGB. If you have to do some major postprocessing, you'll might want to go 16bit to avoid posterization. If you do only the conversion from AdobeRGB to your output profile, I am not sure if 16 bit has any significant advantages. Using ProPhotoRGB makes 16bit a must.

Quote
A large color space also reduces possible clippings as shown in the RAW histogram.
No, not really. Only, if you do not correct the exposure sliders. If your histogram is perfect, when ProPhotoRGB is the output profile, there might be some clipping switching to AdobeRGB. You have to counterbalance this with the exposure settings.

Quote
I wonder what happens thereafter to the image quality obtained when for printing purposes, the image has to be converted to sRGB and 8 bit?
That's, what you have to care for, and that's where large gamuts are really difficult to handle. A perceptive conversion from ProPhotoRGB to sRGB is not possible, so in order to prevent clipping of OOG colors, you have to carefully desaturate the colors using the soft proof function. The final conversion from 16bit to 8bit is not a problem, since you need the 16bit only as a headroom for corrections. When those are done, there's no need for 16bit anymore. As long as your output device works in 8bit, as well. There seem to be printers with 12bit output, regarding Michaels report on the new Canon printer. This of course, would be a definite reason to keep the image in 16bit.

Quote
Does one lose the advantages of editing in large color space 16 bit after converting to sRGB?
As long as your output profile is considerably smaller than your "large color space", there's only very little advantage, but big problems. There is absolutely no use in a large color space, if there are no colors using it. If the colors of your image are within AdobeRGB, a gamut like ProPhotoRGB is a waste.

With my limited knowlidge and experience, I find it more usefull, to compress the colors into AdobeRGB, than using ProPhoto and struggeling with clipped colors later. Since we in Germany have a pretty well organized CM with well supported ISO CMYK profiles (not the crap PS profile named "ISO Coated Fogra 27" but the real ISO stuff), I find it a bit annoying not to be able to choose an ECI-RGB or any printer profile directly in ACR - that would solve the whole large gamut 16bit discussion.
Logged

Best Regards

Dennis.
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #2 on: July 08, 2006, 06:48:17 AM »
ReplyReply

Quote
No, not really. A large gamut per se helps nothing. Mor important is, how the colors are treated, are out-of-gamut (OOG) colors clipped or mapped into the working gamut. And if you do not want to do any post processing, you don't need the 16bit color depth. If you can do all corrections within your Raw conversion software, it would be ideal to convert it directly to the output profile. That's unfortunately not possible with Photoshop, so there you have to convert the image into an intermediate color space first. Due to the architecture of the RGB profiles in PS, there is no use in ProPhotoRGB as an intermediate color space, if you want to convert it to another RGB space afterwords. It only makes sense, if you convert directly from ProPhotoRGB to you printer profile and this profile (and your image!) contains a considerable amount of colors outside AdobeRGB. If this is not the case, you loose nothing converting directly to AdobeRGB. If you have to do some major postprocessing, you'll might want to go 16bit to avoid posterization. If you do only the conversion from AdobeRGB to your output profile, I am not sure if 16 bit has any significant advantages. Using ProPhotoRGB makes 16bit a must.

I strongly disagree with most of this. There are several advantages to using ProPhoto as your editing space. First, when editing an image, there is no advantage to converting from RAW directly to a specific output device profile, and many disadvantages. Most photographers print their images on more than one printer, whether personally owned, the minilab around the corner, or an online service. You want to have at least one master version of your image that is not compromised in any way by output device gamut limitations, and then create a device-specific version for a particular press or printer only when necessary. Converting RAW to ProPhoto first, then editing for a specific device allows much greater control over how the gamut compression/reduction is done. Why cripple your master image to accommodate the current limits of rapidly-advancing printer technology?

Quote
If your histogram is perfect, when ProPhotoRGB is the output profile, there might be some clipping switching to AdobeRGB. You have to counterbalance this with the exposure settings.

This is really bad advice. The exposure slider is at best a very crude way to control color gamut. It only helps control gamut in the highlights, and has no effect on saturated colors in shadows, like green leaves or grass in the shadows of a landscape image. These colors are often more prone to fall out-of-gamut than bright colors, unless you're shooting autumn leaves or car shows or flowers or suchlike. What you're advocating will result in unnecessarily darkening the image when saturated colors are present, which means more and unnecessary adjustments later in Photoshop.

