ACR 4.4 and Clarity

[dwdallam]

Local contrast enhancement is not sharpening - the confusion is understandable, as LCE has often been made with sharpening tools (used outside the manual's recommendations), but it's as far as using a screwdriver to drive screws on one hand, and to open beer cans on the other hand.
As said, it would be a pity not to make LCE, which is only a tone correction, in a raw processor.

And talking about sharpening, did you read this article by Bruce Fraser? It explains well (and I've heard it did define) the input/output sharpening concepts of which Michael wrote about.

The sharpening tool in Lightroom is very powerful.

If I may add a little something : the input sharpening is remarkable indeed. For creative or output sharpening in LR, we all put a great hope in LR2.0!
[a href=\"index.php?act=findpost&pid=183495\"][{POST_SNAPBACK}][/a]

Nope, I didn't know that article existed. However, it is dated 2003, and I'm wondering if that is still "the way" to do it. Incidentally, that's pretty much how I was doing it from   intuition, but with a variation on the first step--and I'm not saying I'm doing it correctly. 1. Initially sharpen the entire image if necessary using a very fine tuned sharpening approach that fits the iamge itself--after it's resized. 2. Sharpen specific areas, like the tried and true example, the eyes, if needed. 3. Sharpen the entire image for output/printing if needed. This is just what my raw logical ability told me to do, but like I said, never read any article putting it in those three steps.

[ErikKaffehr]

Hi,

the idea is that capture sharpening would compensate for what is lost in capture. This would compensate for:

1) Lens MTF
2) MTF of the anti aliasing filter
3) MTF of the sampling process

Ideally the capture sharpening would give an ideal image as a start of creative work. To do that correctly you would need to know the 'transfer function' or 'point spread function' of the lens/sensor system involved. this should be done as early in the process as possible.

The next step is content related sharpening. Here you can use unsharp masking on eyelashes and so on.

Finally you would sharpen for output. You would modify edge contrast to compensate for loss of contrast during printing. To put it simple:

1) Try to make an optimal capture.
2) Improve details as needed.
3) Increase edge contrast to compensate the expected loss of edge contrast in printing. This is depending on the size of your print and the technique used.

Best regards

Erik

Actually, Bruce didn't use Camera Raw's sharpening because before CR 4.1, it was a crude & blunt instrument. Which is why Adobe hired Bruce to consult with them about improving it. CR 4.1 embodies pretty much all the principles of PhotoKit Sharpener's Capture Sharpening and is arguably better (although requiring a bit more fiddling to nail the needed results).

But you are also mixing up the don't sharpen in Camera Raw with the old school thought to only sharpen at the end of a process. As other have stated, the ideal method is to use a sharpening workflow with the only end stage sharpening being output sharpening once the image is in its final size & resolution and the method of output is known.

In the grand scheme of things, Thomas Knoll, the primary author of Camera Raw, wants Camera Raw to be the best raw processor out there. As a result, Camera Raw has grown considerably since it's early functions and capability. Rather than fight the concepts and the logic of why things are the way they are, it behooves users to actually learn to be fluent in the use of the varied functionalities so they can take full advantage of what is now possible in Camera Raw.
[a href=\"index.php?act=findpost&pid=183507\"][{POST_SNAPBACK}][/a]

[dwdallam]

Actually, Bruce didn't use Camera Raw's sharpening because before CR 4.1, it was a crude & blunt instrument. Which is why Adobe hired Bruce to consult with them about improving it. CR 4.1 embodies pretty much all the principles of PhotoKit Sharpener's Capture Sharpening and is arguably better (although requiring a bit more fiddling to nail the needed results).

But you are also mixing up the don't sharpen in Camera Raw with the old school thought to only sharpen at the end of a process. As other have stated, the ideal method is to use a sharpening workflow with the only end stage sharpening being output sharpening once the image is in its final size & resolution and the method of output is known.

In the grand scheme of things, Thomas Knoll, the primary author of Camera Raw, wants Camera Raw to be the best raw processor out there. As a result, Camera Raw has grown considerably since it's early functions and capability. Rather than fight the concepts and the logic of why things are the way they are, it behooves users to actually learn to be fluent in the use of the varied functionalities so they can take full advantage of what is now possible in Camera Raw.
[a href=\"index.php?act=findpost&pid=183507\"][{POST_SNAPBACK}][/a]

You said:
1. ". . . the old school thought to only sharpen at the end of a process."
2. " . . . the ideal method is to use a sharpening workflow with the only end stage sharpening being output sharpening once the image is in its final size & resolution and the method of output is known.

Can you explain the difference between those two comments please?

[dwdallam]

OK I have a question now that comes up since we've hit all over the park on this one:

What is the function of clarity?

For example, is it a (1) sharpening tool, (2) a contrast tool, or (3) alien technology that escapes definition, (4) none of the above?

I think we got off on sharpening, which is also an important topic, maybe even more so than Clarity?

However, uh, Clarity?

[Schewe]

Can you explain the difference between those two comments please?
[{POST_SNAPBACK}][/a]

A sharpening workflow (ala Fraser, even if it's based on an article from 2003) means you do capture sharpening in the beginning after major tone and color adjustments, you do creative sharpening for effect during the imaging process and the final, output sharpening is only done at the end of the process once you know the final output resolution and output type.

Read Bruce's article to better understand a sharpening workflow and read this article for more about Camera Raw 4.1+ [a href=\"http://photoshopnews.com/2007/05/31/about-camera-raw-41/]About Camera Raw 4.1[/url]. Alternatively you can wait for the Camera Raw video tutorial from Michael and I or go out and get Real World Camera Raw for Photoshop CS3 now. Available from Amazon here Real World Camera Raw with Adobe Photoshop CS3

[digitaldog]

What is the function of clarity?

To make your images look better (or not). Seriously, its a creative rendering control. Its roots date to techniques to produce a similar Midtone contrast boost in Photoshop.

At one time during the beta, it was called Punch, a more photocentric yet also violent sounding description of the rendering.

[01af]

Apparently some of the confusion in this thread---which originally wasn't about sharpening at all---is coming from the misconception that unsharp masking (USM for short) was the same as sharpening. It's not.

Unsharp masking can be used to the effect of sharpening. It can be used to other effects as well ... for example, to the effect of local contrast enhancement. And that's what "Clarity" in Adobe Camera Raw does. It all depends on the parameters of the unsharp-masking procedure. Large amount and small radius = sharpening; small amount and large radius = local contrast enhancement. The latter is also known as the hiraloam technique---high radius, low amount ... a phrase coined by Dan Margulis, as far as I know.

By the way, this doesn't necessarily mean "Clarity" in ACR was exactly hiraloam USM and nothing else ... instead, it also might be some clever variation thereof. I don't know. Maybe someone else knows ...?

-- Olaf

[Schewe]

The latter is also known as the hiraloam technique---high radius, low amount ... a phrase coined by Dan Margulis, as far as I know.
[{POST_SNAPBACK}][/a]

The idea of using a small amount/large radius in reduced amounts in USM has been around since Biedny & Monroy's Photoshop book back in the early 1990's.

The phrase hiraloam may have been made up by Margulis but it ain't his technique...Michael Reichmann and Mac Holbert of Nash Editions have been advocating it for several years. Michael called it [a href=\"http://www.luminous-landscape.com/tutorials/contrast-enhancement.shtml]Local Contrast Anhancement[/url] and Mac called it Mid-Tone Contrast Adjustment. Both of which are far more useful names (and techniques).

Thomas' implementation is a hybrid of the two (Mike & Mac) and is based upon a 100 radius blur and using an overlay blend while rolling off the highlights and shadows to concentrate the effect in the mid-tones.

The fact that Margulis has written about it is incidental (and the name: "hiraloam" is goofy/stupid).

[eronald]

In the grand scheme of things, Thomas Knoll, the primary author of Camera Raw, wants Camera Raw to be the best raw processor out there. As a result, Camera Raw has grown considerably since it's early functions and capability. Rather than fight the concepts and the logic of why things are the way they are, it behooves users to actually learn to be fluent in the use of the varied functionalities so they can take full advantage of what is now possible in Camera Raw.
[a href=\"index.php?act=findpost&pid=183507\"][{POST_SNAPBACK}][/a]

Yes, and Adobe also seems to fight tooth and nail against ANY possibility of external camera calibration (input matrices) being accessible to users of ACR or Lightroom. Exactly why I cannot input a matrix into either of these pieces of software is something I fail to understand - or does Adobe seems to subscribe to the "Nanny knows best" school of thought.

