Ad
Ad
Ad
Pages: [1] 2 3 »   Bottom of Page
Print
Author Topic: ProPhoto RGB used better in a flexible workflow?  (Read 12764 times)
Hendrik
Jr. Member
**
Offline Offline

Posts: 93


WWW
« on: June 06, 2006, 06:58:26 AM »
ReplyReply

Iím always trying to educate myself (reading Danís LAB-book, reading PS and color management books and threads). The purpose: maximize the quality of my images within the practical boundaries. I have the time to fine-tune every image I make.

In the past I used Adobe RGB as my working color space, now I use ProPhoto RGB as my working color space. One of the reasons is LLís ĎUnderstanding ProPhoto RGBí article.

Since I mostly shoot portraits, ProPhoto RGB has not always benefits, since all colors captured fits often in a smaller color space. I want to use a workflow that combines the benefits of a large color space (not throwing colors away) with the benefits of a smaller color space (finer control over color and tone, because the data points are packed closer together).

Question: It this a sensible workflow? I want to convert my RAW images to a large color space, say ProPhoto RGB and check with PS the out-of-gamut colors in a smaller color space like Adobe RGB. When there are no out-of-gamut colors I can safely convert my RAW image to Adobe RGB without losing any colors. Maybe even convert to sRGB when my out-of-gamut test shows no problems. I continue my editing in this smaller color space. This way I try to fit my image as tight as possible in a color space, maximizing all benefits.

Best,

Hendrik  
Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #1 on: June 06, 2006, 07:24:14 AM »
ReplyReply

Quote
In the past I used Adobe RGB as my working color space, now I use ProPhoto RGB as my working color space. One of the reasons is LLís ĎUnderstanding ProPhoto RGBí article.
There is a problem with PS and ProPhotoRGB, as discussed here in this thread: As I understand it, PS is not able to do a perceptual rendering with ProPhotoRGB, since it's a matrix based profile. It does relative colormetric instead. Your intention is, to compress or map the OOG colors smoothly into your destination color space, that's what a perceptive rendering intent is supposed to do. RelCol on the other hand means, that OOG colors are simply clipped, so at a first sight, there's no use in ProPhotoRGB, since all the extra color is clipped. In the above linked thread, there's a procedere from Peter Lange to get the tonalities to the smaller gamut. I am still evaluating on this. Another method is, to reduce the saturation locally with a saturation mask, using the gamut warning feature in PS.

Quote
I want to convert my RAW images to a large color space, say ProPhoto RGB and check with PS the out-of-gamut colors in a smaller color space like Adobe RGB. When there are no out-of-gamut colors I can safely convert my RAW image to Adobe RGB without losing any colors.
This is no problem: Open your image with the ProPhotoRGB profile and do a soft proof: Go to View>Proof Setup>Custom... and set your destination color space, eg. AdobeRGB. Then activate Proof Colors and Gamut Warning, and you'll see all the OOG colors as grey.
Logged

Best Regards

Dennis.
Hendrik
Jr. Member
**
Offline Offline

Posts: 93


WWW
« Reply #2 on: June 06, 2006, 11:56:30 AM »
ReplyReply

Quote
There is a problem with PS and ProPhotoRGB, as discussed here in this thread: As I understand it, PS is not able to do a perceptual rendering with ProPhotoRGB, since it's a matrix based profile. It does relative colormetric instead. ...

Thank you for the interesting thread; it seems another aspect to keep in mind (will it ever end?  )


More and more photographers use ProPhoto RGB as there standard working space, regardless the contents of the image (that includes me). Articles like ĎUnderstanding ProPhoto RGB with subtitle ĎPreferred Working Space for Digital Photographersí suggests such a way of thinking.

I think itís better to use the color space that fits your image best, whether itís ProPhoto RGB, Adobe RGB or sRGB. Maybe I overlook something, thatís why I put it here for evaluation.