Quote
That's, what you have to care for, and that's where large gamuts are really difficult to handle. A perceptive conversion from ProPhotoRGB to sRGB is not possible, so in order to prevent clipping of OOG colors, you have to carefully desaturate the colors using the soft proof function.

That's how you should always adjust gamut to fit a specific output device; adjusting individual color ranges with the hue/saturation control, preferably with a saturation and/or luminance mask to focus the effect on the color ranges that need to be shoehorned without affecting the rest of the image any more than absolutely necessary. Wide-gamut editing spaces are only "difficult" if you don't know how to use them effectively.

Quote
The final conversion from 16bit to 8bit is not a problem, since you need the 16bit only as a headroom for corrections. When those are done, there's no need for 16bit anymore. As long as your output device works in 8bit, as well. There seem to be printers with 12bit output, regarding Michaels report on the new Canon printer. This of course, would be a definite reason to keep the image in 16bit.

Keep your image 16-bit anyway, unless sending it to a client that can't handle 16-bit images. Even if you are printing to an 8-bit device, you'll still get better results printing a 16-bit file. When converting from 16-bit ProPhoto RGB to 8-bit printer data during the printing process, all 256 levels of the color channel data going to the printer can be used; the color profile conversion is perfomed first, then the data is rounded to the nearest 8-bit value and sent to the printer. But if the source file is 8-bit, then you may be only using 100-200 of the possible 8-bit values in a color channel (think of the "toothcomb" histogram you get when doing a curve adjustment in 8-bit mode in Photoshop) and you'll have more posterization and banding in your print as a result.

Quote
As long as your output profile is considerably smaller than your "large color space", there's only very little advantage, but big problems. There is absolutely no use in a large color space, if there are no colors using it. If the colors of your image are within AdobeRGB, a gamut like ProPhotoRGB is a waste.

This is really irrelevant when working with 16-bit files. ProPhoto files are no larger than Adobe RGB, and neither are susceptible to banding/posterization in 16-bit mode.

Quote
With my limited knowlidge and experience, I find it more usefull, to compress the colors into AdobeRGB, than using ProPhoto and struggeling with clipped colors later. Since we in Germany have a pretty well organized CM with well supported ISO CMYK profiles (not the crap PS profile named "ISO Coated Fogra 27" but the real ISO stuff), I find it a bit annoying not to be able to choose an ECI-RGB or any printer profile directly in ACR - that would solve the whole large gamut 16bit discussion.

This advice is based on the limitatations of your knowledge and experience, which is why you're "struggeling with clipped colors later". Using RAW converter exposure settings and output color space choice to fix out-of-gamut colors is crude and imprecise, and leaves one with problems that must be solved later. Also, be aware that given the differences in shape between printer color spaces and editing color spaces, you need ProPhoto to use 100% of the printer gamut anyway. Many printers can print some colors that fall outside of Adobe RGB, so you're possibly cheating yourself out of some printer gamut if you use anything less than ProPhoto anyhow.

And if you start using a different printer with a wider gamut, then you have to completely reprocess your image from RAW conversion on to take advantage of it. If you have a master copy of the image in ProPhoto RGB, you only need to perform a hue/saturation adjustment on the master, convert to output profile, and save as a new copy (and even this will be unnecessary on many cases) and you will not have to redo noise reduction, sharpening, curves, compositing, cloning, and other adjustments and edits to the image. This is particularly useful if the image is a stitched panorama, an image shot with mixed lighting requiring multiple RAW conversions with different white balances to be blended together, a multi-shot high-dynamic-range blend, or has some other time-consuming processing that must be done.

Get in the habit of editing a 16-bit ProPhoto master image that is not compromised by output device limitations, and making output-device-specific variations of the master when necessary. It will save you time and hassle in the long run as printer technology improves.
Logged

bjanes
Sr. Member
****
Offline Offline

Posts: 2822



« Reply #3 on: July 08, 2006, 07:52:50 AM »
ReplyReply

Quote
That's how you should always adjust gamut to fit a specific output device; adjusting individual color ranges with the hue/saturation control, preferably with a saturation and/or luminance mask to focus the effect on the color ranges that need to be shoehorned without affecting the rest of the image any more than absolutely necessary. Wide-gamut editing spaces are only "difficult" if you don't know how to use them effectively.

