Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: how to select out of gamut colors in Photoshop?  (Read 7611 times)
smilem
Full Member
***
Offline Offline

Posts: 216


« on: August 20, 2013, 03:46:58 PM »
ReplyReply

Hello, there is the Select -> Color range then choose out of gamut, but it usually selects nonsense and in CS3 there is not such thing at all.

Isn't there some sort of plugin that would do this right? Select out of gamut colors marked by Photoshop "Gamut warning" ?

If selected properly I could select convert to profile only the areas out of gamut to desired profile.
Logged
MarkM
Sr. Member
****
Online Online

Posts: 284



WWW
« Reply #1 on: August 20, 2013, 07:09:23 PM »
ReplyReply

If selected properly I could select convert to profile only the areas out of gamut to desired profile.

First, I don't think you can convert the profile of a selection. It's all or nothing.

Second, selecting out-of-gamut colors in a file doesn't really make sense. All the colors in the file are, by definition, within the gamut of the image's color space. These colors might be out of gamut of a different space, but in order to select these the color range filter would need to know which space to use for the gamut test. I suppose it could use the space selected in proof setup, but Photoshop doesn't seem to do this. Instead it looks like it assumes colors on the edge of the space are clipping and selects those. This is not a particularly useful thing in my opinion and these colors are not technically out-of-gamut.

If you are trying to convert to a smaller space, just convert the image and allow the rendering intent to do what it's designed to do. This is very often the best solution. If you find this isn't working, the next thing to try is working in the soft proof view and making adjustments that look better to you—then convert.
Logged

jmlphotography
Newbie
*
Offline Offline

Posts: 16


« Reply #2 on: August 22, 2013, 07:40:44 PM »
ReplyReply

Here's one way that might work:

http://www.thethinnegative.com/ttn-blog/2011/7/22/the-photoshop-hsbhsl-plug-in.html
Logged
Oldfox
Jr. Member
**
Offline Offline

Posts: 76


« Reply #3 on: August 25, 2013, 12:58:45 AM »
ReplyReply

This is a good question.

(In my CS3 there is Select -> Color Range -> Select -> Out of Gamut).

I have always thought that this selects the out-of-gamut pixels that are outside the gamut selected in View -> Proof Setup -> Custom. But it seems that this is not the case.

Does anyone really know what is happening in Select -> Color Range -> Select -> Out of Gamut ?
Logged
smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #4 on: September 02, 2013, 02:43:24 PM »
ReplyReply


There is no such plugin for CS3 AFAIK
Logged
digitaldog
Sr. Member
****
Online Online

Posts: 8954



WWW
« Reply #5 on: September 02, 2013, 02:48:35 PM »
ReplyReply

The gamut warning and selection isn't accurate (one could say buggy) and unnecessary. The output profile is going to do this for you anyway, and a perceptual intent is going to do so with a lot more smarts, assuming a good profile than you could ever do yourself.
Logged

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

Posts: 76


« Reply #6 on: September 03, 2013, 11:59:57 PM »
ReplyReply

and unnecessary.
unnecessary?

Really? No. If it would work and you could choose the profile, it could be very useful. If you are using Relative instead of Perceptual, you could tweak the out-of-bounds pixels without affecting the in-of-bounds pixels.

 
Logged
D Fosse
Sr. Member
****
Offline Offline

Posts: 322



« Reply #7 on: September 04, 2013, 01:07:55 AM »
ReplyReply

It seems to me there are two schools here. One is to let the profile handle the clipping. The other is to bring everything into gamut before conversion.

Personally I'm in the latter camp, but I'm not fanatic about it. The main argument is that the display limits the usefulness of proofing (many or most output profiles partially exceed the gamut of the display, even wide gamut ones). But it's time-consuming.
Logged
smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #8 on: September 04, 2013, 04:05:04 AM »
ReplyReply

Let me toss a wrench here  Tongue

Why Photoshop colorpicker does not show color delta E ? It you also be very useful.
The same goes for out of gamut colors, you can't specify delta E when a color is out of gamut. There are no boundaries shown (like a weather map) etc.

If you know any plugin that would do those 2 things, I'm all ears.

Professional tools ? Photoshop is not that professional after all if in all those versions they lack this basic functionality.

Hey Adobe, should I use Excel spreadsheet to do the calculations? No thanks.
Logged
digitaldog
Sr. Member
****
Online Online

Posts: 8954



WWW
« Reply #9 on: September 04, 2013, 09:11:43 AM »
ReplyReply

unnecessary?
Really?