Thanks.  
Logged
bruce fraser
Jr. Member
**
Offline Offline

Posts: 51


« Reply #3 on: June 06, 2006, 03:55:27 PM »
ReplyReply

Quote
There is a problem with PS and ProPhotoRGB, as discussed here in this thread: As I understand it, PS is not able to do a perceptual rendering with ProPhotoRGB, since it's a matrix based profile. It does relative colormetric instead. Your intention is, to compress or map the OOG colors smoothly into your destination color space, that's what a perceptive rendering intent is supposed to do. RelCol on the other hand means, that OOG colors are simply clipped, so at a first sight, there's no use in ProPhotoRGB, since all the extra color is clipped.
[a href=\"index.php?act=findpost&pid=67527\"][{POST_SNAPBACK}][/a]

This is true if, and only if, the profile to which you are converting is also a matrix profile (which is one of the reasons I think it's daft to go from ProPhoto to an intermediate working space). If you go direct from ProPhoto to print space, you have complete use of all the rendering intents, but more importantly, you have complete use of all the data you captured.
Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #4 on: June 06, 2006, 04:06:51 PM »
ReplyReply

Quote
This is true if, and only if, the profile to which you are converting is also a matrix profile
Thanks for the correction and clarification. I knew it couldn't be the whole story... I was just looking at a picture of yellow flower, converted via ACR to PrpPhotoRGB, which was completely out of gamut in sRGB. At least the gamut warning showed this:

So, if the OOG color would be clipped via conversion from ProPhotoRGB to sRGB, the whole flower should become a plain yellow spot.

But how do I know, which proiles are matrix based, and thus are problematic? And when I choose RelCol with the above picture, the flower is rendered fine. Why? Alle the yellow colors are OOG, and thus should be clipped. Is the gamut warning not working correctly?
Logged

Best Regards

Dennis.
Hermie
Full Member
***
Offline Offline

Posts: 207


« Reply #5 on: June 06, 2006, 04:46:16 PM »
ReplyReply

Quote
But how do I know, which proiles are matrix based,
[{POST_SNAPBACK}][/a]

File size is an indication, something like 1KB is matrix.
Most working spaces are matrix based. There's one lut based working space afaik (PhotoGamut RGB).

Use a free utility like ICC Profile Inspector to explore more details of profiles:
[a href=\"http://www.color.org/profileinspector.html]http://www.color.org/profileinspector.html[/url]

Herman
« Last Edit: June 06, 2006, 05:10:51 PM by Hermie » Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #6 on: June 06, 2006, 05:20:24 PM »
ReplyReply

Quote
File size is an indication, something like 1KB is matrix.
Aha, thanks. This is correct for all my RGB profiles.

So the above mentioned problem is valid for all RGB-to-RGB conversions, right? That would explain, why there's litteraly no difference in converting a picture perceptive or RelCol from ProPhotoRGB to any RGB, while the differncies are pretty obvious converting to a CMYK profile.

But again, I don't understand, why OOG colors are not clipped going from ProPhotoRGB to sRGB, if the true RI is supposed to be RelCol - whatever RI you might choose.
Logged

Best Regards

Dennis.
Hendrik
Jr. Member
**
Offline Offline

Posts: 93


WWW
« Reply #7 on: June 07, 2006, 07:37:10 AM »
ReplyReply

Quote
...
But again, I don't understand, why OOG colors are not clipped going from ProPhotoRGB to sRGB, if the true RI is supposed to be RelCol - whatever RI you might choose.
[a href=\"index.php?act=findpost&pid=67570\"][{POST_SNAPBACK}][/a]

Yes they are clipped. I can see clipping (loosing detail) when I soft proof a highly saturated ProPhoto RGB image to sRGB.

You can see/test it yourself when you use the flower image from bjanes from the thread Dennis quoted.
Logged
Hendrik
Jr. Member
**
Offline Offline

Posts: 93


WWW
« Reply #8 on: June 07, 2006, 07:55:00 AM »
ReplyReply

I hope you donít mind bringing this thread back on-topic, since I need more input for my own reassurance.  

I think Andrew Rodney answered my question, although itís quoted from another thread: ďA working space is a container for holding you image data and it has a fixed gamut. You can shoot a gray card in RAW and encode that data into two working spaces of differing gamuts. It might fully fit within sRGB and it will of course fit in ProPhoto with a heck of a lot of additional gamut around this data. So looking at the scene gamut before even deciding what working space to use for encoding is useful. If you can fit this gray into sRGB, using ProPhoto buys you nothing

ProPhoto RGB has no advantages if the color gamut of your image fits in another smaller color space. Therefore I think itís wise to check for OOG colors and adjust your working color space accordingly.