Get in the habit of editing a 16-bit ProPhoto master image that is not compromised by output device limitations, and making output-device-specific variations of the master when necessary. It will save you time and hassle in the long run as printer technology improves.
[a href=\"index.php?act=findpost&pid=70053\"][{POST_SNAPBACK}][/a]

Jonathin makes excellent points, but if the gamut of the capture fits into aRGB, I see no advantage in ProPhotoRGB and some disadvantages. If you are mapping to an output device using the controls Jonathin mentions, there is no problem. However, if you are taking the easy way and using perceptual rendering intent, there may be excessive compression of the gamut. The rendering engine (as currently implemented) does not examine what colors are in the file, but merely compresses the larger space by a predetermined amount.
Logged
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #4 on: July 08, 2006, 08:05:15 AM »
ReplyReply

Quote
Jonathin makes excellent points, but if the gamut of the capture fits into aRGB, I see no advantage in ProPhotoRGB and some disadvantages. If you are mapping to an output device using the controls Jonathin mentions, there is no problem. However, if you are taking the easy way and using perceptual rendering intent, there may be excessive compression of the gamut. The rendering engine (as currently implemented) does not examine what colors are in the file, but merely compresses the larger space by a predetermined amount.

Which is why I hardly ever use perceptual rendering, and use relative colorimetric instead. I've found that printing the ProPhoto master file using relative colorimetric rendering to the output profile without hand-converting delivers excellent results more than 95% of the time. Perceptual rendering often introduces color casts as well as the issue you mention. Large-gamut color spaces have no disadvantages if you use them properly and avoid sloppy processing techniques which can get you into trouble even if you use sRGB.
« Last Edit: July 08, 2006, 08:07:57 AM by Jonathan Wienke » Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 9186



WWW
« Reply #5 on: July 08, 2006, 09:01:22 AM »
ReplyReply

Do not pass go, do not collect $200. Go directly to Lightroom poscast #8 and listen to the geeks discuss some of these issues:

http://photoshopnews.com/2006/07/07/lightr...pisode-8-posted

The first 10 minutes will do of course, enjoy the rest if you want.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #6 on: July 08, 2006, 06:22:53 PM »
ReplyReply

Quote
I strongly disagree with most of this.
And I do with most of your comments.
« Last Edit: July 08, 2006, 06:38:00 PM by Dennis » Logged

Best Regards

Dennis.
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #7 on: July 08, 2006, 06:23:29 PM »
ReplyReply

Quote
Get in the habit of editing a 16-bit ProPhoto master image that is not compromised by output device limitations, and making output-device-specific variations of the master when necessary. It will save you time and hassle in the long run as printer technology improves.
Get in the habit of doing well exposed photographs, that do only need some minor corrections, which can be done inside the raw conversion software
« Last Edit: July 08, 2006, 06:37:14 PM by Dennis » Logged

Best Regards

Dennis.
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #8 on: July 09, 2006, 09:53:36 AM »
ReplyReply

Quote
Get in the habit of doing well exposed photographs, that do only need some minor corrections, which can be done inside the raw conversion software

I have never, ever advocated sloppiness in exposure, focus, lighting, or any other photographic technique. An image well-exposed in-camera is only the first step toward a good print. Most images benefit from curves, sharpening, B&W conversion and toning, and other post-processing adjustments that are best done in Photoshop rather than the RAW converter. Limiting yourself to RAW converter adjustments, especially when adjusting color gamut to a specific output device is crude and imprecise; rather like climbing a mountain by hopping on one foot when you have two normal and healthy arms and legs.

You are also espousing a workflow that limits your images to the capabilities of whatever printer you happen to be using at the moment, with no thought to inevitable technological improvements. As new printers are introduced with greater dynamic range and wider gamut, you will have to rework your images from scratch, starting with RAW conversion and redoing all subsequent edits and adjustments, if you wish to use the capabilities of the new printer.

