ACR 4.4 and Clarity

[Mark D Segal]

It seems that many of us need to go over our ACR skills and bring them up to date also.
[{POST_SNAPBACK}][/a]

Doug, yes, the changes in ACR from 3.7 to 4.x are so fundamental and so powerful that it really behoves one to commit real time to mastering it - pays handsome dividends. There are two approaches - (1) experiment with all of it, extensively, and (2) read Fraser/Schewe "Real World Adobe Camera Raw with Adobe Photoshop CS3". It is an excellent resource - you can see my review of it on-line at Amazon.com [a href=\"http://www.amazon.com/review/product/0321518675/ref=sr_1_1_cm_cr_acr_txt?%5Fencoding=UTF8&showViewpoints=1]Real World Camera Raw - Reviews[/url], where the book is well-priced. And no - this is not a sales plug - I have no reason to - it's what I really think.

Mark

[digitaldog]

No, I'm pissed because if my lighting happens to be spiky, or my camera happens to be a strange sample, or, or, or, I cannot just compute a matrix and tell LR or ACR to honor that and be done.

And you've not answered a simple question about your ability or inability to render an image as you desire. I'm not sure what the spikyness of an illuminant has to do with any of this and doesn't address the fundamental and simply question.

[digitaldog]

External calibration means supplying such a matrix numerically directly to the software. Such a user-supplied matrix can be informed by sophisticated knowledge of the camera and the lighting.  The Fors you refer to scripts just "plays around" with the adjustments available in the software itself, when given a colorchecker image.

Edmund
[a href=\"index.php?act=findpost&pid=183636\"][{POST_SNAPBACK}][/a]

So again, with or without the script, the ONLY way you can render an image to a color appearance you so desire is to use a package that allows you to use a custom camera profile? If so, some of us want to examine your Raw files and what you're attempting to render.

You may have had conversations with Thomas, but it appears like so many before you, your desire for a function in an Adobe product was either ill defined, is a solution is search of a problem or, you don't have the problem you think you do.

Lots of Adobe users ask for feature requests, so few know how to do so correctly and based on the very smart people at Adobe who decide how to design a product. You need to demonstrate scientifically (or supply the math, good luck) that without the ability to build and use custom ICC profiles in ACR, the product is unable to produce the results most users require. The group grumbling about this are often those who sell profiles, or software to build profiles. That's not at all been a convincing group so far, probably because they have totally failed to illustrate other than a want, why this is necessary. Thomas didn't leave out ICC profiles to piss off companies trying to sell ineffective camera profiling software, it did it because none of these people have illustrated why its at all useful to anyone but those companies.

[Panopeeper]

You need to demonstrate scientifically (or supply the math, good luck) that without the ability to build and use custom ICC profiles in ACR, the product is unable to produce the results most users require

I don't think even Adobe would come up with such a nonsense. If something is not working as expected, then I as the customer do not need to prove, that it could not work whatever I triad - the software manufacturer has to show, that it *can* work and *how* it can be done.

As it is now, this has not happened. Adobe has not provided a straightforward, exact way to achieve the best color conversion. The profiles created with the Forst and Rags scripts are based on three (or four?) squares of the Gretag checker and the result is still not as good as for example Canon's DPP is creating.

The new DNG specification adds some color conversion related tags, I wonder what these are addressing.

[Ray]

I was a great fan of RSP before Adobe bought them out. I still use RSP because their 'detail' and 'vibrancy' controls are capable of a much more dramatic effect than any of these new controls in ACR, such as 'clarity', 'detail' and 'vibrancy', which I presume are concepts that were borrowed from RSP.

I sort of feel that Adobe should be a bit bolder with its control options in ACR. On the other hand, I can see a problem here. When controls can have a dramatic effect, the options for producing really bad results also increase.

[digitaldog]

I don't think even Adobe would come up with such a nonsense. If something is not working as expected, then I as the customer do not need to prove, that it could not work whatever I triad - the software manufacturer has to show, that it *can* work and *how* it can be done.