Put into practice:
No OOG colors in sRGB, then sRGB is the preferred working color space.
No OOG colors in Adobe RGB, but OOG colors in sRGB, use Adobe RGB.
No OOG colors in ProPhoto RGB, but OOG colors in Adobe RGB, use ProPhoto RGB.

Considerations:

- There are no disadvantages when using smaller color spaces (if there are no OOG colors), but is there an advantage? A wider gamut gives you a wider range of color, but it doesn't give you more colors. The size of the working space's gamut determines the spacing of all possible values of each channel. The same number of colors is simply stretched over a larger color range. This means using a smaller color space gives you finer control over the color.

- High-bit editing gives you 32768 levels per color channel, but your tools use 256 possible values. Maybe (banding) artifacts are reduced, but your color control while editing is not any better. Öor is it? This brings me to Ö

- Does using ProPhoto RGB have any real disadvantages when you work with small gamut 48-bit images. For example, I remember something that perceptual rendering (to the printer profile) attempts to compress the gamut of the source space into the gamut of the target space. Does that mean there is also a change in saturation when the image's gamut fits easily into the target space and there are no OOG colors?

I hope some can give more insight in this matter. I lack the knowledge to fully understand all possible consequences, but Iím tempted to leave ProPhoto RGB as my standard working space and use the space that fits my image.

Thanks!

Hendrik
Logged
PeterLange
Guest
« Reply #9 on: June 07, 2006, 03:13:17 PM »
ReplyReply

Quote
Yes they are clipped. I can see clipping (loosing detail) when I soft proof a highly saturated ProPhoto RGB image to sRGB.

You can see/test it yourself when you use the flower image from bjanes from the thread Dennis quoted.
Is that so ??
Somehow a key question.

Once again Iíve opened the nice flower from bjanes' earlier post (in ProPhoto RGB), changed to 16 bit/ch.  Then, I visited the main Color Settings tab in Photoshop, advanced settings, in order to reduce the saturation of monitor colors by 50%.  This is to get this whole story within the limited capabilities of my humble monitor (tested in advance via a Granger Rainbow).

Now, after installing a SoftProof to sRGB, Iím toggling it via Ctrl + Y.

That way, I do NOT see any loss of details...(even at some hundred % magnification).

The only conclusion I have, is that all visible &  relevant details survive the conversion to sRGB.  Though channel clipping occurs, former out-of-sRGB colors seem to find enough 'room' on the surface of sRGB to be still distinguishable - in particular considering that merged colors do not necessarily have to be located next to each other in the image.

Iíd be pleased if we could sort this out here.

Peter

--
Logged
Hermie
Full Member
***
Offline Offline

Posts: 207


« Reply #10 on: June 07, 2006, 05:06:14 PM »
ReplyReply

> The only conclusion I have, is that all visible & relevant details survive the conversion to sRGB. Though channel clipping occurs, former out-of-sRGB colors seem to find enough 'room' on the surface of sRGB to be still distinguishable - in particular considering that merged colors do not necessarily have to be located next to each other in the image.

> Iíd be pleased if we could sort this out here.

Peter,

I thought I found a clue when I examined the bjanes image in ColorThink. ColorThink includes a very nifty feature to visualize the mapping/clipping of OOG colors
It's probably due to the chromatic adaptation transform from ProPhoto (D50) to sRGB (D65), that shifts OOG *AND* and in-gamut colors.

Only problem is that I can't reproduce this in Photoshop. I can't see a difference between absolute and relative colorimetric conversion in PS. In ColorThink the distinction is obvious. Any further clues?