With the workflow I recommend, all you have to do is (possibly) redo the final gamut tweak, convert to the new profile, and save a new version of the master, and the curves, sharpening, noise reduction, compositing, and other adjustments to the image do not have to be redone. This is especially valuable when a single image needs to be output on several devices, or even using multiple paper types in a single printer, as each paper type will have a different gamut, dynamic range, and associated profile.
Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8899


« Reply #9 on: July 10, 2006, 01:12:38 AM »
ReplyReply

I tend to agree with Jonathan on these point, so he must be right   .
Logged
Stephen Best
Guest
« Reply #10 on: July 10, 2006, 03:16:55 AM »
ReplyReply

Quote
The rendering engine (as currently implemented) does not examine what colors are in the file, but merely compresses the larger space by a predetermined amount.
[a href=\"index.php?act=findpost&pid=70058\"][{POST_SNAPBACK}][/a]

More precisely, perceptual is compressing the colours within the reference space (PCS), not the input space. The size/shape of the input space is irrelevant for the conversion. Take an image in sRGB and print it out. Convert the image to ProPhoto and print it out. Compare the two. There should be no difference. The gamut mapping diagrams are merely a simplification ... and the source of some confusion it seems.
« Last Edit: July 10, 2006, 06:22:50 AM by Stephen Best » Logged
dlashier
Sr. Member
****
Offline Offline

Posts: 518



WWW
« Reply #11 on: July 10, 2006, 03:47:54 AM »
ReplyReply

Quote
Get in the habit of doing well exposed photographs, that do only need some minor corrections, which can be done inside the raw conversion software
[{POST_SNAPBACK}][/a]
I tend to agree with this, for many or most images. Examples are [a href=\"http://www.lashier.com/home.cfm?dir_cat=26101]do it in C1 or PS[/url] and River Hybrid. OTOH there are cases (as in some of my other adjustment examples) where Photoshop is indispensible. Adobe is finally recognizing this by even producing a product such as Lightroom.

- DL
« Last Edit: July 10, 2006, 03:50:15 AM by dlashier » Logged

bjanes
Sr. Member
****
Offline Offline

Posts: 2822



« Reply #12 on: July 10, 2006, 07:05:12 AM »
ReplyReply

Quote
More precisely, perceptual is compressing the colours within the reference space (PCS), not the input space. The size/shape of the input space is irrelevant for the conversion. Take an image in sRGB and print it out. Convert the image to ProPhoto and print it out. Compare the two. There should be no difference. The gamut mapping diagrams are merely a simplification ... and the source of some confusion it seems.
[{POST_SNAPBACK}][/a]

In the context of my statment, we were talking about converting into ProPhotoRGB and then mapping it into a smaller space such as a printer space or sRGB for web output, and were not considering converting from sRGB to Prophoto. Furthermore, in converting from one space to another, Photoshop does not convert the input file to the PCS (CIE LAB) and then to the output space. This would be too slow. Andrew Rodney covers this in this color management book, I think, and perhaps he will jump in here. Where CIE LAB comes in is that the RGB coordinates for the conversion are expressed in CIE LAB and a three-by-three matrix transform is done. See Pointon for an example:

[a href=\"http://www.poynton.com/notes/colour_and_gamma/ColorFAQ.html#RTFToC20]http://www.poynton.com/notes/colour_and_ga...Q.html#RTFToC20[/url]

In any case, in your example, convterting from sRGB into ProPhoto, there can be a difference in the prints. If you print the converted ProPhoto with relative colorimetric, there will be no difference, since there could be no out of gamut colors. If you used perceptual rendering, then there would be a difference, since the entire gamut of ProPhoto would be compressed by a predetermined amount, even though there are no out of gamut colors.
Logged
Stephen Best
Guest
« Reply #13 on: July 10, 2006, 08:03:32 AM »
ReplyReply

Quote
If you used perceptual rendering, then there would be a difference, since the entire gamut of ProPhoto would be compressed by a predetermined amount, even though there are no out of gamut colors.
[a href=\"index.php?act=findpost&pid=70239\"][{POST_SNAPBACK}][/a]

Try it.
Logged
Hermie
Full Member
***
Offline Offline

Posts: 207


« Reply #14 on: July 10, 2006, 01:31:58 PM »
ReplyReply

Quote
More precisely, perceptual is compressing the colours within the reference space (PCS), not the input space. The size/shape of the input space is irrelevant for the conversion.
[a href=\"index.php?act=findpost&pid=70221\"][{POST_SNAPBACK}][/a]

Gamut mapping of perceptual rendering intent (lut profiles only!) is fixed and done when profiling software generates the profile. The gamut mapping is incorporated in the lookup tables.
This also means that the profiling software has to make an assumtion on the source colors (size of gamut, shape).