When a piece of software does not offer crucial functionality - calibrated color in this case - I turn to the alternatives.

Edmund

[digitaldog]

Yes, and Adobe also seems to fight tooth and nail against ANY possibility of external camera calibration (input matrices) being accessible to users of ACR or Lightroom. Exactly why I cannot input a matrix into either of these pieces of software is something I fail to understand - or does Adobe seems to subscribe to the "Nanny knows best" school of thought.

Simple question before we go too OT. 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?

[Schewe]

Simple question before we go too OT. 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?
[a href=\"index.php?act=findpost&pid=183590\"][{POST_SNAPBACK}][/a]

Naw, he's pissed cause he can't sell camera profiles to people using Camera Raw/Lightroom...

:~)

[interpolator]

A sharpening workflow (ala Fraser, even if it's based on an article from 2003) means you do capture sharpening in the beginning after major tone and color adjustments, you do creative sharpening for effect during the imaging process and the final, output sharpening is only done at the end of the process once you know the final output resolution and output type.

Real World Camera Raw with Adobe Photoshop CS3
[a href=\"index.php?act=findpost&pid=183560\"][{POST_SNAPBACK}][/a]

Should noise reduction in ACR 4.4 be performed befor or after input sharpening.

[Mark D Segal]

The idea of using a small amount/large radius in reduced amounts in USM has been around since Biedny & Monroy's Photoshop book back in the early 1990's.

The phrase hiraloam may have been made up by Margulis but it ain't his technique...Michael Reichmann and Mac Holbert of Nash Editions have been advocating it for several years. Michael called it Local Contrast Anhancement and Mac called it Mid-Tone Contrast Adjustment. Both of which are far more useful names (and techniques).

Thomas' implementation is a hybrid of the two (Mike & Mac) and is based upon a 100 radius blur and using an overlay blend while rolling off the highlights and shadows to concentrate the effect in the mid-tones.

The fact that Margulis has written about it is incidental (and the name: "hiraloam" is goofy/stupid).
[{POST_SNAPBACK}][/a]

Jeff,

Indeed.

But my take on the history of the techniques is a bit different (unless I'm missing a piece - not impossible). From what I've read and have been doing, there would appear to be two basic approaches: one uses USM, the other does not. The one using USM, according to Michael's article which you linked is credited there to Thomas, but looks like an earlier version of Thomas's approach because it doesn't use the overlay mode - from what I mention next, that bit would appear to come from Mac Holbert's design. Now, John Paul Caponigro is also a player in this field, but the technique he describes here: [a href=\"http://www.johnpaulcaponigro.com/downloads/technique/documents/HighPassContrast.pdf]JPC High Pass[/url] doesn't use USM - it uses a stamp layer in overlay mode at opacity 20% with the shadows and highlights masked using Blend If, and runs a High Pass filter at radius 50 & Desaturates. J-P credits this approach to Mac Holbert. Each of them creates a slightly different effect. It's handy to make Actions of both. I find myself making more use of the Action from J-P's workflow, because it has a lot of ex-post flex to adjust the effect to taste. But both work fine and are very useful. The one caution I would recommend to users of either approach is that when it comes to output sharpening of images with that effect implemented, one does need a bit of extra care to not over-sharpen - the image can become too "brittle".

[eronald]

Naw, he's pissed cause he can't sell camera profiles to people using Camera Raw/Lightroom...

:~)
[a href=\"index.php?act=findpost&pid=183593\"][{POST_SNAPBACK}][/a]

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.

This is not a software difficulty, it's a conscious decision by Adobe not to allow the user to input a matrix. And also ***not honoring whatever matrix is already there in the Raw file's tags***.


It's not a technical issue, all the hooks are already there. ACR and LR use a matrix, but it's always their own (interpolated) matrix -except maybe when reading DNG. The rest of the world would like to be able to mandate the use of a different matrix.

And yes, incidentally, this closed door policy does lock out third party camera calibration tools, and color consultants.

This issue keeps cropping up again and again in the small world of camera color management. I raised it with Thomas at Photokina, asking for profiles - no luck. I raised it with Manish at CIC in Albuquerque, asking for matrix input on the behalf of Xrite; it was again raised at the  ICC meeting in Munich last month. It will probably again surface at the meeting in June if the camera guys decide to write their preferred matrix into the TIFF/EP files. It's a political problem with Adobe intent on keeping their system locked down and the rest of the world wanting it opened a bit.

