Highlighted

compensating in color profile for 'levels' adjustments on document's pixels?

Explorer ,
Sep 06, 2020

Copy link to clipboard

Copied

I have a scanner, and an IT8 calibration card, and have used them (and the program "Rough Profiler") to generate a color profile specific to my scanner. So far so good; I can scan an image and then tag it with that color profile, to have its colors match the scanned material (when viewed on a properly-calibrated screen).

The thing is, though: any given thing that I scan, is likely to only use a subset of the scanner's gamut; the histogram is going to have empty space on either side of the 'mountain range' in the middle. So in the interest of how the image looks in programs that don't consider embedded calibration data, I'd like to adjust the image's levels to use the entire 0-255 range... but of course, if I did that, then it wouldn't actually match the scanned material's brightness any more.

However: concievably, couldn't I then just adjust the document's color profile information in *exactly* the *opposite* way that I just changed the pixel data, so the two operations would cancel out, and the result would still match the scanned material in a calibrated context (minus any tiny rounding errors of course)?

 

Here's an illustration that will hopefully demonstrate what I'm hoping to achieve:

whatiwant.png

 

I want to convert an image from the "my scanner's gamut" histogram to the "this document's gamut" histogram, but I want the end result to look the same as before, on a color-managed monitor. I want to change the literal numbers representing each pixel inside the file, and to NOT change the colors that each pixel would resolve to in a color-managed context -- which seems like it should be possible by adjusting the document's color profile information to take those changes into account; I just don't know any program that does that.

 

Can Photoshop do this? Does anyone know any other program that can? Can anyone suggest somewhere else to ask where people might know of one? (And: is there some reason not to do this, that hasn't occurred to me?)

 

Topics

Feature request, How to

Views

138

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more

compensating in color profile for 'levels' adjustments on document's pixels?

Explorer ,
Sep 06, 2020

Copy link to clipboard

Copied

I have a scanner, and an IT8 calibration card, and have used them (and the program "Rough Profiler") to generate a color profile specific to my scanner. So far so good; I can scan an image and then tag it with that color profile, to have its colors match the scanned material (when viewed on a properly-calibrated screen).

The thing is, though: any given thing that I scan, is likely to only use a subset of the scanner's gamut; the histogram is going to have empty space on either side of the 'mountain range' in the middle. So in the interest of how the image looks in programs that don't consider embedded calibration data, I'd like to adjust the image's levels to use the entire 0-255 range... but of course, if I did that, then it wouldn't actually match the scanned material's brightness any more.

However: concievably, couldn't I then just adjust the document's color profile information in *exactly* the *opposite* way that I just changed the pixel data, so the two operations would cancel out, and the result would still match the scanned material in a calibrated context (minus any tiny rounding errors of course)?

 

Here's an illustration that will hopefully demonstrate what I'm hoping to achieve:

whatiwant.png

 

I want to convert an image from the "my scanner's gamut" histogram to the "this document's gamut" histogram, but I want the end result to look the same as before, on a color-managed monitor. I want to change the literal numbers representing each pixel inside the file, and to NOT change the colors that each pixel would resolve to in a color-managed context -- which seems like it should be possible by adjusting the document's color profile information to take those changes into account; I just don't know any program that does that.

 

Can Photoshop do this? Does anyone know any other program that can? Can anyone suggest somewhere else to ask where people might know of one? (And: is there some reason not to do this, that hasn't occurred to me?)

 

Topics

Feature request, How to

Views

139

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Adobe Community Professional ,
Sep 07, 2020

Copy link to clipboard

Copied

Hi,

If you want an approximation of how an image looks in a program that doesn’t use an ICC profile (you call that ignoring the embedded calibration data, right?) simply [TEMPORARILY] assign sRGB (for standard gamut screens) or Adobe RGB (for wide gamut screens) to your image - rather than assigning the scanner profile

 

Its hard to know what you hope to gain with the edits based on histogram changes? 

Firstly, if you edit an image based on its histogram, that’s image specific of course.

Such large edits will likely intruduce quantisation errors unkless you work in 16 bit.

I fear that your proposed edits will mean that on any system the images will look crazily contrasty and saturated 

[unless your acanner profile is rather inaccurate]

I am very surprised that when expressed by scanning the IT8 and making a profile, your scanner's 'gamut' is smaller than sRGB 

[have you checked]?

If you want to experiment with editing the ICC profile then try editing the captured target in Photoshop before making the profile.

Move the target image in the opposite direction. so in this case you'd change the levels in the opposite direction to the edit you do on images 

 

intriguing, but I don't think you're going anywhere interesting, sadly. 

 

I hope this helps

if so, please "like" my reply and if you're OK now, please mark it as "correct", so that others who have similar issues can see the solution

thanks

neil barstow, colourmanagement.net :: adobe forum volunteer

[please do not use the reply button on a message within the thread, only use the blue reply button at the top of the page, this maintains the original thread title and chronological order of posts]

 

 

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Adobe Community Professional ,
Sep 07, 2020

Copy link to clipboard

Copied

I'm not sure I understand what you are trying to achieve. If an application does not use color management, then it will send the colors 'as is' to the monitor. Assuming that the user has an sRGB monitor, that means the image will be treated as if it were sRGB. To get the image to look (more or less) correctly in this case, you should use 'Convert to Profile' (not 'Assign Profile') in Photoshop and convert to sRGB. On color managed systems it will obviously also look correctly if you do this.

-- Johan W. Elzenga, http://www.johanfoto.com

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Explorer ,
Sep 07, 2020

Copy link to clipboard

Copied

Yeah, my intention was to scan in 16-bit, then apply this process, *then* drop to 8-bit; I was seeing it as a way to get the most precision out of 8 bits.

 

Filling the gamut on an image without a color profile doesn't make images look "crazily contrasty and saturated" generally speaking; this is what I've always done with my scans in the years before getting the IT8 card. I was just hoping for a way to continue to do so, without it throwing away information about what the image's real-world brightness is (and while getting the most precision possible out of 8-bit channels).

 

About my scanner's gamut being smaller than sRGB: honestly, I haven't looked into that, I just assumed that it would be. I was just trying to reference that the information would look the same relative to a standard gamut.

 

About converting the image to sRGB: wouldn't that just make the image paler (or crush the extremes) unless the image's range happens to line up with sRGB? What I want to do is increase the contrast when viewed without a profile, but preserve the original contrast when viewed with one.

 

Anyway, after asking around, I guess I'm getting the idea that trying to store profiled images in 8 bits just isn't done; maybe I should be looking into good compression formats for 16bpc images instead. Thanks!

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Adobe Community Professional ,
Sep 07, 2020

Copy link to clipboard

Copied

If the scanner gamut it smaller than sRGB, then converting to sRGB will not change the appearance of the image in any way. So no, it won't become paler. Visual changes are usually the result of doing this the wrong way: by using Assign Profile instead of Convert to Profile. 

 

-- Johan W. Elzenga, http://www.johanfoto.com

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Adobe Community Professional ,
Sep 07, 2020

Copy link to clipboard

Copied

The bit depth is totally independent of the color gamut. 

The scanner actually doesn't have a color gamut but that's another story. The gamut of the scanner profile is utterly limited by the gamut of the material used to make that profile (the target). Then there is the gamut of whatever you're scanning which might exceed the gamut of the target but you'll never know or get that. So the widest gamut you can produce is whatever you get from the scanner and is then defined by the assignment of the scanner profile. 

On top of all this is the color gamut of your display of which you say you wish to visually match to..... ?

The display color gamut and the gamut of the scanner defined by the scanner display will likely have a widely differing shape (and if you upload both, I can plot them in 3D for you). 

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Adobe Community Professional ,
Sep 08, 2020

Copy link to clipboard

Copied

Hi,

some of your reply comments made me think you'd not completely understood my post so I have edited it a little. please re read it.

 

The best way to optimise an image for viewing on systems which do not use a colour profile [on standard gamut screens] is very simple - convert a copy to sRGB.

[which, BTW - if the defined gamut IS larger than sRGB will "clip" the image. The gamut of most scanner profiles is larger that even Adobe RGB in some areas. I suggest you read up Joseph Holmes work his "Chrome" colour space, which expands on this subject]

 

As I mentioned, if you want to see how your unconverted image [i.e. in scanner colourspace] will appear without colour management then TEMPORARILY assign sRGB.

 

As Andrew mentioned, converting from 16 bit to 8 bit does not change gamut. 

 

hope it helps

 

neil barstow, colourmanagement.net :: adobe forum volunteer

 

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Explorer ,
Sep 13, 2020

Copy link to clipboard

Copied

So I think I've just realised one thing that makes my confusion about this matter make more sense -- it absolutely hadn't occurred to me that color profiles don't really concern themselves with absolute brightness. (Right?) So my intention of "preserving" the brightness of the image captured by my scanner is meaningless -- the intention of the color profile is to make sure that the individual channels' brightness values are mapped from one color space to another correctly in terms of the curves that map the values between the 0.0 and 1.0 of each profile, NOT to remap where the 0.0 and 1.0 fall relative to some real-world standard of brightness. So creating a new color profile just to "compensate" for playing with the levels of an image -- when all three channels are being changed together, which is the scenario I was picturing, which I'm now realising might not have been clear -- is meaningless, because a color profile wouldn't change that anyway...

 

Am I closer to having my head around this? I'm sorry for leaping into this question so uninformed -- I thought I understood this better than it turns out I did, a tricky situation to become aware of at the best of times... 🙂

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...
Adobe Community Professional ,
Sep 13, 2020

Copy link to clipboard

Copied

Brightness is a perceptual phenomena. Luminance (Luminosity) is a measure of the total radiant energy from a body. It has nothing to do with what a human perceives but rather describes the total radiant energy, such as watts/second of a source (the surface of a radiating object like a display). 

 

"Lightness" is a scaled value relative to the dynamic range of a system. In the CIE L*a*b* system L is "Lightness." Lightness is the perception of color without the chromaticity component and is scaled to a particular media or scene, such as
a display, a print, or a real natural scene. L50 is perceptually halfway between the white and black of a system. L50 may be 100 cd/m^2 or 1000 cd/m^2 you don’t know unless you also know luminance of white and black. Analogy: 50 paces would be a unit perceptually similar to both an ant and a human.

Likes

Translate

Translate

Report

Report
Community Guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
Reply
Loading...