That's my story and I'm sticking to it <g>.

You can use screwdriver as a kitchen knife or even for brain surgery but I don't recommend it. The tools in Photoshop for OOG control are buggy (not accurate) and crude. It treats 1% and 100% OOG the same. It predates soft proofing introduced in 1998. I've tried it, it isn't very effective and I've illustrated the differences in my LR soft proof video. I'm working in a very wide gamut space that exceeds my wide gamut display as does my printer. There's a lot I can't see the profile can affect and does affect. Let alone having an ugly overlay over my image as I attempt to edit that ugly overlay. IOW, the soft proof while not perfect shows me a lot more than a sold overlay that treats all OOG colors as if they were the same. The tools to adjust the overlay is as crude as that screwdriver in an operating room. Note that some could say that it is more accurate to use all Photoshop brush tools set to a single pixel for editing but I don't recommend it! Now if you charge by the hour, futzing with the OOG overlay seems like a good business plan.

IF Photoshop were to improve the feedback, selection and tools to control OGG would I use them? Perhaps. I don't see that happening anytime soon and in the meantime, good ICC profiles have vastly more 'precision' to affect individual pixels than anyone can manually and the soft proof at least shows me, within the gamut limitations of the display (but not necessarily the limitations of the output device) what I'll get and DO get.

Quote
Why Photoshop colorpicker does not show color delta E ? It you also be very useful.

Agreed and that's something you can see in ColorThink Pro (but forget editing that image).
Logged

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

Posts: 3572


« Reply #10 on: September 04, 2013, 10:12:31 AM »
ReplyReply

The tools in Photoshop for OOG control are buggy (not accurate) and crude. It treats 1% and 100% OOG the same. It predates soft proofing introduced in 1998.

Why not do the following (I know it is not elegant, as it should be, but it allows to stay in Photoshop):
- Make a duplicate of the image
- Convert that duplicate to the destination profile.
- Shift and Drag a copy of that back onto the original as a new layer and accept the Profile mismatch warning (preserve appearance).
- Change blending mode of that layer to difference, and create an ALT + Merge Visible layer of that.
- Disable the intermediate copy, or delete it if you are confident.
- Then change that new difference layer's blending mode to e.g. Hue or Color blending.

You now have a display of the OOG areas in color on a Black and white original, with intensity as a guide to the largest differences and in which colors to do the adjustments in the original, or rather on an adjustment layer. The actual corrections may differ depending on the best remedy, maybe brightness, maybe saturation, maybe a hue shift, etc.

Made into an Action, it becomes very simple to repeat after some corrections have been made. As I said, not elegant, but accurate and it can be tuned for different situations.

Cheers,
Bart
Logged
digitaldog
Sr. Member
****
Online Online

Posts: 8954



WWW
« Reply #11 on: September 04, 2013, 10:52:06 AM »
ReplyReply

You now have a display of the OOG areas in color on a Black and white original, with intensity as a guide to the largest differences and in which colors to do the adjustments in the original, or rather on an adjustment layer. The actual corrections may differ depending on the best remedy, maybe brightness, maybe saturation, maybe a hue shift, etc.

Not sure I understand the benefit of all this. You've used the output profile and RI you'll use anyway to handle the OOG it will deal with anyway. Why not just convert using the profile based presumably on the soft proof you viewed to select the RI and be done?
Logged

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

Posts: 3572


« Reply #12 on: September 04, 2013, 01:10:07 PM »
ReplyReply

Not sure I understand the benefit of all this. You've used the output profile and RI you'll use anyway to handle the OOG it will deal with anyway. Why not just convert using the profile based presumably on the soft proof you viewed to select the RI and be done?

Hi Andrew,

After converting, it is hard to know what exactly has changed and by how much. Checking the original against the converted version will allow to exactly see the magnitude of the losses because they won't come back after converting back to the original colorspace,  especially with the Relative Colorimetric RI. What was lost (quantized to same values) remains lost, and what changed only a bit will still be only a bit different, as with some parts of a Perceptual RI.

It is a very good predictor of problematic colors.