There are increasingly color management technologies that could be brought to bear on Raw color fidelity in much the same way as spectros and printer linearisation can be brought to bear on print color. Various camera calibration technologies have been published or demonstrated, the latest of which is Dietmar Wueller's spectral response measurement device which was described to the ICC digital photo workgroup in Munich. But nobody can get these things to the user employing ACR and Lightroom because of the lockdown.

Oh, and by the way - hint, hint, I'd be delighted to talk in person to anyone ...

Edmund

[dwdallam]

Look what I started. Shame on me

[Mark D Segal]

Look what I started. Shame on me
[a href=\"index.php?act=findpost&pid=183618\"][{POST_SNAPBACK}][/a]

Why shame on you? What did you start? You asked a simple, legitimate question and got a slew of answers and some piffy debate. What else is new?  

[KeithR]

Yes, and Adobe also seems to fight tooth and nail against ANY possibility of external camera calibration (input matrices) being accessible to users of ACR or Lightroom. Exactly why I cannot input a matrix into either of these pieces of software is something I fail to understand - or does Adobe seems to subscribe to the "Nanny knows best" school of thought.

When a piece of software does not offer crucial functionality - calibrated color in this case - I turn to the alternatives.

Edmund
[a href=\"index.php?act=findpost&pid=183576\"][{POST_SNAPBACK}][/a]

I thought that this was what the Calibrate Tab in ACR/LR were for. Utilizing something like the Thomas Fors or Rags Gardner script you can, so to speak, calibrate you camera to ACR/LR.

From the Rags Gardner site:

"...is a script for Adobe Photoshop that can be used to calibrate a camera's color response in an Adobe Camera Raw / Lightroom workflow. It performs the calibration based on the standardized colors of the GretagMacBeth ColorChecker target. It is based on Thomas Fors' ACR Calibrator script, which was the first automatic implementation of a calibration procedure suggested by Bruce Fraser."

Since each and every camera can be different, Adobe has, by putting in the Calibrate Tab, a way for calibrating each camera. If they were fighting tooth and nail against it, they simply could have left it out. I'm sure that Mr. Fraser had a strong hand in making sure that this was available in ACR/LR and is explained in all three versions of his-and Jeff's book. The scripts just automate the process.

[KeithR]

Should noise reduction in ACR 4.4 be performed befor or after input sharpening.
[a href=\"index.php?act=findpost&pid=183600\"][{POST_SNAPBACK}][/a]

ACR 4.4(and the LR version) have been pulled and is being recommended that you should roll back to v 4.3x. Noise reduction should be done before sharpening if possible. Why would you want to sharpen noise?

[eronald]

Keith,
 To effect the Raw conversion, ACR determines primaries that will be associated with each of the Bayer filtered sensor cells. 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

I thought that this was what the Calibrate Tab in ACR/LR were for. Utilizing something like the Thomas Fors or Rags Gardner script you can, so to speak, calibrate you camera to ACR/LR.

From the Rags Gardner site:

"...is a script for Adobe Photoshop that can be used to calibrate a camera's color response in an Adobe Camera Raw / Lightroom workflow. It performs the calibration based on the standardized colors of the GretagMacBeth ColorChecker target. It is based on Thomas Fors' ACR Calibrator script, which was the first automatic implementation of a calibration procedure suggested by Bruce Fraser."

Since each and every camera can be different, Adobe has, by putting in the Calibrate Tab, a way for calibrating each camera. If they were fighting tooth and nail against it, they simply could have left it out. I'm sure that Mr. Fraser had a strong hand in making sure that this was available in ACR/LR and is explained in all three versions of his-and Jeff's book. The scripts just automate the process.
[a href=\"index.php?act=findpost&pid=183628\"][{POST_SNAPBACK}][/a]

[dwdallam]

Why shame on you? What did you start? You asked a simple, legitimate question and got a slew of answers and some piffy debate. What else is new? 
[a href=\"index.php?act=findpost&pid=183622\"][{POST_SNAPBACK}][/a]

Hahaha, true. I was really joking when I said that. It seems this was the topic of the week, and someone just simply needed to post it.

It seems that many of us need to go over our ACR skills and bring them up to date also.