Herman
« Last Edit: June 07, 2006, 05:11:35 PM by Hermie » Logged
Stephen Best
Guest
« Reply #11 on: June 07, 2006, 07:07:44 PM »
ReplyReply

Quote
I remember something that perceptual rendering (to the printer profile) attempts to compress the gamut of the source space into the gamut of the target space. Does that mean there is also a change in saturation when the image's gamut fits easily into the target space and there are no OOG colors?
[a href=\"index.php?act=findpost&pid=67609\"][{POST_SNAPBACK}][/a]

It doesn't work like that. The conversion (from working to output space) is a two-stage process where the source colours (and not gamut) first get converted into the connection space and then from there to output. So the gamut of each is independent. You can verify this by creating an sRGB image filled with say 255,0,0 and using Photoshop's Info box to readout Proof Color. Now use Convert Profile to convert it to ProPhoto RGB. The image RGB values will change, but the Proof Color will be the same.
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6926


WWW
« Reply #12 on: June 07, 2006, 08:35:45 PM »
ReplyReply

My understanding of this is that the rendering intent is designed to handle out of gamut colours in different ways depending on the choice of rendering intent. Most of the technical discussion of rendering intent focuses on how the rendering intent handles the relationship between out of gamut colours and in-gamut colours. This relationship is not that simple according to Fraser et. al. in "Real World Color Management". When faced with OOG colours, Perceptual rendering intent will move the OOG colours into gamut at the same time that it shifts many of the ingamut colours so that overall colour RELATIONSHIPS are preserved (on the principle that human perception is more sensitive to the relationship between colours rather than to their absolute hues). In so doing, it will shift some colours that are in gamut, and it will also desaturate some colours. Fraser et. al. go on to say that the visible differences between these rendering intents is quite subtle. I have not seen discussion of the impact on colour saturation from changing rendering intent for images that have no OOG colours, but the information that is available would tend to suggest you should see extremely little - if any - saturation impact, say moving between Relative Colorimetric and Perceptual because there are no OOG adjustments to be made that would impact other colours in the image.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Stephen Best
Guest
« Reply #13 on: June 07, 2006, 09:49:08 PM »
ReplyReply

Quote
Fraser et. al. go on to say that the visible differences between these rendering intents is quite subtle. I have not seen discussion of the impact on colour saturation from changing rendering intent for images that have no OOG colours, but the information that is available would tend to suggest you should see extremely little - if any - saturation impact, say moving between Relative Colorimetric and Perceptual because there are no OOG adjustments to be made that would impact other colours in the image.
[a href=\"index.php?act=findpost&pid=67662\"][{POST_SNAPBACK}][/a]

The process isn't adaptive, that is it doesn't know/care whether all colours are in the output gamut or not. It just remaps everything according to fixed tables. (This after the input colours have been converted to the connection space.) The difference between colorimetric and perceptual depends entirely on the perceptual tables themselves. Some could be designed for more compression at the output gamut boundary, others could spread the shifts more evenly throughout the space. The only way to know is to soft-proof or print.
« Last Edit: June 07, 2006, 10:31:39 PM by Stephen Best » Logged
Hendrik
Jr. Member
**
Offline Offline

Posts: 93


WWW
« Reply #14 on: June 08, 2006, 04:09:51 AM »
ReplyReply

Lets assume that soft proofing on a calibrated monitor has some predictive value. Maybe itís not a good assumption, but what purpose does it have if we didnít.

I see clearly a loss of detail when I soft proof, maybe even some very slight banding. I attached a screen capture. Itís more apparent on my screen when I toggle between the two versions.

Soft proof settings: RelCol, Ďblack point compensationí checked.

@Peter, if I do the same as you did (50% desaturation of monitor colors and 16bit/channel), I still see the same, maybe even better noticeable.

@MarkDS, the differences between Perceptual rendering and RelCol isnít valid here, my (newly) understanding is that it doesnít matter with conversions between matrix profiles. Whatever you select, you only can use Colorimetric rendering.

Btw, anyone of the experts a comment on my topic?  
« Last Edit: June 08, 2006, 04:11:14 AM by Hendrik » Logged
Stephen Best
Guest
« Reply #15 on: June 08, 2006, 05:40:29 AM »
ReplyReply

Quote
Btw, anyone of the experts a comment on my topic? 
[{POST_SNAPBACK}][/a]

I'm no expert, but I think the answer to your original post is that it will work fine (provided ProPhoto is in 16-bits) ... but maybe it doesn't really matter that much. While I can think of colours from the original outside of Adobe RGB that I've managed to transfer intact to print it has been pretty rare. The downsides of larger spaces are of course that any adjustments you do will be coarser (sometimes too coarse) and you can't actually see what you're doing. You may want to read the following:

[a href=\"http://www.jeremydaalder.com/singleArticle.php?articleID=6]http://www.jeremydaalder.com/singleArticle.php?articleID=6[/url]

Note that in the above, the gamut of the current K3 inkset is larger that for the 2100 so this point no longer applies.

