Luminous Landscape Forum

Raw & Post Processing, Printing => Adobe Camera Raw Q&A => Topic started by: MOMOP on March 27, 2012, 07:25:52 pm

Title: Lab in ACR
Post by: MOMOP on March 27, 2012, 07:25:52 pm
In the info palette in Photoshop it is possible to display color information, e.g., from the cursor point, in not only RGB values, but also HSB, CMYK, and importantly (for me) in Lab values.  By contrast, color information in ACR is rendered in RGB values, and I do not see any way to alter this to display (for example) the equivalent Lab values.  I'd like to monitor WB changes using Lab values within the ACR interface.  Is there any way to change the default behavior of ACR?  Is there any intrinsic/technical reason why this could not be done?

Thanks in advance for any insights anyone can share.

Cheers,

P.
Title: Re: Lab in ACR
Post by: Schewe on March 27, 2012, 08:14:02 pm
By contrast, color information in ACR is rendered in RGB values, and I do not see any way to alter this to display (for example) the equivalent Lab values.

Since ACR outputs RGB, it can't read nor display other color spaces. ACR reads out in the color space that it's set to output. You'll have to wait till you get into Photoshop. And it's highly unlikely it ever will because it would require putting other color spaces in ACR/LR and I'm pretty sure what the engineers would say, "no"...
Title: Re: Lab in ACR
Post by: MOMOP on March 27, 2012, 08:49:48 pm
Thanks for that response.  During the last few beta test cycles of ACR I've been requesting Lab read-out as a new/enhanced feature.  Not that I expected that the engineers would add a new feature during a beta test per se, but I had hoped that they might consider it for a some future release.  I abandon hope.

P.
Title: Re: Lab in ACR
Post by: deejjjaaaa on March 27, 2012, 11:06:52 pm
Schewe, Schewe... you are supposed to know how ACR is working, aren't you ?

wrote Eric Chan on May 24, 2011 in ColorSync mailing list = "ACR does use L*a*b* for some internal color difference estimates, e.g., for auto-calculated masks"

so I think somebody is just stubborn in Adobe Labs, and I bet it is not Eric  ::)
Title: Re: Lab in ACR
Post by: Schewe on March 28, 2012, 12:12:44 am
wrote Eric Chan on May 24, 2011 in ColorSync mailing list = "ACR does use L*a*b* for some internal color difference estimates, e.g., for auto-calculated masks"

I know ACR/LR moves between RGB and HSL for certain operations and does a lot of luminance based operations but that is internal pipeline operations. It wouldn't surprise me they moved into and out of Lab for some things internally.

But since it can't output in Lab or CMYK, it can't display anything other than RGB readouts...could it?

I suppose, Thomas and Eric can do anything they want...but the question is, why should ACR readout in a color space it can't output to? Once you open the image in Photoshop you can read the Lab or CMYK colors and it's my experience that reading the Lab or CMYK numbers for RGB color correction is, well, not really useful.
Title: Re: Lab in ACR
Post by: MOMOP on March 28, 2012, 01:14:40 am
Thanks to Schewe and deejjjaaa for their comments.  Perhaps a bit of background on my original point is in order...  About 5 years back I picked up a copy of a text by Margulis on the use of Lab in photoshop.  Based on the practical lessons in the book I found that for WB corrections specifically, I could learn to manipulate the a/b controls in photoshop to make rather effective/intuitive adjustments.  I also recall a few frames of a video that Margulis worked up for the NAPP site which specifically emphasized the use of Lab in color/WB correction.  I thought, well, if I can do this in photoshop, why not ACR?  Not trying to make a major point here at all, just sort of providing motivation.

Honestly, in practice I usually either make a shot with a WHiBal card or else just make a creative decision on the WB.  Also, I can nearly read three numbers as well as two ;) 

Cheers,

P.
Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 01:22:30 am
I know ACR/LR moves between RGB and HSL for certain operations and does a lot of luminance based operations but that is internal pipeline operations. It wouldn't surprise me they moved into and out of Lab for some things internally.

But since it can't output in Lab or CMYK, it can't display anything other than RGB readouts...could it?


it can... math is coded there already as they have right now a code that translates color pixel by pixel from internal working colorspace into LAB and it does not take a lot of CPU resources to display LAB values for a single given pixel (or several pixels if several color samplers are used)... does it slow down Photoshop to display a lot of different colorspaces values like this ? no, so it won't slow down ACR either... peanuts calculation-wise... you do not recalculate that for the whole picture to display values just for one pixel or several pixels.

I suppose, Thomas and Eric can do anything they want...but the question is, why should ACR readout in a color space it can't output to?

because some people 'd like to see the readings in LAB or HSV/HSL, may be they do certain adjustments by numbers (and LAB or HSL/HSV are better then RGB for that matter)... and they want to do this in ACR (or LR)... why deprive them ?

PS: And why LR readout in a certain colorspace that has nothing to do w/ its export/output ?


Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 01:33:26 am
Based on the practical lessons in the book I found that for WB corrections specifically, I could learn to manipulate the a/b controls in photoshop to make rather effective/intuitive adjustments.  I also recall a few frames of a video that Margulis worked up for the NAPP site which specifically emphasized the use of Lab in color/WB correction.  I thought, well, if I can do this in photoshop, why not ACR?  Not trying to make a major point here at all, just sort of providing motivation.

if you are so concerned about WB then use raw converter (donation ware) like RPP (http://www.raw-photo-processor.com/RPP/Overview.html) - it does allow you to work w/ multipliers directly for all 4 channels (and your beloved LAB readout, upLAB)... along w/ such religious things as applying WB, exposure corrections,etc before demosaicking (and not after like Adobe does), etc... or a little (actually more than a little) more user friendly (but commercial) Iridient Rawdeveloper (www.iridientdigital.com) - it also has per channel WB (and LAB readout)
Title: Re: Lab in ACR
Post by: Schewe on March 28, 2012, 02:27:52 am
because some people 'd like to see the readings in LAB or HSV/HSL, may be they do certain adjustments by numbers (and LAB or HSL/HSV are better then RGB for that matter)... and they want to do this in ACR (or LR)... why deprive them ?

PS: And why LR readout in a certain colorspace that has nothing to do w/ its export/output ?

Just because ACR/LR may use a certain color space breakdown on tone & color for internal processing doesn't mean the "numbers" of those spaces are available in the ACR/LR UI. Fact is, it isn't.

And truth be told (as indicated by the OP's request) the "reasons" for trying to attain certain numbers in various spaces isn't really what I would call "mission critical".

With regards to Margulis' approach to "color correcting" based on Lab numbers, uh, no. Not a strong use case...Dan still suggests people don't use ACR/LR adjustments but suggests a complicated post processing in Photoshop (using Lab). The problem with that approach is you first make your image look bad before you make the image look better after lots of Photoshop post processing (which defeats the use of ACR/LR controls in the first place). Not a use case that would gain much traction with Thomas.

Unless you can make a use case for reading out numbers that ACR/LR can't output to then I think the odds of getting those reading in ACR/LR is nil. I'm only saying what I expect the results will be, no...
Title: Re: Lab in ACR
Post by: Christoph C. Feldhaim on March 28, 2012, 06:46:51 am
Lab wasn't really designed for colorspace conversions or manipulations.
Lab was made for estimating distances between subjective colors.
Especially the hue shift with L modification is a huge problem if you want to manipulate in Lab.
So - to me it looks quite limited and not really useful in the RAW conversion pipeline where you usually do a lot of Luminance based adjustments.
Title: Re: Lab in ACR
Post by: madmanchan on March 28, 2012, 09:22:00 am
Correct, this is why we sometimes use L*a*b* internally for measurements of color differences (e.g., mask calculations or other cases where colors needs to be distinguished and separated), but not much else.
Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 10:19:36 am
Just because ACR/LR may use a certain color space breakdown on tone & color for internal processing doesn't mean the "numbers" of those spaces are available in the ACR/LR UI. Fact is, it isn't.

yes, they are not available in UI... your initial argument was that they are not available behind the GUI - it turned out they are.

And truth be told (as indicated by the OP's request) the "reasons" for trying to attain certain numbers in various spaces isn't really what I would call "mission critical".

well, I have heard that you were using ACR version 1 ? so I guess everything else that was added since then is not mission critical either, because you were able to get your work done, right ?

With regards to Margulis' approach to "color correcting" based on Lab numbers, uh, no. Not a strong use case...Dan still suggests people don't use ACR/LR adjustments but suggests a complicated post processing in Photoshop (using Lab). The problem with that approach is you first make your image look bad before you make the image look better after lots of Photoshop post processing (which defeats the use of ACR/LR controls in the first place). Not a use case that would gain much traction with Thomas.

well, his approach has certain merits for some people, even if you do not agree w/ it...


Unless you can make a use case for reading out numbers that ACR/LR can't output to then I think the odds of getting those reading in ACR/LR is nil. I'm only saying what I expect the results will be, no...

what about readouts in LR w/ its Melissa ?
Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 10:24:17 am
Lab wasn't really designed for colorspace conversions or manipulations.
Lab was made for estimating distances between subjective colors.

yet a lot of people are using it for what it was not intended and quite happy they are w/ the results

Especially the hue shift with L modification is a huge problem if you want to manipulate in Lab.

lame argument, converters like RPP are using upLab = http://www.brucelindbloom.com/index.html?UPLab.html


So - to me it looks quite limited and not really useful in the RAW conversion pipeline where you usually do a lot of Luminance based adjustments.

you are certainly entitled to your opinion, but reread the points above.

Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 10:29:29 am
Correct, this is why we sometimes use L*a*b* internally for measurements of color differences (e.g., mask calculations or other cases where colors needs to be distinguished and separated), but not much else.
but OP was not asking for anything else except just the readout for a probed pixel in alternative spaces... he was not asking to LAB curves or whatever.
Title: Re: Lab in ACR
Post by: digitaldog on March 28, 2012, 10:46:45 am
it can... math is coded there already as they have right now a code that translates color pixel by pixel from internal working colorspace into LAB and it does not take a lot of CPU resources to display LAB values for a single given pixel (or several pixels if several color samplers are used)...

You are probably right. But the OP wants Lab to preform some of the Lab ‘magic’ Dan M proposes and if you look at his proposals, he’s actually processing the data IN Lab (after what can be a nasty conversion from god knows what original color space). Having Lab feedback is one thing. Having data in Lab to process is I believe something different and something that ACR isn’t doing.

And there are quite a few issues with editing in Lab which we can go into. It has it’s own set of issues.
Title: Re: Lab in ACR
Post by: Christoph C. Feldhaim on March 28, 2012, 11:58:50 am
yet a lot of people are using it for what it was not intended and quite happy they are w/ the results
....
lame argument, converters like RPP are using upLab = http://www.brucelindbloom.com/index.html?UPLab.html
....
you are certainly entitled to your opinion, but reread the points above.

I have rarely had such an impolite, arrogant and self-righteous answer on a posting here on LuLa.
You are fighting windmills.
For me the case is closed due to a lack of basic requirements for civilized discussion.
I'm out of this thread.
~C.
Title: Re: Lab in ACR
Post by: jrp on March 28, 2012, 05:57:31 pm
Unless you can make a use case for reading out numbers that ACR/LR can't output to then I think the odds of getting those reading in ACR/LR is nil. I'm only saying what I expect the results will be, no...

I tend to use Lab settings for assessing color and correcting casts.  You only have to look at two numbers rather than three, which makes things a bit quicker.

Since the output colour space is known, it surely can't be that difficult to do an on-the-fly conversion from RGB to Lab for use in a gauge. 

The work-around is to wait to do the colour adjustments in Photoshop, although that does not help Lightroom users.
Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 06:45:56 pm
You are probably right. But the OP wants Lab to preform some of the Lab ‘magic’ Dan M proposes and if you look at his proposals, he’s actually processing the data IN Lab (after what can be a nasty conversion from god knows what original color space). Having Lab feedback is one thing. Having data in Lab to process is I believe something different and something that ACR isn’t doing.

And there are quite a few issues with editing in Lab which we can go into. It has it’s own set of issues.

well, the question is narrow - just additional readouts in LAB/HSL/HSV... I certainly agree that ACR is not the raw converter where you want to do anything w/ LAB or so - but the readouts of pixel values in such colorspaces in addition are still more informative than RGB readouts alone, even if ACR is totally oriented on a different workflow.
Title: Re: Lab in ACR
Post by: deejjjaaaa on March 28, 2012, 06:49:37 pm
I have rarely had such an impolite, arrogant and self-righteous answer on a posting here on LuLa.
You are fighting windmills.
For me the case is closed due to a lack of basic requirements for civilized discussion.
I'm out of this thread.
~C.

so what you are unhappy with exactly ? that some people work in LAB (and are happy w/ the results) and/or that there is upLAB that corrects certain LAB issues ?
Title: Re: Lab in ACR
Post by: MOMOP on March 29, 2012, 08:21:14 pm
True, what I was originally asking about was only the Lab read-out.  In a subsequent comment I ran on way too long and made it seem that I wanted to conduct Lab manipulations in ACR (which I do not) - sorry about the confusion.  Back to the point, and it really is a small one: I have worked enough in Lab (in Photoshop) that I find the a/b channels values provide intuitive feedback, and it would be nice to have them in ACR.

P.
Title: Re: Lab in ACR
Post by: tom60634 on June 08, 2012, 05:22:43 pm
I understand your use of the lab color space for information, not necesarilly actual adjustments.

I use an Apple Mac system.

One of the more useful features is the DigitalColor Meter.
This useful program can be set to open automatically, with it you can measure L*a*b values from any program that your system can display i.e. all browsers, any imaging program, email attachments etc.etc..

I know nothing of the Windows operating system, but it should have something similar, and similarly useful.

As an aside "Raw Developer" a Mac only raw developer program can display L*a*b values and also allows you to make adjustments in L*a*b.
Title: Re: Lab in ACR
Post by: digitaldog on June 08, 2012, 05:40:21 pm
DigitalColor Meter used to be a reliable way of finding out what RGB values were being sent to the video card, and to the display. That’s about it. That doesn’t tell you about the data in the document.
Title: Re: Lab in ACR
Post by: tom60634 on June 09, 2012, 12:39:43 pm
The read outs seem to match well with the Photoshop info panel when checking lab values.

I've never used if to for RGB values.
Title: Re: Lab in ACR
Post by: digitaldog on June 09, 2012, 01:15:02 pm
The read outs seem to match well with the Photoshop info panel when checking lab values.
I've never used if to for RGB values.

What’s wrong with this picture?

(http://digitaldog.net/files/digicolor_readout.jpg)

And under Lion, no way I see to get Lab values. Considering the readings are based on what is spit out from the video system to the display, probably a good thing.