I actually believe there's some good stuff in the PP5E book, but it's way harder to get at those jewels and I find myself dreading the work required to try to get something out of it. And, some of what he preaches in PP5E, I just don't agree with. Separating color from contrast (learned in the LAB book) just makes intuitive sense to me, is easy to use and works way, way better for me than separate R, G and B or separate C, M, Y and K curves. I just can't make the separate curves work for me without constantly fighting color changes that I don't want. I think this is a really, really advanced concept that requires you to implicitly know how to move all three curves so that you get what you are looking for without undesirable color changes and that's way beyond me.
Dan actually addresses this issue in Chapter 9 of PP5E. He recommends a somewhat complicated work with separate steps for contrast and for color. This turns out not to be hard. You do your curving for contrast then change the mode of the curve layer to luminosity. No color effects. Then you use other techniques to get the color right (in color mode).
I sympathize with your frustration here, but I also understand Dan's difficulties in getting a complex subject across.
The LAB techniques in his book are easy to comprehend and to use, but they have somehwat limited use. Dan remarked that people exposed to these techniques tend to over-utilitze them. I agree, having been one of them.
If you want to improve your curving, I'd recommend doing it in CMYK. It's
more precise, and so it doesn't present the color changing problems quite
as much as RGB. Lots of caveats, of course, depending on the image.
Note to Rutt: One idea to kick start things in the reading group, would be for someone to put together a summary of the top 5 things that can be quickly learned and applied from the PP5E book and perhaps that would get some people reinvigorated and get the conversation going again. That's what was so great about the LAB book - you got a high return early on with things that are simple to learn and easily usable by many. I haven't found those things in PP5E yet.
Excellent suggestion, but the application may be a little tricky. One of the values of the LAB books was a recipe that works in a lot of cases (even though you can get better results in some cases using other methods). PP5E doesn't offer that. Instead, it offers a set of principles (first half of the book) and a set of techniques that apply to some images but not others (last half of the book).
I don't know how this would work online, but you might try to replicate what Dan does in his basic class, perhaps using the images distributed with the book. Just thinking out lowd here, but it might work in a reading group context.
Dan actually addresses this issue in Chapter 9 of PP5E. He recommends a somewhat complicated work with separate steps for contrast and for color. This turns out not to be hard. You do your curving for contrast then change the mode of the curve layer to luminosity. No color effects. Then you use other techniques to get the color right (in color mode).
I already do this (apply curves to individual RGB channels with a blend mode of Luminosity) and I learned it in the LAB book. It works great in some images. I even wrote up Chapter 14 in the LAB book Once for Color and Once for Contrast.
I would like to see the chapter for false gammas gone through slowly, in fine detail.
I see there is a lot of value in the technique, but find I am really unable to avail myself of it on my own. I think it will take me several days of intense work to really incorporate it in my own skillset, and yet I think that is something I need to do.
I already do this (apply curves to individual RGB channels with a blend mode of Luminosity) and I learned it in the LAB book. It works great in some images. I even wrote up Chapter 14 in the LAB book Once for Color and Once for Contrast.
If you're already doing it, then what's the problem?
Don't take me wrongly; I'm not being argumentative - just seeking clarification.
Indeed, I myself sometimes have experienced dissatisfaction trying to separate luminosity from color by working in those two layer modes in RGB & CMYK. I can't put my finger on it, but often things just don't "look right" after working in those modes, and I go back to Normal mode and get better results working on individual channels juggling tonality & color simultaneously. Maybe it's because I've been used to doing it that way for so long.
There's sometimes something wrong with working especially in Color mode (Luminosity, not so much). Can you put into words the shortcomings you encounter? I can't figure it out.
I would like to see the chapter for false gammas gone through slowly, in fine detail.
I see there is a lot of value in the technique, but find I am really unable to avail myself of it on my own. I think it will take me several days of intense work to really incorporate it in my own skillset, and yet I think that is something I need to do.
Last time I saw Dan, he gave us a midterm grade on our PP5E reading group which would take some generosity to interpret as a C. We discussed some remedial actions. Here is my best reconstruction of this conversation. (Keep in mind that this happened after a bit of a wine tasting...)
Skip chapters 11-14. These are entertaining enough but really of limited appeal. They aren't going to help anyone become a better retoucher except for targeting high volume commercial presses.
Chapters 15-18 are really cool and can make a huge difference in plenty of images. If you want your flowers to look spectacular, for example, this is a good place to start. The ideas here also form the basis of Dan's current experimental ideas, which I tried to summarize here. I'm having a lot of fun playing with these.
Chapters 10 and 20 are just total virtuoso mindblowers, designed, I think to emphasize the fact that Dan is so good at what he does. That last example with the beer is like one of those beautiful chess puzzles, so sweet it's almost art. I'd love to see someone do a take on these, but I understand why it's hard to apply to just any old images which happen to be lying around.
Chapter 9 is a crucial chapter. I've been thinking that I'd take it, but I'm so busy, I can say when that will be. One of the most common criticisms of Dan from people with only a surface understanding is that his methods won't help images without neutral white and dark points. This is far from true and this is where you find out why. Also I'm totally sick of finding purple horses and green skys and red oceans and bright orange people and having to explain each time about casts and how to check for them and how you stupid calibrated monitor doesn't amount to a hill of beans. I really do have a full scale rant in mind and the summary of this chapter would be the perfect place for it. Then I could just link to the rant every time instead of making it up from scratch every time.
When Dan was writing those late chapters, he was reaching for something that he didn't exactly get at the time but which he understands better now. I've outlined this workflow a few times: false profile to lighten, CMYK to get the shadows right, LAB for midtones and highlights with a K channel mask from the previous step, then USM with 3 channel HIRALOAM. The results of this workflow look lots different then LAB book portrait workflow and often more natural. I'd like to work this in somehow.
Last time I saw Dan, he gave us a midterm grade on our PP5E reading group which would take some generosity to interpret as a C.
And what I would tell Dan is that the reading group is a reflection of the effectiveness of his book. As great as the information in there may be, it is a failure for people like me. People who are not color theorists, but want some help understanding it and how to control it. He's lost touch with the need for an on-ramp, or at least with how to design a well-constructed on-ramp.
I would like to see the chapter for false gammas gone through slowly, in fine detail.
I see there is a lot of value in the technique, but find I am really unable to avail myself of it on my own. I think it will take me several days of intense work to really incorporate it in my own skillset, and yet I think that is something I need to do.
And what I would tell Dan is that the reading group is a reflection of the effectiveness of his book. As great as the information in there may be, it is a failure for people like me. People who are not color theorists, but want some help understanding it and how to control it. He's lost touch with the need for an on-ramp, or at least with how to design a well-constructed on-ramp.
I've told him so in so many words, and I think he mostly agrees. He does have the experience of his classes which he is trying to translate into these books....
I'd very much like to see the two books merged together with the early LAB material as an on-ramp. There are people (we won't name names here) who are NEVER going to accept the idea that LAB is actually easier than RGB, and I think these nameless people expected PP5E to be easier than the LAB book. But what can you do? Can't please all the people. Have to put a stake in the ground.
Dan says there was a lot of time pressure in writing PP5E because PP4E was going out of print.
Anyway, if you sign up to take a literature course and they assign Last of the Mohicans , it's no good saying you didn't finish it and write a paper about it just because it sucked (Mark Twain's opinion, not mine re Cooper.) We've seen our midterm grade and it needs improvement. In the best of all possible worlds, we'll inspire Dan to do better himself.
Anyone want to sign up to write up Ch 15, "The Art of the False Profile?" It's a great way to master this difficult material (well not that difficult, really) and I'll help you all I can, which, depending on how you look at it, is either a big bonus or drag.
I'm going to do something about Chapter 9. Don't know when, but I'm going to do it.
Duffy, it's not to soon for you to write up Ch 16. I know you don't agree with it 100% and that's actually a good thing. Also has CS3 been a deal changer? You can kick off a great discussion. Dan's opinion here is controversial and probably designed to be controversial. Let's have some fun with it.
Skip chapters 11-14. These are entertaining enough but really of limited appeal. They aren't going to help anyone become a better retoucher except for targeting high volume commercial presses.
Skipping Chapters 11-14 was what I did; Then I found Dan advocated using false gamma as the preferred way to correct major global exposure problems in later, 'cool' chapters, which left me puzzled. I mean, I've always heard people saying the middle slider in Levels adjustment is for gamma, so what's the advantage, especially since Levels allows continuous adjustment with instant feedback? I'm sure the answer lies somewhere in those skipped chapters, but at the moment I don't have the will to take them on; so I agree with prior posts that some help on the false gamma concept is needed.
Last time I saw Dan, he gave us a midterm grade on our PP5E reading group which would take some generosity to interpret as a C. We discussed some remedial actions. Here is my best reconstruction of this conversation. (Keep in mind that this happened after a bit of a wine tasting...)
Was Dan's "gentleman's C" grade based only on the subjects the reading group has focused on (which is all you discussed in your summary of the conversation), or also on other factors (quality of the reviews, amount of discussion, etc.)?
I'm a newcomer to the forum, and I've read many (but not all) of the threads. I've been quite impressed by effort that's gone into the chapter reviews and the originality of some of them. But there doesn't seem to be a lot of discussion/debate over the techniques presented. For some yes, for many no.
I guess it's not particularly important whether Dan thinks this a problem, but whether the reading group thinks this a problem.
Was Dan's "gentleman's C" grade based only on the subjects the reading group has focused on (which is all you discussed in your summary of the conversation), or also on other factors (quality of the reviews, amount of discussion, etc.)?
I made up the grade to describe what he said (after the aforementioned wine tasting, so don't expect me to quote verbatim.) He was comparing the PP5E group with the LAB book. We blew his socks off in that group. Here we are struggling to reinvigorate the PP5E book. The LAB group had so many participants who became proficient. Participants did so many great retouches.
The business with the master vs individual curves frustrated him and I'm not sure how much he has followed since. I'm going to have a chance to pin him down (maybe with no wine tasting going on.) I'll try to exchange constructive feedback (though he does read a lot of this stuff.)
Anyone want to sign up to write up Ch 15, "The Art of the False Profile?" It's a great way to master this difficult material (well not that difficult, really) and I'll help you all I can, which, depending on how you look at it, is either a big bonus or drag.
I reread Chapter 15 because it sounds like something that could be useful and I even wondered if I might volunteer to do the write-up on it, but I'm still not "getting it". I know there isn't a write-up of it yet and that would be the best place to ask questions about it, but I don't know when that will be and I have a question now.
Here goes the question. Let me know if there's a better place to put this question. I understand how a false profile can be used to change the luminosity gamma or the color space. What I don't understand is how messing with the luminosity gamma is any different than a luminosity curve? I'm one of those people who only really "gets" something when I know how it works under the hood. I intimately understand what transformation is happening with a curve. I imagine that a gamma adjustment targetted at luminosity is just a particular shaped curve. If that's the case, then I don't understand why false gamma profiles are any better than curves. If there's more to the luminosity part of a false gamma than a curve, then I'd love to understand how it works.
I do understand how certain color situations might benefit from assignment into a false color profile to give you a larger (perhaps even fictitious) color space to work in, manipulation in the larger color space to accomplish what you want while you have lots of room, then conversion back to a real color space. But, I don't understand that for luminosity. Why can't you just reproduce any gamma luminosity changes from a false profile with the appropriate set of luminosity-targetted curves. I could never find where Dan actually explained why they were better than curves. And, since I feel like I understand curves really well, I'm stuck as to why I should use false profiles for fixing overly dark shots.
I... I could never find where Dan actually explained why they were better than curves. And, since I feel like I understand curves really well, I'm stuck as to why I should use false profiles for fixing overly dark shots.
This is really a great question. The answer, I think is that it's particularly easy to create a large library of these false profiles and use them when you need them. You could do the same with curves, but it's more work to make them. With the variant color spaces, you can just plug in one number and you are done. I would be interested to see the resulting L curve, but it's not vital.
Dan loves to tell the story of a newspaper guy he knows who uses ONLY false profiles for image correction. It becomes a matter of knowing your library, not much different from people who like the filters, which can also be represented as curves (RGB, not LAB).
This is really a great question. The answer, I think is that it's particularly easy to create a large library of these false profiles and use them when you need them. You could do the same with curves, but it's more work to make them. With the variant color spaces, you can just plug in one number and you are done. I would be interested to see the resulting L curve, but it's not vital.
Dan loves to tell the story of a newspaper guy he knows who uses ONLY false profiles for image correction. It becomes a matter of knowing your library, not much different from people who like the filters, which can also be represented as curves (RGB, not LAB).
OK, I can accept that and that makes sense. It's a different tool that can be used to solve a problem. Depending upon the problem and your efficiency with other tools, it will be an individual choice whether this tool works better for you than curves.
But, other than the efficiency or convenience factor, it seems that a different gamma is just a pre-defined curve (with less control over the shape). This Wikipedia artile on gamma was useful for me and shows some gamma curves.
OK, I can accept that and that makes sense. It's a different tool that can be used to solve a problem. Depending upon the problem and your efficiency with other tools, it will be an individual choice whether this tool works better for you than curves.
But, other than the efficiency or convenience factor, it seems that a different gamma is just a pre-defined curve (with less control over the shape). This Wikipedia artile on gamma was useful for me and shows some gamma curves.
I had exactly the same reaction when I read about the overlay layers in the portrait technique, and it drove me to reverse engineer the related curves. So I basically agree with you. These days I'm finding I use both the overlay move from the portrait technique and the false profiles more and more. They really are big time savers once you get used to them. Get them on a layer and move the opacity slider as in the MFM technique. Just a few clicks, compared to writing a new curve or set of curves.
Keep in mind that LAB curves are not all powerful the way that RGB curves are. You cannot write one curve which causes a color shift in the highlights only and a different one in the shadows. Some false profiles can do this, but you have to be a wizard at photoshop's colorspace definitions (not me.)
I had exactly the same reaction when I read about the overlay layers in the portrait technique, and it drove me to reverse engineer the related curves. So I basically agree with you. These days I'm finding I use both the overlay move from the portrait technique and the false profiles more and more. They really are big time savers once you get used to them. Get them on a layer and move the opacity slider as in the MFM technique. Just a few clicks, compared to writing a new curve or set of curves.
Keep in mind that LAB curves are not all powerful the way that RGB curves are. You cannot write one curve which causes a color shift in the highlights only and a different one in the shadows. Some false profiles can do this, but you have to be a wizard at photoshop's colorspace definitions (not me.)
Got it. I guess I'm more interested in techniques that let me do new things I didn't know how to do before (like enhance colors in LAB or use a color channel for a mask or HIRALOAM sharpening or a red channel blend to enhance a sky) than I am in a faster way to do something I already know how.
It's not that I don't value speed, but it's not near as important to me as capability and I'm not a pro so time isn't money in my photography. Because I do so much event photography with lots of photos of a large number of people (often sports teams), my time on PP is dominated by my workflow to process a 300 image shoot, not by the time to PP a few of the extraordinary shots.
For your example of color correcting shadows and highlights differently, I have no problem just doing that with RGB curves or using two curve layers in LAB and targetting them to a luminosity range either with blend-if settings or a luminosity mask or a manually created mask.
What would be practical difference of having a library of false profiles VS having a library of curve presets?
I mean, to each is own, but is there a difference in results?
And if not, what would be the benefits doing it via - rather obscure and uncontrollable - false profiles compared to controlled to a T set of curves (except habit/sentimental purposes/feeling like dining with Gods/etc.)?
What would be practical difference of having a library of false profiles VS having a library of curve presets?
I mean, to each is own, but is there a difference in results?
And if not, what would be the benefits doing it via - rather obscure and uncontrollable - false profiles compared to controlled to a T set of curves (except habit/sentimental purposes/feeling like dining with Gods/etc.)?
Got me. I'm liking curves myself for luminosity changes. I already know them and I get to control their shape. I also tend to do big luminosity fixes in ACR before I get to PS anyway.
I can see some advantage to using false color profiles. Of particular interest is the wide gamut CMYK which lets you get to CMYK adjustments, channels, etc... without clipping colors like regular CMYK. But that's just as a temporary color space for the purposes of doing some CMYK adjustment before you go back to RGB.
I have been wondering about "dining with the Gods". I guess I wasn't invited .
I see three advantages to starting with a false profile move rather than a curves adjustment.
1) Non-destructive editing. Using a false profile changes appearance without changing brightness levels.
2) Make subsequent corrections easier. An image that's too light or too dark is already at an extreme end of what can be adjusted using curves. One goal of applying a false profile is to bring the image into a range in which subsequent corrections will be much easier.
3) Time savings. If subsequent corrections are easier they'll also be faster.
False profiles only seem rather obscure and uncontrollable until one acquires a feel for what they do. I felt the same way about using curves at first. Eventually I became comfortable using curves. The same thing happened with false profiles.
I think of false profiles as a jump start to using curves. As such, using a library of false profiles is for me a better workflow than using a library of saved curves adjustments.
Regards,
Michael
Michael D. Aery, KSC | Prepress Manager
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
I see three advantages to starting with a false profile move rather than a curves adjustment.
[...]
I think of false profiles as a jump start to using curves. As such, using a library of false profiles is for me a better workflow than using a library of saved curves adjustments.
Regards,
Michael
Great post! Thanks! So glad to have you on board. (Want to write a chapter summary?!)
Great post! Thanks! So glad to have you on board. (Want to write a chapter summary?!)
You're welcome!
I'm glad to be here.
Yes, I'd very much like to write a chapter summary.
I've finished reading the book, but I haven't finished reading all the other posts in this thread nor have I caught up with the posts in the reading group chapters. I'm real sorry to be joining the party so late. I wish I had discovered this site long ago!
I'm even more disappointed at missing the LAB reading group. I've read that book several times, it's one of my favorite books! Is it too late to contribute to that thread as well?
Obviously I'm new here, so I'm not sure what the gameplan is. How would you like me to proceed?
Regards,
Michael
Michael D. Aery, KSC | Prepress Manager
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
Yes, I'd very much like to write a chapter summary.
I was thinking about the "Art of the False Profile" chapter which is the topic of the moment and the beginning of a pretty interesting cycle of chapters. How about it?
I've finished reading the book, but I haven't finished reading all the other posts in this thread nor have I caught up with the posts in the reading group chapters. I'm real sorry to be joining the party so late. I wish I had discovered this site long ago!
Our "marketing department" wants to know how you found us.
I'm even more disappointed at missing the LAB reading group. I've read that book several times, it's one of my favorite books! Is it too late to contribute to that thread as well?
I was thinking about the "Art of the False Profile" chapter which is the topic of the moment and the beginning of a pretty interesting cycle of chapters. How about it?
Our "marketing department" wants to know how you found us.
Not at all.
"Art of the False Profile" it is! A topic near and dear to my heart.
I was told about dgrin by someone I met at Photoshop World in Boston while we were talking about Dan Margulis and the amazing presentations he did.
I noticed there doesn't appear to be a chapter 10 entry for the LAB group, maybe I could take a stab at that one as well? Was it skipped for a particular reason?
If I'm asking questions that've already been covered please note that I've only just started digging into these threads.
Regards,
Michael
Michael D. Aery, KSC | Prepress Manager
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
I see three advantages to starting with a false profile move rather than a curves adjustment.
1) Non-destructive editing. Using a false profile changes appearance without changing brightness levels.
2) Make subsequent corrections easier. An image that's too light or too dark is already at an extreme end of what can be adjusted using curves. One goal of applying a false profile is to bring the image into a range in which subsequent corrections will be much easier.
3) Time savings. If subsequent corrections are easier they'll also be faster.
False profiles only seem rather obscure and uncontrollable until one acquires a feel for what they do. I felt the same way about using curves at first. Eventually I became comfortable using curves. The same thing happened with false profiles.
I think of false profiles as a jump start to using curves. As such, using a library of false profiles is for me a better workflow than using a library of saved curves adjustments.
Regards,
Michael
Good points. I'm fine with the fact that some people will reach the conclusion that a false profile is a faster tool for them. As of this moment, I haven't reached that conclusion for myself so I thought I'd give you my curves answer to your three points.
Non-destructive editing. When I start work on an image that is seriously off in luminosity, the very first thing I do is apply a curve adjustment layer to get the image in the general ballpark for editing. I often use a different curve layer for getting the image in the right brightness ballpark than I do for tweaking/optimizing it. I think this is done for the same goal as a gamma adjustment with a false profile and is also non-destructive.
Make subsequent adjustments easier. If I've already gotten the image in the right ballpark with the first curve layer, then I think we're even here.
Time Savings. This is a personal experience thing. Whatever you believe for your own skills and workflow is what matters to you. Once the image is in the right ballpark, I don't think it matters whether that happened with a false gamma or a curve adjustment layer, so the potential time savings is in the initial gamma adjustment or initial curve adjustment. That will simply depend upon which tool a given PS user knows how to use more effectively. False gammas might be simpler to make a library of because they are just a number, but they are also less flexible in that they have no flexibility in shape.
I was told about dgrin by someone I met at Photoshop World in Boston while we were talking about Dan Margulis and the amazing presentations he did.
I's the culprit.
Dan Dill
"It is a magical time. I am reluctant to leave. Yet the shooting becomes more difficult, the path back grows black as it is without this last light. I don't do it anymore unless my husband is with me, as I am still afraid of the dark, smile.
This was truly last light, my legs were tired, my husband could no longer read and was anxious to leave, but the magic and I, we lingered........" Ginger Jones
Non-destructive editing. When I start work on an image that is seriously off in luminosity, the very first thing I do is apply a curve adjustment layer to get the image in the general ballpark for editing. I often use a different curve layer for getting the image in the right brightness ballpark than I do for tweaking/optimizing it. I think this is done for the same goal as a gamma adjustment with a false profile and is also non-destructive.
But it's not the same degree of non-destructive. An adjustment layer is non-destructive only in the sense that the changes made can be reversed. You are still changing brightness levels with your initial curve and any curve adjustment comes with a price. Eventually a curves adjustment will run out of room. In other words, curving has a potential to clip an image where false profiling does not.
A false profile move makes absolutely no changes to brightness levels, it changes their definition or what a given level means, and consequently does not carry the cost of a curves adjustment. This is truely non-destructive editing!
Make subsequent adjustments easier. If I've already gotten the image in the right ballpark with the first curve layer, then I think we're even here.
Not at all. Adjusting gamma is not the only objective of a false profile move.
Consider this; If an RGB image comes to me and it's too dark, it's also going to be too gray. The first thing I do is Edit>Assign Profile and choose a profile which will lower the gamma and boost the colors. This will bring the image into a range where subsequent corrections will be much easier, if they're needed at all.
In an 8bit RGB file, there are only 256 brightness levels per pixel per channel. This holds true whether one is working in sRGB, Adobe RGB, or Wide Gamut RGB. I am using these three profiles here because they all have the same 2.2 gamma.
Here's an analogy:
I have 256 dominoes; half inch thick, one inch wide, two inches tall.
If I pile them front to back I get a stack 128 inches high. This is my sRGB stack.
If I pile them side to side I get a stack 256 inches high. This is my Adobe RGB stack.
If I pile them top to bottom I get a stack 512 inches high. This is my Wide Gamut RGB stack.
Note that with all this stacking, I have not lost a single domino. This is truely non-destructive editing.
This is what Assign Profile does. It changes the appearance of the numbers while leaving the numbers unchanged.
That image which is too dark, being also too gray, comes to me from a digital camers that assigns sRGB. I assign a false profile, in this case I would use a custom Wide Gamut RGB with let's say a 1.4 gamma. With one move I have not only lightened the image, but I have also boosted the colors from their initial gray appearance. Again, all without changing a single brightness level.
In addition, any and all curves adjustments made in Wide Gamut RGB have much, much, greater impact on the image because the scale of the brightness levels has been widened.
Time Savings. This is a personal experience thing. Whatever you believe for your own skills and workflow is what matters to you. Once the image is in the right ballpark, I don't think it matters whether that happened with a false gamma or a curve adjustment layer, so the potential time savings is in the initial gamma adjustment or initial curve adjustment. That will simply depend upon which tool a given PS user knows how to use more effectively. False gammas might be simpler to make a library of because they are just a number, but they are also less flexible in that they have no flexibility in shape.
Again, a change in gamma is not my only goal. If that were the case it would be easier to use the Image>Adjustment>Exposure command.
By changing the scale of the brightness levels, going from a relatively narrow gamut profile to a wide gamut profile, I have far more leverage to engineer changes into my image. This is not to say one should always use a wide gamut working space. That would be like using a sledge hammer to drive in a thumbtack. At some point in the workflow I will eventually convert to some other profile either for output or further editing. But if the image requires a drastic first move, false profiling is a far more effective first step than a curve adjustment layer.
Note: I used the term "convert" just now very specifically. Assigning a profile and converting to a profile are very different moves. For the benefit of others who may read this post I will explain the difference...
There are two commands in PS under the Edit menu; Assign Profile & Convert to Profile. They both allow you to choose an ICC profile from a list of the profiles installed on your machine.
Assign Profile changes the appearance of an image without changing the numbers, by numbers I mean channel structure (RGB, LAB, CMYK) and brightness levels. A brightness level of 255r in sRGB is not the same color red as 255r in ProPhoto RGB.
Convert to Profile changes the numbers while attempting to not change the appearance, because out of gamut colors will change based on your choice of rendering intent.
To put it another way; Assign Profile tells PS what a given "red" means, Convert to Profile tells PS what numbers to use to get that "red" if it is possible to get that "red" using a specific channel structure.
In closing, I'm not advocating using false profiles instead of curves. But there is no question that the "Art of the False Profile" is a very valuable technique. Even if the above discussion has not convinced you of their worth, I would still advise adding this technique to your skill set simply because it gives you more choices in your workflow, and more choices means greater flexibility.
You've got me using up all my stuff for the chapter summary. D'oh!
Regards,
Michael
Michael D. Aery, KSC | Prepress Manager
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
You've got me using up all my stuff for the chapter summary. D'oh!
I look forward to your chapter writeup. For luminosity changes, I'm still not seeing a workflow advantage or finding an actual technical advantage of gamma profiles over curves (a gamma adjustment is technically described as a curve), but I'll let it sink in some and wait for your writeup. I do see some interesting possibilities for certain types of color manipulations because I don't know how to do the same thing other ways. My prior arguments above were about luminosity adjustments.
You've got me using up all my stuff for the chapter summary. D'oh!
Regards,
Michael
Thank you very much! I'll be looking for your write up. Hopefully with your explanations I'll be able to actually start using this powerful stuff..
(Great analogy with domino chips!)
Comments
I sympathize with your frustration here, but I also understand Dan's difficulties in getting a complex subject across.
The LAB techniques in his book are easy to comprehend and to use, but they have somehwat limited use. Dan remarked that people exposed to these techniques tend to over-utilitze them. I agree, having been one of them.
If you want to improve your curving, I'd recommend doing it in CMYK. It's
more precise, and so it doesn't present the color changing problems quite
as much as RGB. Lots of caveats, of course, depending on the image.
Excellent suggestion, but the application may be a little tricky. One of the values of the LAB books was a recipe that works in a lot of cases (even though you can get better results in some cases using other methods). PP5E doesn't offer that. Instead, it offers a set of principles (first half of the book) and a set of techniques that apply to some images but not others (last half of the book).
I don't know how this would work online, but you might try to replicate what Dan does in his basic class, perhaps using the images distributed with the book. Just thinking out lowd here, but it might work in a reading group context.
I already do this (apply curves to individual RGB channels with a blend mode of Luminosity) and I learned it in the LAB book. It works great in some images. I even wrote up Chapter 14 in the LAB book Once for Color and Once for Contrast.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
I see there is a lot of value in the technique, but find I am really unable to avail myself of it on my own. I think it will take me several days of intense work to really incorporate it in my own skillset, and yet I think that is something I need to do.
Moderator of the Technique Forum and Finishing School on Dgrin
Don't take me wrongly; I'm not being argumentative - just seeking clarification.
Indeed, I myself sometimes have experienced dissatisfaction trying to separate luminosity from color by working in those two layer modes in RGB & CMYK. I can't put my finger on it, but often things just don't "look right" after working in those modes, and I go back to Normal mode and get better results working on individual channels juggling tonality & color simultaneously. Maybe it's because I've been used to doing it that way for so long.
There's sometimes something wrong with working especially in Color mode (Luminosity, not so much). Can you put into words the shortcomings you encounter? I can't figure it out.
Last time I saw Dan, he gave us a midterm grade on our PP5E reading group which would take some generosity to interpret as a C. We discussed some remedial actions. Here is my best reconstruction of this conversation. (Keep in mind that this happened after a bit of a wine tasting...)
So much to do, so little time.
And what I would tell Dan is that the reading group is a reflection of the effectiveness of his book. As great as the information in there may be, it is a failure for people like me. People who are not color theorists, but want some help understanding it and how to control it. He's lost touch with the need for an on-ramp, or at least with how to design a well-constructed on-ramp.
Dgrin FAQ | Me | Workshops
Same here. False gammas seem promising.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
I've told him so in so many words, and I think he mostly agrees. He does have the experience of his classes which he is trying to translate into these books....
I'd very much like to see the two books merged together with the early LAB material as an on-ramp. There are people (we won't name names here) who are NEVER going to accept the idea that LAB is actually easier than RGB, and I think these nameless people expected PP5E to be easier than the LAB book. But what can you do? Can't please all the people. Have to put a stake in the ground.
Dan says there was a lot of time pressure in writing PP5E because PP4E was going out of print.
Anyway, if you sign up to take a literature course and they assign Last of the Mohicans , it's no good saying you didn't finish it and write a paper about it just because it sucked (Mark Twain's opinion, not mine re Cooper.) We've seen our midterm grade and it needs improvement. In the best of all possible worlds, we'll inspire Dan to do better himself.
I'm going to do something about Chapter 9. Don't know when, but I'm going to do it.
Duffy, it's not to soon for you to write up Ch 16. I know you don't agree with it 100% and that's actually a good thing. Also has CS3 been a deal changer? You can kick off a great discussion. Dan's opinion here is controversial and probably designed to be controversial. Let's have some fun with it.
I'm a newcomer to the forum, and I've read many (but not all) of the threads. I've been quite impressed by effort that's gone into the chapter reviews and the originality of some of them. But there doesn't seem to be a lot of discussion/debate over the techniques presented. For some yes, for many no.
I guess it's not particularly important whether Dan thinks this a problem, but whether the reading group thinks this a problem.
I made up the grade to describe what he said (after the aforementioned wine tasting, so don't expect me to quote verbatim.) He was comparing the PP5E group with the LAB book. We blew his socks off in that group. Here we are struggling to reinvigorate the PP5E book. The LAB group had so many participants who became proficient. Participants did so many great retouches.
The business with the master vs individual curves frustrated him and I'm not sure how much he has followed since. I'm going to have a chance to pin him down (maybe with no wine tasting going on.) I'll try to exchange constructive feedback (though he does read a lot of this stuff.)
I reread Chapter 15 because it sounds like something that could be useful and I even wondered if I might volunteer to do the write-up on it, but I'm still not "getting it". I know there isn't a write-up of it yet and that would be the best place to ask questions about it, but I don't know when that will be and I have a question now.
Here goes the question. Let me know if there's a better place to put this question. I understand how a false profile can be used to change the luminosity gamma or the color space. What I don't understand is how messing with the luminosity gamma is any different than a luminosity curve? I'm one of those people who only really "gets" something when I know how it works under the hood. I intimately understand what transformation is happening with a curve. I imagine that a gamma adjustment targetted at luminosity is just a particular shaped curve. If that's the case, then I don't understand why false gamma profiles are any better than curves. If there's more to the luminosity part of a false gamma than a curve, then I'd love to understand how it works.
I do understand how certain color situations might benefit from assignment into a false color profile to give you a larger (perhaps even fictitious) color space to work in, manipulation in the larger color space to accomplish what you want while you have lots of room, then conversion back to a real color space. But, I don't understand that for luminosity. Why can't you just reproduce any gamma luminosity changes from a false profile with the appropriate set of luminosity-targetted curves. I could never find where Dan actually explained why they were better than curves. And, since I feel like I understand curves really well, I'm stuck as to why I should use false profiles for fixing overly dark shots.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
This is really a great question. The answer, I think is that it's particularly easy to create a large library of these false profiles and use them when you need them. You could do the same with curves, but it's more work to make them. With the variant color spaces, you can just plug in one number and you are done. I would be interested to see the resulting L curve, but it's not vital.
Dan loves to tell the story of a newspaper guy he knows who uses ONLY false profiles for image correction. It becomes a matter of knowing your library, not much different from people who like the filters, which can also be represented as curves (RGB, not LAB).
OK, I can accept that and that makes sense. It's a different tool that can be used to solve a problem. Depending upon the problem and your efficiency with other tools, it will be an individual choice whether this tool works better for you than curves.
But, other than the efficiency or convenience factor, it seems that a different gamma is just a pre-defined curve (with less control over the shape). This Wikipedia artile on gamma was useful for me and shows some gamma curves.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
I had exactly the same reaction when I read about the overlay layers in the portrait technique, and it drove me to reverse engineer the related curves. So I basically agree with you. These days I'm finding I use both the overlay move from the portrait technique and the false profiles more and more. They really are big time savers once you get used to them. Get them on a layer and move the opacity slider as in the MFM technique. Just a few clicks, compared to writing a new curve or set of curves.
Keep in mind that LAB curves are not all powerful the way that RGB curves are. You cannot write one curve which causes a color shift in the highlights only and a different one in the shadows. Some false profiles can do this, but you have to be a wizard at photoshop's colorspace definitions (not me.)
Got it. I guess I'm more interested in techniques that let me do new things I didn't know how to do before (like enhance colors in LAB or use a color channel for a mask or HIRALOAM sharpening or a red channel blend to enhance a sky) than I am in a faster way to do something I already know how.
It's not that I don't value speed, but it's not near as important to me as capability and I'm not a pro so time isn't money in my photography. Because I do so much event photography with lots of photos of a large number of people (often sports teams), my time on PP is dominated by my workflow to process a 300 image shoot, not by the time to PP a few of the extraordinary shots.
For your example of color correcting shadows and highlights differently, I have no problem just doing that with RGB curves or using two curve layers in LAB and targetting them to a luminosity range either with blend-if settings or a luminosity mask or a manually created mask.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
I mean, to each is own, but is there a difference in results?
And if not, what would be the benefits doing it via - rather obscure and uncontrollable - false profiles compared to controlled to a T set of curves (except habit/sentimental purposes/feeling like dining with Gods/etc.)?
I can see some advantage to using false color profiles. Of particular interest is the wide gamut CMYK which lets you get to CMYK adjustments, channels, etc... without clipping colors like regular CMYK. But that's just as a temporary color space for the purposes of doing some CMYK adjustment before you go back to RGB.
I have been wondering about "dining with the Gods". I guess I wasn't invited
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
1) Non-destructive editing. Using a false profile changes appearance without changing brightness levels.
2) Make subsequent corrections easier. An image that's too light or too dark is already at an extreme end of what can be adjusted using curves. One goal of applying a false profile is to bring the image into a range in which subsequent corrections will be much easier.
3) Time savings. If subsequent corrections are easier they'll also be faster.
False profiles only seem rather obscure and uncontrollable until one acquires a feel for what they do. I felt the same way about using curves at first. Eventually I became comfortable using curves. The same thing happened with false profiles.
I think of false profiles as a jump start to using curves. As such, using a library of false profiles is for me a better workflow than using a library of saved curves adjustments.
Regards,
Michael
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
Great post! Thanks! So glad to have you on board. (Want to write a chapter summary?!)
I'm glad to be here.
Yes, I'd very much like to write a chapter summary.
I've finished reading the book, but I haven't finished reading all the other posts in this thread nor have I caught up with the posts in the reading group chapters. I'm real sorry to be joining the party so late. I wish I had discovered this site long ago!
I'm even more disappointed at missing the LAB reading group. I've read that book several times, it's one of my favorite books! Is it too late to contribute to that thread as well?
Obviously I'm new here, so I'm not sure what the gameplan is. How would you like me to proceed?
Regards,
Michael
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
I was thinking about the "Art of the False Profile" chapter which is the topic of the moment and the beginning of a pretty interesting cycle of chapters. How about it?
Our "marketing department" wants to know how you found us.
Not at all.
[PM sent]
"Art of the False Profile" it is! A topic near and dear to my heart.
I was told about dgrin by someone I met at Photoshop World in Boston while we were talking about Dan Margulis and the amazing presentations he did.
I noticed there doesn't appear to be a chapter 10 entry for the LAB group, maybe I could take a stab at that one as well? Was it skipped for a particular reason?
If I'm asking questions that've already been covered please note that I've only just started digging into these threads.
Regards,
Michael
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
Good points. I'm fine with the fact that some people will reach the conclusion that a false profile is a faster tool for them. As of this moment, I haven't reached that conclusion for myself so I thought I'd give you my curves answer to your three points.
Non-destructive editing. When I start work on an image that is seriously off in luminosity, the very first thing I do is apply a curve adjustment layer to get the image in the general ballpark for editing. I often use a different curve layer for getting the image in the right brightness ballpark than I do for tweaking/optimizing it. I think this is done for the same goal as a gamma adjustment with a false profile and is also non-destructive.
Make subsequent adjustments easier. If I've already gotten the image in the right ballpark with the first curve layer, then I think we're even here.
Time Savings. This is a personal experience thing. Whatever you believe for your own skills and workflow is what matters to you. Once the image is in the right ballpark, I don't think it matters whether that happened with a false gamma or a curve adjustment layer, so the potential time savings is in the initial gamma adjustment or initial curve adjustment. That will simply depend upon which tool a given PS user knows how to use more effectively. False gammas might be simpler to make a library of because they are just a number, but they are also less flexible in that they have no flexibility in shape.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
"It is a magical time. I am reluctant to leave. Yet the shooting becomes more difficult, the path back grows black as it is without this last light. I don't do it anymore unless my husband is with me, as I am still afraid of the dark, smile.
This was truly last light, my legs were tired, my husband could no longer read and was anxious to leave, but the magic and I, we lingered........"
Ginger Jones
But it's not the same degree of non-destructive. An adjustment layer is non-destructive only in the sense that the changes made can be reversed. You are still changing brightness levels with your initial curve and any curve adjustment comes with a price. Eventually a curves adjustment will run out of room. In other words, curving has a potential to clip an image where false profiling does not.
A false profile move makes absolutely no changes to brightness levels, it changes their definition or what a given level means, and consequently does not carry the cost of a curves adjustment. This is truely non-destructive editing!
Not at all. Adjusting gamma is not the only objective of a false profile move.
Consider this; If an RGB image comes to me and it's too dark, it's also going to be too gray. The first thing I do is Edit>Assign Profile and choose a profile which will lower the gamma and boost the colors. This will bring the image into a range where subsequent corrections will be much easier, if they're needed at all.
In an 8bit RGB file, there are only 256 brightness levels per pixel per channel. This holds true whether one is working in sRGB, Adobe RGB, or Wide Gamut RGB. I am using these three profiles here because they all have the same 2.2 gamma.
Here's an analogy:
I have 256 dominoes; half inch thick, one inch wide, two inches tall.
If I pile them front to back I get a stack 128 inches high. This is my sRGB stack.
If I pile them side to side I get a stack 256 inches high. This is my Adobe RGB stack.
If I pile them top to bottom I get a stack 512 inches high. This is my Wide Gamut RGB stack.
Note that with all this stacking, I have not lost a single domino. This is truely non-destructive editing.
This is what Assign Profile does. It changes the appearance of the numbers while leaving the numbers unchanged.
That image which is too dark, being also too gray, comes to me from a digital camers that assigns sRGB. I assign a false profile, in this case I would use a custom Wide Gamut RGB with let's say a 1.4 gamma. With one move I have not only lightened the image, but I have also boosted the colors from their initial gray appearance. Again, all without changing a single brightness level.
In addition, any and all curves adjustments made in Wide Gamut RGB have much, much, greater impact on the image because the scale of the brightness levels has been widened.
Again, a change in gamma is not my only goal. If that were the case it would be easier to use the Image>Adjustment>Exposure command.
By changing the scale of the brightness levels, going from a relatively narrow gamut profile to a wide gamut profile, I have far more leverage to engineer changes into my image. This is not to say one should always use a wide gamut working space. That would be like using a sledge hammer to drive in a thumbtack. At some point in the workflow I will eventually convert to some other profile either for output or further editing. But if the image requires a drastic first move, false profiling is a far more effective first step than a curve adjustment layer.
Note: I used the term "convert" just now very specifically. Assigning a profile and converting to a profile are very different moves. For the benefit of others who may read this post I will explain the difference...
There are two commands in PS under the Edit menu; Assign Profile & Convert to Profile. They both allow you to choose an ICC profile from a list of the profiles installed on your machine.
Assign Profile changes the appearance of an image without changing the numbers, by numbers I mean channel structure (RGB, LAB, CMYK) and brightness levels. A brightness level of 255r in sRGB is not the same color red as 255r in ProPhoto RGB.
Convert to Profile changes the numbers while attempting to not change the appearance, because out of gamut colors will change based on your choice of rendering intent.
To put it another way; Assign Profile tells PS what a given "red" means, Convert to Profile tells PS what numbers to use to get that "red" if it is possible to get that "red" using a specific channel structure.
In closing, I'm not advocating using false profiles instead of curves. But there is no question that the "Art of the False Profile" is a very valuable technique. Even if the above discussion has not convinced you of their worth, I would still advise adding this technique to your skill set simply because it gives you more choices in your workflow, and more choices means greater flexibility.
You've got me using up all my stuff for the chapter summary. D'oh!
Regards,
Michael
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
Hi Dan!
Thanks again for introducing me to dgrin.
As you can see, I'm havin' a great time!
Michael
Graphics23 - Design, Illustration, Restoration & Retouching
What's a Pirate's favorite color mode? Arrrr, G, B!
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
Thank you very much! I'll be looking for your write up. Hopefully with your explanations I'll be able to actually start using this powerful stuff..
(Great analogy with domino chips!)