Cheers,
Bart
Logged
digitaldog
Sr. Member
****
Online Online

Posts: 8954



WWW
« Reply #13 on: September 04, 2013, 01:19:46 PM »
ReplyReply

After converting, it is hard to know what exactly has changed and by how much. Checking the original against the converted version will allow to exactly see the magnitude of the losses because they won't come back after converting back to the original colorspace,  especially with the Relative Colorimetric RI. What was lost (quantized to same values) remains lost, and what changed only a bit will still be only a bit different, as with some parts of a Perceptual RI.

Sorry, I don't see the practical implications. You have an output profile and you have to use it. You soft proof and decide visually what RI to use because it looks better. So after this exercise you do what? Use a different RI, don't convert? OOG is a fact of life, I don't see the point agonizing over it. I don't see the point editing based on this. I do believe it's a very good idea to soft proof to select the RI you prefer. And I think it's a good idea to edit in the original working space with the soft proof on while viewing a copy with the soft proof off in an attempt to make them match closer within reason. But short of that, I don't see what the exercise does that's useful.
Logged

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

Posts: 3572


« Reply #14 on: September 04, 2013, 01:32:18 PM »
ReplyReply

Sorry, I don't see the practical implications. You have an output profile and you have to use it. You soft proof and decide visually what RI to use because it looks better. So after this exercise you do what?

Adjust the 'original' before converting to the destination colorspace, and avoid problems such as lost of color detail.

Quote
I don't see the point editing based on this. I do believe it's a very good idea to soft proof to select the RI you prefer. And I think it's a good idea to edit in the original working space with the soft proof on while viewing a copy with the soft proof off in an attempt to make them match closer within reason. But short of that, I don't see what the exercise does that's useful.

Photoshop is not accurate enough to do well dosed good corrections based on its OOG indicators, not all displays have a sufficient gamut to preview the gamut issues of the output medium.

Cheers,
Bart
Logged
digitaldog
Sr. Member
****
Online Online

Posts: 8954



WWW
« Reply #15 on: September 04, 2013, 01:36:41 PM »
ReplyReply

Adjust the 'original' before converting to the destination colorspace, and avoid problems such as lost of color detail.
Which is what I do. Easier in LR which at least provides simple before and after previews. I see the soft proof while in the working space, I convert. That's again what I see. What I get isn't 100%, usually due to the limitations of a display gamut among a few other things. But the output isn't a surprise either. Dealing with OOG overlays in any mode just isn't as useful to me as a soft proof. Blocking my image and then expecting me to somehow make that overlay disappear is very 20th century Photoshop.
Logged

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

Posts: 3572


« Reply #16 on: September 04, 2013, 01:42:48 PM »
ReplyReply

Which is what I do. Easier in LR which at least provides simple before and after previews. I see the soft proof while in the working space, I convert. That's again what I see.

Exactly, and I thought that that is what the tread was a bout, how to get a meaningful prediction in Photoshop. LR is already much better in predictability, not perfect, but usable.

Quote
What I get isn't 100%, usually due to the limitations of a display gamut among a few other things. But the output isn't a surprise either. Dealing with OOG overlays in any mode just isn't as useful to me as a soft proof. Blocking my image and then expecting me to somehow make that overlay disappear is very 20th century Photoshop.

I agree, but one has to make do with the available tools, which is a compromise, but doable.

Cheers,
Bart
Logged
Oldfox
Jr. Member
**
Offline Offline

Posts: 76


« Reply #17 on: September 08, 2013, 01:47:20 AM »
ReplyReply

@Andrew. How would do you handle the out-of-gamut pixels using the Relative rendering intent?
« Last Edit: September 08, 2013, 02:20:57 AM by Oldfox » Logged
digitaldog
Sr. Member
****
Online Online

Posts: 8954



WWW
« Reply #18 on: September 08, 2013, 12:07:59 PM »
ReplyReply

@Andrew. How would do you handle the out-of-gamut pixels using the Relative rendering intent?

Again, with the ICC profile I'll use to convert the data. Assuming again that I prefer the gamut clip behavior of that profile on each image, based on the soft proof.
Logged

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

Posts: 76


« Reply #19 on: September 10, 2013, 01:01:26 AM »
ReplyReply

So if you dont like the clipping, you'll accept the Perceptual one?

In theory that maybe right. Here is an example of an image (tight crop). The Relative is very ugly because the shirt is totally ruined. The Perceptual is washed out (hopefully you can see it that from the upper corners).

For me choosing the shirt and tweaking with saturation, hue and lightness in an adjustment layer plus using the Relative gave the best result.

Therefore it would be nice to be able to select the out-of-gamut pixels properly.

Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad