Color corrections LAB or RAW

[bart1986]

I'm reading the Dan Margulis LAB book and there's one little thing that i find odd. Dan makes an entire chapter about color correction in LAB, but what are the advantages compared to doing it in the RAW converter? Which happens to be non-destructive?

Are there advantages to doing it in LAB or is the book slightly out-dated on this subject?

Sorry if my english is messed up, its not my native language.

Thank!

[John Wheeler]

Hallo Bart en ik geloof dat je uit België. Is uw taal het Nederlands, Frans of Duits? Ik neem een gok dat het Nederlands is. Welkom bij ReTouchPro. (Courtesy Google translate).

The tools available are rarely a one vs the other choice. Personally, I find the tools in Photoshop ACR or Adobe Lightroom quite robust, intuitive for the photographer and not only are you working with the original RAW files non-destructively, you can open them in Photoshop as a Smart Object so you can go back to the RAW file and make additional edits. I opt to make most of my changes with ACR or Lightroom, take them into Photoshop as Smart Objects and if still desired, use the techniques available in Lab Color Mode/Space.

LAB has many parallels to learning a new language. Tough at first yet when mastered, it is another tool that opens up possibilities and new ways to communicate not always possible in other languages. LAB's color mode/space (some would call it a religion) is another paradigm of how to get your work done. Whether it has benefit for any one person's situation/workflow I believe is decided individually. I personally appreciate having the LAB avenue available yet find that I use it sparingly. For others, I am sure it is a key part of their workflow. All of course IMHO

[bart1986]

Originally Posted by John Wheeler

Hallo Bart en ik geloof dat je uit België. Is uw taal het Nederlands, Frans of Duits? Ik neem een gok dat het Nederlands is. Welkom bij ReTouchPro. (Courtesy Google translate).

The tools available are rarely a one vs the other choice. Personally, I find the tools in Photoshop ACR or Adobe Lightroom quite robust, intuitive for the photographer and not only are you working with the original RAW files non-destructively, you can open them in Photoshop as a Smart Object so you can go back to the RAW file and make additional edits. I opt to make most of my changes with ACR or Lightroom, take them into Photoshop as Smart Objects and if still desired, use the techniques available in Lab Color Mode/Space.

LAB has many parallels to learning a new language. Tough at first yet when mastered, it is another tool that opens up possibilities and new ways to communicate not always possible in other languages. LAB's color mode/space (some would call it a religion) is another paradigm of how to get your work done. Whether it has benefit for any one person's situation/workflow I believe is decided individually. I personally appreciate having the LAB avenue available yet find that I use it sparingly. For others, I am sure it is a key part of their workflow. All of course IMHO

Thanks! It always amazes me how badly the google translate thing works. You might as well just take some words from the dictionary, put them in a random order and type that lol. Anyhow, yes i speak Dutch.

So you are looking at it from a personal taste, and your taste is more towards RAW. But lets say you don't have a personal taste, are there any real benefits to doing it completely in LAB, stuff you can't possibly do in RAW and what is it?

[Tony W]

IMO, RAW should be your first stop to maximising image quality. As it is non destructive this is a huge benefit. Once you have gone as far as you are able in RAW then consider further work in Photoshop. Think carefully about the image and if it will benefit from LAB specifically - the canyon scenario from Dans book for instance or gross changes to colour. Can you achieve the same in RGB which will save a trip in and out of colour spaces - you will likely have to convert after LAB for your output device!

I think overall Dans books are excellent (if a little hard going at times) but he does seem to dismiss other options too easily at times. From memory ACR is one of those.

I cannot think of a benefit of ignoring RAW and going straight to LAB unless the image type absolutely demands it.

[bart1986]

Originally Posted by Tony W

IMO, RAW should be your first stop to maximising image quality. As it is non destructive this is a huge benefit. Once you have gone as far as you are able in RAW then consider further work in Photoshop. Think carefully about the image and if it will benefit from LAB specifically - the canyon scenario from Dans book for instance or gross changes to colour. Can you achieve the same in RGB which will save a trip in and out of colour spaces - you will likely have to convert after LAB for your output device!

I think overall Dans books are excellent (if a little hard going at times) but he does seem to dismiss other options too easily at times. From memory ACR is one of those.

I cannot think of a benefit of ignoring RAW and going straight to LAB unless the image type absolutely demands it.

So is this something about the book being out-dated? Were there any RAW possibilities at the time the book was written? Because seriously after reading the first 5 chapters i am like, i can do this with photoshop/ACR and get the same results. Apart from the little thing he does in chapter one where he just makes a slight curve adjustment in the A&B channel in LAB mode which is quick, but again, can be done with PS. It's basically getting the colors to pop which you might as well do with a layer set to hue, take a foreground of 100% hue and a back of 100% de-saturation and start painting with different opacities.

[RobertAsh]

Originally Posted by bart1986

So is this something about the book being out-dated? Were there any RAW possibilities at the time the book was written? Because seriously after reading the first 5 chapters i am like, i can do this with photoshop/ACR and get the same results. .

The book is not outdated. And sometimes you can't always get the same results, certainly not nearly so easily (at least for me personally). That's one of the benefits of Lab color.


That said, Lightroom and ACR do incorporate Lab color already, to some degree, at least approximately:

• The color balance sliders (Temp and Tint) in the white balance section manipulate the B channel (Temp = yellow / blue slider) and the A channel (Tint = magenta / green slider)
• The Exposure / Brightness sliders and the tone curve manipulate the L channel.

But effectively you're limited to one set of L-a-b curves (one curves layer) with no masking.

So Dan's book is not obsolete but there are now ways to do similar things through different means. But of course - that's Photoshop for you

[RobertAsh]

Originally Posted by RobertAsh

...But effectively you're limited to one set of L-a-b curves (one curves layer) with no masking....

And even this functionality that I shared above in Lightroom and Adobe Camera Raw is, at best, "a poor man's Lab color." It has pretty stark limitations:

• You can't manipulate Lab curves for the colors (A and B channels in Lab). The sliders induce or reduce color casts for basic white balance. If you think about the curves layers in Photoshop, the Temp and Tint sliders only allow you the equivalent of moving one point on the A or B channel up or down.
• You can't stack layers. In Photoshop you can stack multiple Lab curves adjustment layers. Can't do that in LR or ACR.
• You can't do selective application. Sliding the Temp or Tint sliders affects the entire image.

One strong upside to using LR or ACR is the Vibrance slider. Lab mode in Photoshop does not support the Vibrance adjustment layer. Would be great if Adobe made that happen in the next release of Photoshop.

So Dan's book is still very much on track, and which technique to use, or whether to use Lab or not, depends on your needs and working style.

There is a section in his book where he compares the strengths and weaknesses of all 3 major color spaces (RGB, CMYK, Lab). It's well done. Also, he cites on example where in a class of his a participant did best him in a challenge to do some technique in RGB. He said that was educational. So he's open to at least some degree to learning and to acknowledging the limitations of what is after all a human tool.

But for me, Lab has enabled me to do things I couldn't do before, and has enabled me to do things much more easily, more intuitively and faster than I could do in RGB. Nowadays I use Lab for landscapes and travel and RGB for portraits, and CMYK for skin tone color correction when I have a tough color cast I'm trying to eliminate, or (learned this from Dan's book) when there's some shadow detail I want to extract that I can't do in either of the other color spaces (that's rare but it's happened).

[DWThomp]

Originally Posted by RobertAsh

Nowadays I use Lab for landscapes and travel and RGB for portraits, and CMYK for skin tone color correction when I have a tough color cast I'm trying to eliminate...

Do you go from one to another and then back again? Ex. - start out in RGB, switch to CMYK for some correction, and then back to RGB. Is there any harm doing that. I usually stay in RGB, but will change the sampler readout to CMYK at times.

[RobertAsh]

Originally Posted by DWThomp

Do you go from one to another and then back again? Ex. - start out in RGB, switch to CMYK for some correction, and then back to RGB. Is there any harm doing that. I usually stay in RGB, but will change the sampler readout to CMYK at times.

Sometimes I do. I often start out in Lightroom which uses an RGB variant. If I like something I do there, I'll move into Photoshop and convert profile to Lab color.

No harm done in moving between color spaces that I've personally encountered. The one problem is that when you change color spaces you lose all your adjustment layers in the previous color space and can't re-trace your steps (or you have to save multiple copies of the file, one for each color space). Making it into a Smart Object should work to avoid that, but for me it that doesn't work.

So I'll either save a different copy of the file or, more often, will do:

• Image>Duplicate,
• Convert the duplicate to the other color space (e.g. CMYK)and work in the other color space.
• Convert the duplicate back to my original color space (e.g., Lab) and flatten or Merge Visible.
• Drag the resulting layer into my original file as a new layer.

Gets tricky but it's manageable when the need is there.

[Tony W]

Originally Posted by bart1986

So is this something about the book being out-dated? Were there any RAW possibilities at the time the book was written? Because seriously after reading the first 5 chapters i am like, i can do this with photoshop/ACR and get the same results. Apart from the little thing he does in chapter one where he just makes a slight curve adjustment in the A&B channel in LAB mode which is quick, but again, can be done with PS. It's basically getting the colors to pop which you might as well do with a layer set to hue, take a foreground of 100% hue and a back of 100% de-saturation and start painting with different opacities.

The information given in the books is not outdated IMO. He is presenting different ways to reach your final goals.

ACR was around at the time he wrote the book no idea which version. As far as I remember he spent no time in ACR and gave the impression that he dismissed it as not being valuable. On the other hand somewhere he did say that RAW converters were coming of age and a hotly fought for area. So I suspect that his opinion may have changed to date?

Only you will be able to decide which method to make colours pop. In my view LAB is capable of doing a quicker and possibly better job in most cirumstances.

Here are Dans opinions from Professional Photoshop fifth edition 2007 on LAB

• Excellent for quick correction when no time for fine tuning
• Best sharpening space for images not dominated by one colour
• Allows cost free blurring of colour noise
• Easy to eliminate gross colour casts
• Steepening the A dn B is the most natural looking way of creating colour variation or more intense colours

If your image falls within any of the above categories then perhaps you should consider LAB as an option. Even so I would still want to assess and take the image as far as possible first in ACR. AFAIK any RAW capture you make will open first in Photoshop via the Camera RAW module (not aware if you can turn this off) so why would you not take advantage of this first?

[andrewrodney]

Much of the issues that need to be ‘fixed’ by moves into Lab are able to be addressed when you render the raw data. Not all but many. There are significant advantages to doing all the heavy lifting you can do in the raw converter and its here that Dan and I strongly disagree! He feels you should zero out the slides, produce a butt ugly rendering and fix that in Photoshop (if all you know is a hammer, everything looks like a nail).

Raw processing is pixel creation. Never forget GIGO:GARBAGE IN GARBAGE OUT. Being careless in a raw converter is akin to the photographer who improperly exposes his film, then has to fix that in the lab/processing. Not ideal. But then Dan’s not a photographer...

The advantages to doing all the work possible in a raw converter is its totally, fully non destructive. You are creating metadata instructions that tell the converter how to build ideal RGB pixels.

Creating metadata is far faster than altering existing pixels. Think about it. You are not altering numeric values of millions of exiting pixels, you are building a text file.

As a raw converter improves, as does the rendering. In Adobe products, we just saw a significant increase in image quality between the 2003 and new 2010 PV processing (noise reduction, demosaicing, just flat out, visually superior). Want that better processing? Easy. Render the raw using the existing settings.

Raw processing is JIT processing meaning, you build the millions of pixels only when you need them. Photoshop is a memory and processor hog because no matter what you want to do there, it has to deal with millions of pixels. For pixel editing, that’s necessary, no question. But for a lot of image processing its not.

All the processing is done in an optimal processing order. Unlike Photoshop, where you can really hose the data or trip yourself up if you don’t process in the correct order, it doesn’t mater at all in Adobe raw processors. You can play with exposure, then saturation or vibrance, then go back to Exposure. The order makes zero difference because when you render the data, Adobe will process the edits in the optimal order. All in high bit, wide gamut color to maintain as much data as possible.

Of course there are tools that require Photoshop, no question. But for 95% of all global and more recently, even selective color and tone editing, you can do it better, faster, cheaper in a raw converter. That’s why we capture raw data too, for the control we have over rendering. Once you render an image, there are things you either can’t do effectively or at all in Photoshop (hint, just try taking an ugly JPEG shot with the wrong white balance and fix that in Photoshop versus using a white balance tool in the raw converter on the raw. Huge differences here).

Don’t build turds and then spend hours fixing them in Photoshop unless you get paid by the hour and need to find ways to build turds. Produce the best possible data with the raw converter before you even think of bring Photoshop to the party.

[andrewrodney]

Originally Posted by Tony W

ACR was around at the time he wrote the book no idea which version. As far as I remember he spent no time in ACR and gave the impression that he dismissed it as not being valuable.

Guilty as charged. That alone is oh so telling. OK, now here did he put that hammer....

[John Wheeler]

Originally Posted by DWThomp

Do you go from one to another and then back again? Ex. - start out in RGB, switch to CMYK for some correction, and then back to RGB. Is there any harm doing that. I usually stay in RGB, but will change the sampler readout to CMYK at times.

Hi Dennis, I am responding to your specific question about round trips from RGB to CMYK color space. I do not do round trips to CMYK color spaces and also just use the color sampler readout. The reason, virtually all CMYK color spaces have a much smaller color gamut (and visually noticeable in many images). Doing a round trip to CMYK will compress many colors to that smaller gamut and it will remain there on the return trip to RGB.

Even if your original image in RGB only has colors that exist in the CMYK color space there is something different in translating to CMYK and back than with other color spaces (I can provide the reference if that were really needed). A round trip from RGB to CMYK (with no other edits) does not do a great job of preserving the pixel numbers.

If you really want to do a round trip to CMYK, there are ways to minimize such impacts yet that would be the topic of another thread.

[Tony W]

Andrew, your first post really made me laugh . Not the overall information about why RAW should be used, which IMO is spot on rather it was your comment 'Don't build turds...'. Having at times throughout my life resorted to semi professional turd polishing I can tell you it gets you nowhere slowly . Sorry but I just could not resist adding comment.

[bart1986]

haha great comment about the turd! But very true! I finally ended up by taking the information about LAB and bringing colors closer together by steepening the curve, that idea i liked and i think it is quicker than in RGB, but defiantly not impossible or better IMO. That idea i'll keep in the back of my head, the rest of the stuff i believe is RAW stuff and stuff i might as well do in RGB

[kav]

Not to derail this but has anyone ever found a truly compelling reason to bother with LAB? I messed around with it back around photoshop 6 (I think it was 6) but LAB to me seems best used for referencing colors rather than working in photoshop. Not all rgb/cmyk gamuts contain the same colors but you can pretty much reference them in terms of LAB numbers. I'm not sure what the advantage would be in choosing to actually work within such a space in photoshop.

[andrewrodney]

Originally Posted by kav

LAB to me seems best used for referencing colors rather than working in photoshop. Not all rgb/cmyk gamuts contain the same colors but you can pretty much reference them in terms of LAB numbers. I'm not sure what the advantage would be in choosing to actually work within such a space in photoshop.

Lab has been around long, long before Photoshop. And yes, that’s its original purpose, as a device independent (kind of) color space for communicating color in an almost non ambiguous way. And its been tweaked over the years due to faults in its design and math.

Keep in mind that Lab was just an attempt to create a perceptually uniform color space where equal steps correlated to equal color closeness based on the perception of a viewer. The CIE didn't claim it was prefect (cause its not). Most color scientists will point out that Lab exaggerates the distance in yellows and consequently underestimate the distances in blues. Lab assumes that hue and chroma can be treated separately. There's an issue where hue lines bend with increase in saturation perceived by viewers as an increase in both saturation and a change in hue when that's really not supposed to be happening. Further, according to Karl Lang, there is a bug in the definition of the Lab color space. If you are dealing with a very saturated blue that's outside the gamut of say a printer, when one uses a perceptual rendering intent, the CMM preserves the hue angle and reduces the saturation in an attempt to make a less saturated blue within this gamut. The result is mathematically the same hue as the original, but the results end up appearing purple to the viewer. This is unfortunately accentuated with blues, causing a shift towards magenta. Keep in mind that the Lab color model was invented way back in 1976, long before anyone had thoughts about digital color management.

[kav]

Now I remember you mentioned this before. What makes a good reference space now? I still do see measured print values referenced by LAB at times. Is there a better way to reference said values from a larger space(note my post in no way suggesting working on your images in LAB)?

[andrewrodney]

Lab is a good reference space because the numbers are fixed in stone (they describe how we humans perceive a color). In sRGB, 0R-255G-0B is a different color than the same values in ProPhoto RGB, or Adobe RGB (1998), or myepson3880RGB. A Lab value doesn’t behave this way, the values describe how the ‘standard observer’ perceives that solid color. So Lab is super useful in color management, for converting values among color spaces and so forth. Its the universal translator of color values because its device independent unless you consider the standard human observer a device.