Ah, you've leaped past the point of my questions about not working as expected. That's not the complaint as yet. Who has yet proven the basic premise that the product isn't working?

As it is now, this has not happened. Adobe has not provided a straightforward, exact way to achieve the best color conversion. The profiles created with the Forst and Rags scripts are based on three (or four?) squares of the Gretag checker and the result is still not as good as for example Canon's DPP is creating.

Some may not see exactly what's going on here. The process aids (some would say ensures) that IF you capture the Macbeth color checker, you can get out the back end (output referred) the correct RGB values in a fixed output referred working space (ProPhoto RGB).

This is useful for all photographers shooting Macbeth color checkers. If you're not, then its questionable if not using OTHER rendering controls to produce a desired color appearance isn't were a user should move. It goes back to the simple question no one wants to answer.

Note, you don't need the scripts, nor did Bruce Fraser who came up with this process to get the correct RGB values. You can, as Bruce originally did, move all the calibrate sliders manually to get the values he was aiming for. Note that there's nothing that says you can't alter non-Calibrate sliders to get the values and save a preset. But the values he was aiming for alter the rendering to produce output referred RGB values of this target.

One can now say "I've calibrated my chip" or something along those lines and they would be correct. You've produced the same Macbeth values in ProPhoto as Thomas did with his camera sample(s). It doesn't guarantee anything else you photograph will automatically be the desired color for any other condition (scene and scene gamut, dynamic range or illuminant).

Those who beg for custom profiles never tell us that using such a profile guarantees every capture is now "correct" either subjectively or colorimetrically as was done using the Fraser technique on this single capture. Or if using a calibration process to tweak your chip to what the two profiles were designed for is a less effective use of a CMS. So we end up with the original and simple question: using or not using such scripts, are you unable to produce a desired rendering?

In the end, we hear people say "we want profiles" but without an ounce of proof why or how it would be any better than the arguably different system we have in ACR/LR.

[KeithR]

The profiles created with the Forst and Rags scripts are based on three (or four?) squares of the Gretag checker ....[a href=\"index.php?act=findpost&pid=183742\"][{POST_SNAPBACK}][/a]

From the Gardner site:
ACR Calibrator L adds a number of enhancements, the most important of which is that it performs a single 6-dimensional optimization on the calibration sliders, using all patches, instead of alternating the calibration of red, green and blue. In some cases, this improves the stability of the algorithm.

[Mark D Segal]

I was a great fan of RSP before Adobe bought them out. I still use RSP because their 'detail' and 'vibrancy' controls are capable of a much more dramatic effect than any of these new controls in ACR, such as 'clarity', 'detail' and 'vibrancy', which I presume are concepts that were borrowed from RSP.

I sort of feel that Adobe should be a bit bolder with its control options in ACR. On the other hand, I can see a problem here. When controls can have a dramatic effect, the options for producing really bad results also increase.
[a href=\"index.php?act=findpost&pid=183746\"][{POST_SNAPBACK}][/a]

Ray, I tried RSP some time before ACR 4.x appeared with these controls. RSP's Vibrancy was arguably more aggressive, but I think the Adobe implementation is more refined and works well. I always find that I can't push Vibrancy and Clarity in ACR too far before the image begins to look "pumped" and artificial. That tells me these tools have all *I* need. YMMV.

Mark

[Mark D Segal]

I don't think even Adobe would come up with such a nonsense. If something is not working as expected, then I as the customer do not need to prove, that it could not work whatever I triad - the software manufacturer has to show, that it *can* work and *how* it can be done.

As it is now, this has not happened. Adobe has not provided a straightforward, exact way to achieve the best color conversion. The profiles created with the Forst and Rags scripts are based on three (or four?) squares of the Gretag checker and the result is still not as good as for example Canon's DPP is creating.

The new DNG specification adds some color conversion related tags, I wonder what these are addressing.
[a href=\"index.php?act=findpost&pid=183742\"][{POST_SNAPBACK}][/a]

Pano, I'm perplexed with this post. What do you mean when you say "something is not working as expected"? What was expected and what isn't working properly in relation to that?

As for the obligations of the software manufacturer - look - it's a commercial enterprise whose shareholders reward or punish its management based on sales and profits, so the management has to produce something the market accepts or they're out. The fact that Photoshop has grown through 10 versions over almost a couple of decades, Adobe stock is doing well and the program (including ACR) has become the world's premier image editing application speaks for itself. Not to say it doesn't have issues, but everything does, and just look at what it CAN do - in capable hands. Adobe doesn't have a tradition of publishing the last word on what Photoshop can do or how (there would be no such thing as a last word with this program anyhow - it's too deep). Rather, a whole industry has developed around the program with numerous very well qualified and experienced authors doing that for them - and for us. This is all to the good. It's win-win-win for Adobe, the Photoshop education industry and us users.

Now you go on to say that "Adobe has not provided a straightforward, exact way to achieve the best color conversion". I don't understand the operational significance of that statement. What do you mean by "straightforward", "exact" and "best color conversion". Unless you explain these things in some tangible, objective manner the meaning of this statement is impenetrable.

Then you bring in the Fors and Rags scripts. I don't know where those fit this discussion. Do you know how Thomas Knoll profiles the cameras and creates the recipes for the display representation and the rendering? Have you seen information suggesting that he uses the Fors or Rags methodology, exclusively, for doing this? I'm curious. Indeed, in all of this discussion, it would be good to see some concrete explanation from any one who knows - in what ways do the Fors/Rags approach differ from the Thomas Knoll approach and whose is likely to deliver more "reliable" conversions (with some objective definition of what that means). I've never seen such an explanation, yet this is what's needed to make much sense of this discussion.

You say the results using ACR are not as "good" as those from Canon DPP. What does "good" mean here - because as it stands, this again is a statement devoid of operational significance - the term "good" is subjective, an adjective, and not an objective descriptor of a desired result.

[Mark D Segal]

................... The process aids (some would say ensures) that IF you capture the Macbeth color checker, you can get out the back end (output referred) the correct RGB values in a fixed output referred working space (ProPhoto RGB).

........................... It goes back to the simple question no one wants to answer.

......................
[a href=\"index.php?act=findpost&pid=183763\"][{POST_SNAPBACK}][/a]

Well, indeed. The devil is in the details of the "IF". I have a very strong sense that unless the GMCC is shot under very controlled conditions it's a pointless exercise.

Now don't say "no-one" wants to answer your question. I'll repeat your question here for convenience, and I'll acknowledge at once that I know it was not addressed to me, but I'll answer it anyhow:

<<Are you saying that the current sets of rendering controls do not allow you to produce a desired rendering from the Raw? On some or all images?>>

I've probably pushed about a couple of thousand images through the ACR 4.x series since it became available - a considerable variety of stuff. It's an extremely capable piece of software and takes me a very long way to the final product. It lacks three fundamental functions which require trips into Photoshop: no masking (except in the sharpening tab), no output sharpening and no soft-proofing. I can think of further refinements which would add more precision targeting capabilities to its global adjustments, but we do have Photoshop for that. With all the permutations and combinations of adjustments this plug-in allows, my sense is that the primary limitation on outcomes is the user's capability to make it deliver. Sure, there will always be those situations where one wished it could do a little more or a little different of this or that - and I could develop this discussion - but no point - there isn't a piece of software on earth that does everything everyone wants all the time. So let's get real.

[Panopeeper]

What do you mean when you say "something is not working as expected"?

The colors produced by ACR for several cameras (for example for Canons, based on *many* users' complaints) are off. If color fidelity is not among your expectations, then I don't know, what subject you are discussing.

The fact that Photoshop has grown through 10 versions over almost a couple of decades, Adobe stock is doing well and the program (including ACR) has become the world's premier image editing application speaks for itself

Valid arguments among lay people (I mean in software development). I guess you are criticlessly satisfied with Windows as well - Microsoft is a much more successful company, than Adobe.

Now you go on to say that "Adobe has not provided a straightforward, exact way to achieve the best color conversion". I don't understand the operational significance of that statement. What do you mean by "straightforward", "exact" and "best color conversion". Unless you explain these things in some tangible, objective manner the meaning of this statement is impenetrable

I, as a software developer don't expect my customers to search in internet for the information necessary to the full utilization of my products

Then you bring in the Fors and Rags scripts. I don't know where those fit this discussion

I really don't understand what you don't understand. Don't you know what these scripts are doing? Don't you think they are useful? Do you know of other ways of computerized calculation of the calibration parameters?

Do you know how Thomas Knoll profiles the cameras and creates the recipes for the display representation and the rendering?

Who cares? That's a black box, as it should be.

Have you seen information suggesting that he uses the Fors or Rags methodology, exclusively, for doing this?

I seriously doubt it, but again, that is irrelevant for the user.

You say the results using ACR are not as "good" as those from Canon DPP. What does "good" mean here

I think it would be better to ask those users, who prefer DPP to ACR, perhaps they have inclination to discuss such questions. I do prefer ACR in most cases, but color fidelity is less important for me than for many other users. My initial impression was, that the subject is the quality of color reproduction of ACR and the viability of third-party vendors on that area, not the question, what "good color" means.

[dwdallam]

Doug, yes, the changes in ACR from 3.7 to 4.x are so fundamental and so powerful that it really behoves one to commit real time to mastering it - pays handsome dividends. There are two approaches - (1) experiment with all of it, extensively, and (2) read Fraser/Schewe "Real World Adobe Camera Raw with Adobe Photoshop CS3". It is an excellent resource - you can see my review of it on-line at Amazon.com Real World Camera Raw - Reviews, where the book is well-priced. And no - this is not a sales plug - I have no reason to - it's what I really think.

Mark
[a href=\"index.php?act=findpost&pid=183724\"][{POST_SNAPBACK}][/a]

Yes but what version of ACR do they cover? CS3 came with 4.0.

[dwdallam]

CLARITY HI-JACKED!

[digitaldog]

The colors produced by ACR for several cameras (for example for Canons, based on *many* users' complaints) are off.

Off (whatever that means) when and how? As a default or no matter what you try to adjust in the UI?

[Mark D Segal]

Yes but what version of ACR do they cover? CS3 came with 4.0.
[a href=\"index.php?act=findpost&pid=183824\"][{POST_SNAPBACK}][/a]

4.1 - and in the book they say they provide a web-site link where they will post up-dated information for future versions. In terms of program features, 4.1 is still "current".

[Mark D Segal]

The colors produced by ACR for several cameras (for example for Canons, based on *many* users' complaints) are off. If color fidelity is not among your expectations, then I don't know, what subject you are discussing.
Valid arguments among lay people (I mean in software development). I guess you are criticlessly satisfied with Windows as well - Microsoft is a much more successful company, than Adobe.
I, as a software developer don't expect my customers to search in internet for the information necessary to the full utilization of my products
I really don't understand what you don't understand. Don't you know what these scripts are doing? Don't you think they are useful? Do you know of other ways of computerized calculation of the calibration parameters?
Who cares? That's a black box, as it should be.
I seriously doubt it, but again, that is irrelevant for the user.
I think it would be better to ask those users, who prefer DPP to ACR, perhaps they have inclination to discuss such questions. I do prefer ACR in most cases, but color fidelity is less important for me than for many other users. My initial impression was, that the subject is the quality of color reproduction of ACR and the viability of third-party vendors on that area, not the question, what "good color" means.
[a href=\"index.php?act=findpost&pid=183822\"][{POST_SNAPBACK}][/a]

What do you mean by "colors are off" - Is it that I photograph a color checker, open it in ACR and the values aren't the same as those of the patches?

No-one with a brain in their head accepts anything "uncritically". So let's not side-track the discussion with personal innuendo. And by the way, I think Windows XP on the whole is a very good operating system. Does it have issues - sure. Are they train-smashes? No.

Have you developed a piece of software as complex as Photoshop and tried to explain it exhaustively to the tens of millions of users at all experience levels? Do you really think you are asking for something sensible in this context? Does Microsoft bundle with Office a complete and transparent manual for using Excel?

I know what the Fors scripts are doing. Been there long ago, read the whole thing. If the whole process is prepared and used with great care perhaps it's value-added. OTH, high-end professional cameras are probably manufactured to perform quite uniformly from copy to copy, so once Thomas Knoll calibrates one of them for ACR it's probably satisfactory for most intents and purposes.

OK, so "good color" means "color fidelity", "color fidelity" means what, going back to the initial question. I downloaded DPP, tried it once and trashed it. No way I would substitute that for Camera Raw.

You say colour fidelity is not that critical for most of what you do. Same here if by that you mean some standard of scene-referred accuracy. I know it is for some folks. If you're doing commercial photography and your client is a brewery who tells you the colour of the beer in the glass needs to be Pantone XXX, then you'd darn well better produce a glass of beer that's Pantone XXX. And I'll bet a dollar to a dime that many of the folks doing that kind of work are using ACR and Photoshop, or if they're using Phase backs they could well be using Capture for the raw conversion. If they don't get the accuracy they need "straight out of the box", there are plenty of controls for achieving it - and the more experienced they are, the less they'll break-out in a sweat over it.

[Ray]

Ray, I tried RSP some time before ACR 4.x appeared with these controls. RSP's Vibrancy was arguably more aggressive, but I think the Adobe implementation is more refined and works well. I always find that I can't push Vibrancy and Clarity in ACR too far before the image begins to look "pumped" and artificial. That tells me these tools have all *I* need. YMMV.

Mark
[a href=\"index.php?act=findpost&pid=183793\"][{POST_SNAPBACK}][/a]

Point taken, Mark. I should not have included vibrancy in that comment. The vibrancy and saturation controls in ACR are more than sufficient with regard to the range of adjustments possible.

It's the 'clarity' and 'detail' controls that seem a bit tame. There's a characteristic quality that RSP's 'detail extraction' can produce that I find difficult to emulate in ACR.

In other words, ACR's 'clarity' and 'detail' controls, both at maximum adjustment and with sharpening at zero, do not equal RSP's 'detail extraction' at maximum with sharpening also at zero.

The closest I can get (in ACR) to that single adjustment in RSP of 'detail extraction' at maximum, is 'detail' and 'clarity' at 100% plus sharpening at 50% with a radius of 1 pixel.

The following 100% crops show the ACR conversion on the left with detail and clarity at 100%, sharpening zero, compared with the RSP conversion with detail extraction at maximum and sharpening at zero.

[attachment=5713:attachment]

[Peter_DL]

What do you mean by "colors are off" ...
[{POST_SNAPBACK}][/a]

Mark,

While I like your contributions,
and that’s really the only reason why I’m going to comment,
you might wish to have a look at my recent [a href=\"http://luminous-landscape.com/forum/index.php?showtopic=24087]feature request[/url].

Please assume that I understand the science part of ACR calibration well enough,
and that I’m also familiar enough with the given ACR rendering controls such as Clarity to derive a somewhat pleasing rendition.

Peter

--

[Mark D Segal]

Point taken, Mark. I should not have included vibrancy in that comment. The vibrancy and saturation controls in ACR are more than sufficient with regard to the range of adjustments possible.

It's the 'clarity' and 'detail' controls that seem a bit tame. There's a characteristic quality that RSP's 'detail extraction' can produce that I find difficult to emulate in ACR.

In other words, ACR's 'clarity' and 'detail' controls, both at maximum adjustment and with sharpening at zero, do not equal RSP's 'detail extraction' at maximum with sharpening also at zero.

The closest I can get (in ACR) to that single adjustment in RSP of 'detail extraction' at maximum, is 'detail' and 'clarity' at 100% plus sharpening at 50% with a radius of 1 pixel.

The following 100% crops show the ACR conversion on the left with detail and clarity at 100%, sharpening zero, compared with the RSP conversion with detail extraction at maximum and sharpening at zero.

[attachment=5713:attachment]
[a href=\"index.php?act=findpost&pid=183868\"][{POST_SNAPBACK}][/a]

Hi Ray, that's a very interesting pair of images. I think what it shows is that while RSP bundled local contrast enhancement and sharpening in one very effective tool, with ACR you would need to use two tools for roughly equivalent effect. Your left image shows pretty much what is to be expected from ACR - at zero Amount you get zero sharpening regardless of where you set Detail.

Beyond that, I think it becomes a matter of personal workflow preference. I'm actually not very attracted to packing much clarity or sharpening into an image at the raw processing stage, because once you render the image, depending on what else you do (especially the interplay between local contrast enhancement [LCE] and sharpening - which could evolve right through to the soft-proofing stage), you could regret the extent of these adjustments and then you need to go back to the raw (unless in ACR/PS you've imported the raw as a Smart Object) to undo it. I tend to prefer the separation of LCE from sharpening, and apply the former conservatively at the raw stage IF I have no doubt I won't regret it. Then mild capture sharpening in the raw conversion process is fine provided one then doesn't repeat capture sharpening in Photoshop, and provided upon Output Sharpening one doesn't make the image brittle.  

So what I'm discussing here really is how to optimize the efficiency and effectiveness of the workflow between the ACR stage and the PS stage, and I find the separation and fine adjustability of LCE and capture sharpening in ACR to be helpful in this pursuit.

[Mark D Segal]

Mark,

While I like your contributions,
and that’s really the only reason why I’m going to comment,
you might wish to have a look at my recent feature request.

Please assume that I understand the science part of ACR calibration well enough,
and that I’m also familiar enough with the given ACR rendering controls such as Clarity to derive a somewhat pleasing rendition.

Peter

--
[a href=\"index.php?act=findpost&pid=183882\"][{POST_SNAPBACK}][/a]

Peter, thanks for those references - and Tindemans' site is very useful indeed.

I think the merits of combining Raw+JPG processing is a subject that extends well beyond the question from the OP, interesting as it is. I'd have to think more about it, but the fundamental premise and the prospective valued-added would be the first areas on which I'd focus.

I guess I'm saying this because when it comes to the difficult type of images (typically those with much deep shade and extreme highlights) where such novel techniques would likely be most called-upon, I've never been terribly impressed with the JPEG renderings of even the most expensive Canon DSLRs, but of course this is an individual judgment and everyones' MMV.

I have experimented shooting raw and JPEG to see whether I could use the JPEG as some kind of reference image. It turned out to be more of a nuisance than an assist.  I just shoot raw and use the controls in ACR 4.3 and Photoshop to reproduce what motivated me to make the photograph. This of course is something the engineers in Canon can't know. It's the creative side of the image creation process.

I find I can obtain more than a "somewhat pleasing" rendition from ACR. I'd say that is an accurate descriptor up to and including ACR 3.7, but with 4.x enough control is there to go well beyond the "somewhat pleasing", including quite respectable division of control between luminosity and saturation (and yes I'm all too painfully knowledgeable about all the controversy surrounding this subject), significant and powerful tools for improving contrast and image detail in the dark tones and of course there is local contrast enhancement - to revert to Doug Dallam's original question!