Personally, I'm coming to the opinion that it's better to just settle on a mid-sized space and concentrate more on differentiation of in-gamut colours (al la Dan's book) than maximizing the gamut itself. It will depend a lot on your subject matter. If you're starting from RAW you can always go back for the colours you've missed if you find some use for them in the future.
Logged
Hendrik
Jr. Member
**
Offline Offline

Posts: 93


WWW
« Reply #16 on: June 08, 2006, 06:32:50 AM »
ReplyReply

@ Stephan

Thank you for the link. This is in essence exactly what I was thinking, though must better written. It answers all my questions. More and more I became uncomfortable using ProPhoto RGB as my standard color space, all I needed was a confirmation to change my way of working.  
« Last Edit: June 08, 2006, 06:34:01 AM by Hendrik » Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6926


WWW
« Reply #17 on: June 08, 2006, 06:46:36 AM »
ReplyReply

Quote
I'm no expert, but I think the answer to your original post is that it will work fine (provided ProPhoto is in 16-bits) ... but maybe it doesn't really matter that much. While I can think of colours from the original outside of Adobe RGB that I've managed to transfer intact to print it has been pretty rare. The downsides of larger spaces are of course that any adjustments you do will be coarser (sometimes too coarse) and you can't actually see what you're doing. You may want to read the following:

http://www.jeremydaalder.com/singleArticle.php?articleID=6

Note that in the above, the gamut of the current K3 inkset is larger that for the 2100 so this point no longer applies.

Personally, I'm coming to the opinion that it's better to just settle on a mid-sized space and concentrate more on differentiation of in-gamut colours (al la Dan's book) than maximizing the gamut itself. It will depend a lot on your subject matter. If you're starting from RAW you can always go back for the colours you've missed if you find some use for them in the future.
[a href=\"index.php?act=findpost&pid=67693\"][{POST_SNAPBACK}][/a]

The most important point you made above is the one about the K3 inkset. Since that point is correct the remainder of Daalder's argument for not using ProPhoto largely (but not unexceptionally) falls apart provided one works in 16-bit depth.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Stephen Best
Guest
« Reply #18 on: June 08, 2006, 07:11:39 AM »
ReplyReply

Quote
The most important point you made above is the one about the K3 inkset. Since that point is correct the remainder of Daalder's argument for not using ProPhoto largely (but not unexceptionally) falls apart provided one works in 16-bit depth.
[a href=\"index.php?act=findpost&pid=67697\"][{POST_SNAPBACK}][/a]

I work exclusively in 16-bit but find fine tuning of images somewhat difficult in ProPhoto, or even Beta RGB which is a better match for my scanned originals. Personally, I think the benefits of ProPhoto are largely overstated. Further that it's possible to get too hung up in absolute gamut size (and Dmax) and lose sight of what you're trying to achieve but this is a different argument!

As I tried to say above, for output it doesn't matter which working space you're using (assuming the same image colours) so use whatever *you* feel comfortable with.
Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #19 on: June 08, 2006, 07:18:45 AM »
ReplyReply

Quote
That way, I do NOT see any loss of details...(even at some hundred % magnification).

The only conclusion I have, is that all visible &  relevant details survive the conversion to sRGB.  Though channel clipping occurs, former out-of-sRGB colors seem to find enough 'room' on the surface of sRGB to be still distinguishable - in particular considering that merged colors do not necessarily have to be located next to each other in the image.
Maybe this due to the fact, that the image was converted to 8bit, which is a nonsense with a large gamut. A raw image would make a far better starting point. With my own experiences, I definately can see a difference:

Raw image converted to ProPhotoRGB at 16bit with an exposure compensation to bring the histogram off the right end. Compared to a straight 8bit sRGB conversion and set the "Reduce Monitor Saturation" to -30%, there are differencies. Using desaturation and gamut warning, it is possible to get those fine tonalities into 8bit sRGB.
Logged

Best Regards

Dennis.
Pages: [1] 2 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad