From: "Maris V. Lidaka, Sr.", mlidaka@ameritech.net
Date: Tue, Nov 13, 2001, 10:44 AM
RE: [colortheory] False Profiles redux
Dan and others,Harking back to your article on false profiles, have you found that this does or does not work in the other direction - namely applying a false profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out image?
And if it does or could work, would you recommend a smaller color space such as sRGB or AppleRGB for it?
Maris Lidaka Sr
From: Dan Margulis,
Date: Wed, Nov 14, 2001, 5:28 AM
RE: [colortheory] False Profiles redux
Maris writes,>>Harking back to your article on false profiles, have you found that this does or does not work in the other direction - namely applying a false profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out image?>>
It works, however one has to be careful about hosing the shadow detail. Many apparently washed-out images in fact have full shadows. But the false profile may still be the best bet.
>>And if it does or could work, would you recommend a smaller color space such as sRGB or AppleRGB for it?>>
As with a false profile at a lower gamma, it would be image-specific. If it's worth making and saving a false profile at 2.6 it's certainly worth making another one with different color characteristics.
Dan Margulis
From: gowens01GO@netscape.net
Date: Wed, Nov 14, 2001, 9:32 AM
RE: [colortheory] Re: False Profiles redux
> Harking back to your article on false profiles, have you found that this does or does not work in the other direction - namely applying a false profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out image?>It seems to me that the great choices we have in terms of gamma and different work spaces is a wonderful creative tool. One should be able to manipulate these tools to achieve whatever illusion one desires.
How do others feel about this? Or is there a tendancy to be limited by the "rules" and hold a hard line? And what kind of wisdom must one invoke in terms of workflow and output? How much is creativity really limited by the "rules" of Photoshop and the printing industry?
Gary Owens
From: Todd Flashner, tflash@earthlink.net
Date: Wed, Nov 14, 2001, 10:36 AM
RE: Re: [colortheory] Re: False Profiles redux
on 11/14/01 12:31 PM, gowens01GO@netscape.net wrote:> what kind of wisdom must one
> invoke in terms of workflow and output? How much is creativity really
> limited by the "rules" of Photoshop and the printing industry?The proof is in the outpudding. ;-)
Todd Flashner
From: Andrew Rodney, andrew@digitaldog.net
Date: Wed, Nov 14, 2001, 11:06 AM
RE: Re: [colortheory] Re: False Profiles redux
on 11/14/01 10:31 AM, Gary Owens wrote:>It seems to me that the great choices we have in terms of gamma and
> different work spaces is a wonderful creative tool. One should be
> able to manipulate these tools to achieve whatever illusion one
> desires.
> How do others feel about this? Or is there a tendancy to be limited
> by the "rules" and hold a hard line? And what kind of wisdom must one
> invoke in terms of workflow and output? How much is creativity really
> limited by the "rules" of Photoshop and the printing industry?IF you understand what Photoshop is doing and how the term "False Profile" is incorrect, you1ll see no rules are being broken.
A file is simply a set of numbers and without a profile have no meaning. Photoshop no longer simply throws raw numbers to the screen like other stupid applications (like a browser for example). It uses the meaning of the numbers it gets from an embedded profile and then looks at the users display profile to produce an accurate preview.
When you send an untagged file to Photoshop 6 it simply makes an educated guess about what the numbers mean. That is, it looks at your preferred RGB, CMYK or Grayscale profiles in the color settings. Simply informing Photoshop 6 what the correct meaning of the numbers is, the preview is now correct.
You can have a file that is untagged and looks absolutely awful on screen. The problem IS NOT THE FILE, it1s the lack of a profile for Photoshop 6 to know how to preview the file. If you Assign the correct profile (or a profile that is close to the real meaning of the file), you get a better preview and ultimately a much better conversion to some other space. The term False Profile is incorrect and misleading. What it really is, is a 3good guess2 profile. If we actually had an accurate custom generated profile for the file and tagged the file, it would now be properly described. But we don1t have a profile. We make one up! Using Dan1s technique is ingenious and very useful. That he labels the technique incorrectly only confuses how Photoshop 6 deals with color files.
If I have a good profile for a file, I get a good preview. I can edit that profile to get a better preview. Is that edited profile false? No, it is simply closer to the ultimate meaning of the data.
If you want to try this, simply take a file and convert it to say Wide Gamut RGB. Now untag the file and it will look awful on screen in virtually any application (raw numbers being sent to the screen). Tag the file with Adobe RGB and it looks a bit better (depending on the working space you1ve got in the color settings). You could open Adobe RGB and enter the numbers that would get it much closer (or actually match) Wide Gamut RGB and your preview would be fine again. Is it false? No, it1s correct!
What this all leads to is how awful dealing with untagged files really are! IF all the files really were tagged correctly, we1d KNOW that the preview on screen looks bad because the file is bad. What people need to realize is that files can look truly awful on screen (because there isn1t a correct profile being assigned) and with the assignment of the correct profile, the file looks, converts and prints just fine. Why try and 3fix2 an awful looking file when the data is just fine but the preview is science fiction?
Andrew Rodney
From: "Maris V. Lidaka, Sr.", mlidaka@ameritech.net
Date: Wed, Nov 14, 2001, 12:32 PM
RE: Re: [colortheory] Re: False Profiles redux
Andrew,I would prefer to discuss the substance rather than the semantics - I really do not care what you choose to call it. Dan called it "false profile" in the article, so I use the term "false profile" so that everyone else reading this list knows what I am talking about.
Maris Lidaka Sr
Riga Company
From: "Jerry L. P'Simer", jpsimer@twmi.rr.com
Date: Wed, Nov 14, 2001, 3:13 PM
RE: Re: [colortheory] Re: False Profiles redux
Andrew Rodney wrote:> What this all leads to is how awful dealing with untagged files really are!
> IF all the files really were tagged correctly, we1d KNOW that the preview on
> screen looks bad because the file is bad. What people need to realize is
> that files can look truly awful on screen (because there isn1t a correct
> profile being assigned) and with the assignment of the correct profile, the
> file looks, converts and prints just fine. Why try and 3fix2 an awful
> looking file when the data is just fine but the preview is science fiction?So if you have an image that looks horrible but it has a correct profile describing its numbers then what? If you can assign a different profile that makes the image look much better should this profile be refered to as the new and improved correct profile? You abviously have a problem with the term "false profile", so how about coming up with a new term that you are happy with.
Jerry P'Simer
From: "Jerry L. P'Simer", jpsimer@twmi.rr.com
Date: Wed, Nov 14, 2001, 2:25 PM
RE: Re: [colortheory] Re: False Profiles redux
Gary Owens wrote:> How do others feel about this? Or is there a tendancy to be limited
> by the "rules" and hold a hard line? And what kind of wisdom must one
> invoke in terms of workflow and output? How much is creativity really
> limited by the "rules" of Photoshop and the printing industry?In terms of image editing, (silk purse from sows ear) I would be inclined to believe that there are no rules.
Jerry P'Simer
From: Andrew Rodney, andrew@digitaldog.net
Date: Wed, Nov 14, 2001, 3:22 PM
RE: Re: [colortheory] Re: False Profiles redux
on 11/14/01 4:10 PM, Jerry L. P'Simer wrote:> So if you have an image that looks horible but it has a correct profile
> describing its numbers then what?Then you use Photoshop 6 to correct the image (change the pixels). That is naturally what Photoshop 6 does.
> If you can assign a different profile that
> makes the image look much better should this profile be refered to as the new
> and improved correct profile?
If it is untagged, yes.> You abviously have a problem with the term "false
> profile", so how about coming up with a new term that you are happy with.Correct profile is far more accurate.
Andrew Rodney
From: "Maris V. Lidaka, Sr.", mlidaka@ameritech.net
Date: Wed, Nov 14, 2001, 3:37 PM
RE: Re: [colortheory] Re: False Profiles redux
Let's all just get along and call it "optimal profile". Then we can move on and deal with the image and not the grammar.Maris
From: Andrew Rodney, andrew@digitaldog.net
To: Color Theory, colortheory@yahoogroups.com
Date: Wed, Nov 14, 2001, 3:56 PM
RE: Re: [colortheory] Re: False Profiles redux
on 11/14/01 4:37 PM, Maris V. Lidaka, Sr. wrote:> Let's all just get along and call it "optimal profile". Then we can move on
> and deal with the image and not the grammar.
I'll buy that. Good name!Andrew Rodney
From: "Jerry L. P'Simer", jpsimer@twmi.rr.com
To: Dan Margulis, 76270,1033
Date: Wed, Nov 14, 2001, 4:07 PM
RE: Re: [colortheory] Re: False Profiles redux
Maris V. Lidaka, Sr. wrote:> Let's all just get along and call it "optimal profile". Then we can move on and deal with the image and not the grammar.
Good choice!
Jerry
From: "Jerry L. P'Simer", jpsimer@twmi.rr.com
Date: Wed, Nov 14, 2001, 4:04 PM
RE: Re: [colortheory] Re: False Profiles redux
Andrew Rodney wrote:> Correct profile is far more accurate.
Original Question:
Harking back to your article on false profiles, have you found that this does or does not work in the other direction - namely applying a false profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out image?New And Improved Question:
Harking back to your article on correct profiles, have you found that this does or does not work in the other direction--namely applying a correct profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out image?To me the question makes more sense in its original form. Personally I like the term "False Profile", it's more colorful. Excuse the pun:).
Jerry
From: Stephen Marsh
Date: Thu, Nov 15, 2001, 5:51 AM
RE: [colortheory] Re: False Profiles redux
Maris writes,
> Harking back to your article on false profiles, have you found that this
> does or does not work in the other direction - namely applying a false
> profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out
> image?
Dan replies,
> It works, however one has to be careful about hosing the shadow detail.
> Many apparently washed-out images in fact have full shadows. But the false
> profile may still be the best bet.
I have not used the high gamma settings (3.0) as much as the low gamma ones (1.0). When I have had to 'bulk up' an image--result did not seem as good as when lightening a dark image.
With regard to Dan's comment on loosing shadow detail, I use a separate layer with the false data copied in with a conversion to that space. Then I use the layer options blend if sliders to quickly isolate or 'mask' the wanted/unwanted detail - then a very quick layer mask can be painted to refine the transition.
Maris writes,
> >>And if it does or could work, would you recommend a smaller color space
> such as sRGB or AppleRGB for it?>>
Dan replies,
> As with a false profile at a lower gamma, it would be image-specific. If
> it's worth making and saving a false profile at 2.6 it's certainly worth
> making another one with different color characteristics.
I agree, it depends on the case - and editing whitepoint or RGB primaries can be useful but not as simple as changing to a false gamma.
Leaving the subject of applying a false profile on an image that has little tone or depth - and now editing the numbers. Often the assigning of a false profile is only the start of extreme edits.
Stacking some duped layers in multiply mode can work - or using a curves adjustment layer set to multiply mode.
Starting with a washed out RGB description is often needed, since multiply mode builds up saturation as well as other faded colour components.
Hope this helps,
Stephen Marsh.
From: samarsh@ozemail.com.au, samarsh@ozemail.com.au
Date: Thu, Nov 15, 2001, 5:15 AM
RE: [colortheory] Re: False Profiles redux
Maris V. Lidaka, Sr. wrote:>> Let's all just get along and call it "optimal profile". Then we can move on and deal with the image and not the grammar.>>
I hate to wade in at this point, just when all current parties are feeling happy about the jargon...but --
For me, false profile is still the term that best describes the theory behind the move.
Back in the 'old days', before Adobe adopted widespread internal use of ICC profile conversions, experienced CMYK users would commonly create a "FALSE SEPARATION".
This was generated on the fly using the old separation setup and ink settings, for the specific conversion and image content.
Examples include spot channel creation from RGB originals via false separation methods, creating special seps for images being used in rich black settings or many more image or output specific situations.
Dan's book PP6 has many examples of a false separation in CMYK - which is really a false profile, since all colour settings are actually ICC profiles in ver6.
Technically, you were providing SWOP seps or whatever, and the the final file was within specifications - its just that you got there via a 'false method' that did not meet the ideal.
It was the final numbers that mattered, not how you got them.
As a long term CMYK user, I never had any issues with this term - since it was an accurate term for describing the intent behind the move.
Now in the 'infamous' article at the EP site - Mr Margulis used the word 'False Profile' to describe an RGB move.
I see no difference between customising CMYK settings on an image or output based case, and altering gamma or RGB or whitepoint descriptions when dealing with problem RGB files.
For a file with an accurate known embedded profile - a totally different or related profile to the correct tagged profile can be used for artistic or technical reasons.
The new tag is clearly false, but it produces some effect which is desirable when this profile describes the numbers to a device or colour transform.
Some parties have issues with tthe term False Profile when it is used to describe an untagged file (or perhaps in general). True, having the correct profile would help - and lets assume that the correct profile can be obtained and assigned to the image.
If you then decide to assign a new profile which you now know is incorrect, but you like the results - I see no problem with the term false profle being applied, since this is known to be true.
Optimal profile might be a good term, but to me the choice of an optimal profile is when you have no idea of the true description of the numbers - and have to guess (both via the numbers and visually).
False profile is a good term to describe a _hack_ to a profile, to create a false outcome for every other image - except the one the hack is written for. So a false profile then becomesa 'true' profile...which is where it gets confusing and could be why some parties dislike the term.
When you understand the theory and intent behind the moves - which I believe I do, the term false seems like the best description, IMHO.
To me, it's not just semantics.
I agree with the recent posts that discussing how one uses the techniques is more important than getting bogged down in arguments on naming conventions. With this in mind, I will make a separate post addressing the original question.<g>
Sincerely,
Stephen Marsh.
From: Dan Margulis,
Date: Thu, Nov 15, 2001, 7:38 AM
RE: Re: [colortheory] Re: False Profiles redux
Jerry writes,>>New And Improved Question:
Harking back to your article on correct profiles, have you found that this does or does not work in the other direction - namely applying a correct profile with gamma of say 2.6 or even 3.0 to an underexposed, bleached-out image?>>Since I ordinarily work at a 1.8 gamma I always have the option of just assigning sRGB or Adobe RGB, both of which are 2.2 gammas, if I think the image is too light. I did show that method in the article.
For me to use something higher than 2.2 would imply something grossly washed out. I don't get too many pictures of that nature, whereas originals that are grossly too dark seem to be increasingly common.
AFAIK assigning a higher gamma will work, subject to my earlier remarks about the possibility of plugging the shadows. The only time I have tried it was with the excruciatingly flat images supplied by Todd Flashner for my tests with 16-bit vs. 8-bit. For the live tests I didn't use the false profile, because the objective was to make extreme changes with curves in 8-bit. However, for comparison's sake I tried assigning a higher gamma profile and it looked like this would be the way to handle a live job that came in that light.
Dan Margulis
From: "Maris V. Lidaka, Sr.", mlidaka@ameritech.net
Date: Thu, Nov 15, 2001, 8:52 AM
RE: Re: [colortheory] Re: False Profiles redux
Thank you - it worked very well on the series of images I had. Levels and curves just did not to the job, even curves in the luminosity channel of LAB, until after the application of 3.0 gamma..Maris Lidaka Sr
| AFAIK assigning a higher gamma will work, subject to my earlier remarks
| about the possibility of plugging the shadows.. The only time I have tried
| it was with the excruciatingly flat images supplied by Todd Flashner for my
| tests with 16-bit vs. 8-bit. For the live tests I didn't use the false
| profile, because the objective was to make extreme changes with curves in
| 8-bit. However, for comparison's sake I tried assigning a higher gamma
| profile and it looked like this would be the way to handle a live job that | came in that light.
From: "Jerry L. P'Simer", jpsimer@twmi.rr.com
Date: Fri, Nov 16, 2001, 2:32 PM
RE: Re: [colortheory] Re: False Profiles redux
Stephen Marsh wrote:> Leaving the subject of applying a false profile on an image that
> has little tone or depth - and now editing the numbers. Often the
> assigning of a false profile is only the start of extreme edits.
> Stacking some duped layers in multiply mode can work - or
> using a curves adjustment layer set to multiply mode.
> Starting with a washed out RGB description is often needed,
> since multiply mode builds up saturation as well as other faded
> colour components.If you have an image that is washed out but the color is otherwise reasonable you can avoid over saturation by first duplicating the base layer. Set the duped layer to luminosity, then set up a clipping group with a curve correction set to multiply. This allows the multiply factor to effect only the luminosity of the base layer. At this point you can use the opacity slider, blend if sliders, a layer mask containing actual image data, or any combination of the three to control the outcome.
Jerry P'Simer
From: Dan Margulis
Date: Fri, Nov 23, 2001, 11:16 PM
RE: [colortheory] False Profile
A list member sent the following to me privately, but in view of the interest in this area recently I thought others might be interested in the answer.--------------- Forwarded Message ---------------
To: Dan Margulis
Date: Fri, 23 Nov 2001 15:23:44 -0500
Subject: FalseProfileDan,
Sorry to bother you with my ignorance. Am having trouble making a useable false profile. Following the steps in your Aug. '01 EP article, i end up with a .csf file, while the files available in the Assign Profile dialog box seem to show only .icc files. Where am i going wrong. . . please?--------------------------------------------------------
You are hitting the "Save" button within the Edit: Color Settings dialog box. This saves *all* of your color-related settings, CMYK, RGB, grayscale, color management, whatever--as a single .csf file. This is a lovely feature, since it means that if somebody else has been working on your machine and you have no clue what he may have done to your settings, you can restore all of your color settings in a single command.
However, it's not the one you're looking for here. Assuming that what you are trying to do is create a profile that uses Adobe RGB primaries but with a gamma of 1.4:
1) Be sure "Advanced Mode" is checked in Edit: Color Settings.
2) Within Color Settings, change RGB to Adobe RGB.
3) Release the mouse and click again on Adobe RGB and hold. Select "Custom RGB".
4) The dialog box will now show the Adobe RGB primaries and 2.2 gamma. Change the gamma to 1.4, and change the name to something like "1.4 gamma Adobe RGB" so that it won't be confused with ordinary Adobe RGB. Click OK.
5) "1.4 gamma Adobe RGB" will now appear as your RGB working space. Click and hold on it, and one of the options will now be "Save RGB". If you do this (rather than save the entire color settings) you will get a profile you can dump into your ColorSync folder.
Dan Margulis
Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.