Herman
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2822



« Reply #15 on: July 10, 2006, 02:05:27 PM »
ReplyReply

Quote
Gamut mapping of perceptual rendering intent (lut profiles only!) is fixed and done when profiling software generates the profile. The gamut mapping is incorporated in the lookup tables.
This also means that the profiling software has to make an assumtion on the source colors (size of gamut, shape).

Herman
[{POST_SNAPBACK}][/a]

Yes, if perceptual rendering from ProPhotoRGB into a smaller space such as a printer profile is done, unlike what Steven Best seems to think, there will be compression according to the LUT in the profile. There is a good article by Mike Chaney on this subject. The entire gamut of ProPhotoRGB is not compressed, but the maker of the profile has to make assumptions on what colors the source is likely to contain.

[a href=\"http://www.steves-digicams.com/techcorner/July_2005.html]http://www.steves-digicams.com/techcorner/July_2005.html[/url]
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2822



« Reply #16 on: July 10, 2006, 02:20:30 PM »
ReplyReply

duplicate message deleted by author
« Last Edit: July 10, 2006, 05:20:48 PM by bjanes » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2822



« Reply #17 on: July 10, 2006, 02:56:53 PM »
ReplyReply

Quote
Try it.
[a href=\"index.php?act=findpost&pid=70246\"][{POST_SNAPBACK}][/a]

To try it, one needs profiles with perceptual rendering intent lookup tables. Photoshop permits perceptual conversions even when the appropriate lookup tables are not present, for example as in matrix based spaces such as ProPhotoRGB and sRGB, where no lookup tables exist. If you perform perceptual and relative colorimetric conversions with these matrix based spaces, you will see no differences, since conversion is always relative colorimetric (clipping) even when one checks perceptual conversion. PSCS2 does not tell you about the problem.

I did try it for an Epson printer profile from DryCreek.com and saw no difference, at least converting to the printer space in both modes and comparing both images brought into photoshop as layers and using the difference blending mode. On investigation, I found that no perceptual tables were present. Someone else will have to do the experiment.
Logged
Stephen Best
Guest
« Reply #18 on: July 10, 2006, 03:22:54 PM »
ReplyReply

Quote
I did try it for an Epson printer profile from DryCreek.com and saw no difference, at least converting to the printer space in both modes and comparing both images brought into photoshop as layers and using the difference blending mode. On investigation, I found that no perceptual tables were present. Someone else will have to do the experiment.
[a href=\"index.php?act=findpost&pid=70277\"][{POST_SNAPBACK}][/a]

A simpler technique is to look at the Proof Color readout in Photoshop's info palette.

As Hermie above has explained, the tables are fixed. There is no mechanism for output rendering to know the gamut of the input (working) space. The individual colours within the space get remapped, not the gamut. It's the same for perceptual and colorimetric.
Logged
Stephen Best
Guest
« Reply #19 on: July 10, 2006, 03:44:35 PM »
ReplyReply

Quote
There is a good article on the Steve's Digicam site by Mike Chaney on this subject.

http://www.steves-digicams.com/techcorner/July_2005.html
[a href=\"index.php?act=findpost&pid=70274\"][{POST_SNAPBACK}][/a]

From the article you linked:

"Some misinformation on the web would lead you to believe that because the CMM cannot account for image gamut, that it simply compresses the entire gamut of the color space used by the image so that it fits inside the printer's gamut. This is, however, also not true. Squashing the entire color space where the image resides into the printer's color space would amount to taking the entire wire frame above and shrinking it in size so that no corners protrude through the printer's solid gamut in the diagram. As you may be able to see by the graphic, that would be an extreme amount of compression that would result in noticeable color desaturation. In addition, it would mean that the same image encoded in two different color spaces of different sizes (say sRGB and Adobe RGB) would result in two different prints with different amounts of desaturation even though the original images should appear the same as they both have all the same colors:  they are just encoded differently."
Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad