"Photoshop LAB Color

[Iliah]

... The Canyon Conundrum and Other Adventures in the Most Powerful Color Space" is the new book by Dan Margulis.

This book "is currently is #1 selling on Amazon title in the entire computer/internet field", to quote Dan's publisher. If you have this book, please share your experience on using the methods suggested there, as 1st edition is running out and the work on the 2nd edition will be started soon to add methods and explanations.

 

[cwilt]

I was going to look for this book locally, today. I'll let you know my novice thoughts as I already do some work in LAB space now.

 

[sarhento]

I ordered the book from amazon.ca and can't wait to receive it.

 

[Iliah]

Dear Friends,

Your feedback on the book will be most appreciated! Dan was very clear about that.

 

[JerseyJay]

Iliah,

After checking few topics from Roman Johnston on DPR, I called local Barnes and Nobles and got myself a copy. Didn't look in the book yet but I have read nothing but excellent feedback about it.

Here are few topics

- http://forums.dpreview.com/forums/read.asp?forum=1021&message=14874477
- http://forums.dpreview.com/forums/read.asp?forum=1021&message=14866745
- http://forums.dpreview.com/forums/read.asp?forum=1021&message=14869821

 

[greyflash]

After looking at some of the examples using Lab space for processing I must admit some skepticism. They talk about having spent 30 to 45 min. to process their image in RGB only to get mediocre results. Spending 5 min. in Lab and they have fantastic results. Something doesn't compute for me. First of all it would be extremely rare for me to spend 30 minutes processing a landscape image and second of all I do not find it particularly difficult to make the colors pop using RGB color space. Either their original image was very poor or they are not too good at using PS CS.

I am not knocking the book as I haven't read it. Based on the comments and the fact that it is #1 on the best seller list at Amazon.com I would guess the book is excellent. Am I missing something.

 

[Iliah]

Dear Gordon,

What IMHO matters is cut in time - 5 minutes vs. 30. First few chapters are exactly for that. The big argument about how the book would be received was whether people without a LAB background would be able to get immediate results from the first few chapters. But if one is versed in colour correction, next chapters are for him.

 

[JeffKohn]

After looking at some of the examples using Lab space for processing I must admit some skepticism. They talk about having spent 30 to 45 min. to process their image in RGB only to get mediocre results. Spending 5 min. in Lab and they have fantastic results. Something doesn't compute for me. First of all it would be extremely rare for me to spend 30 minutes processing a landscape image and second of all I do not find it particularly difficult to make the colors pop using RGB color space. Either their original image was very poor or they are not too good at using PS CS.

I can somewhat relate to your skepticism; some of the feedback is so glowing that it makes you wonder. I too rarely spend anywhere near 30 minutes post-processing an image unless there is some complicated masking involved (which I really hate to do and will only resort to if absolutely necessary). I'm already familiar with things like sharpening the lightness channel, or using layers with a blend mode of "luminosity". What I haven't done is actually work with LAB for color correction, so I do think this will be an interesting book. That said I'm not really expecting it to change my life or anything.

 

[greyflash]

Thanks for your responses Iliah and Jeff, you both make some valid points. I have even thought about buying the book but like you Jeff I don't expect it to change my life.

 

[Iliah]

Since first versions of PS one of the most standard workflows in pre-press was to convert scans to Lab (previously it was LCH found in scanner packages for producing colour separations). The main reason for that was usually poor quality of colour and tonality on those originals we were to work with and necessity for pretty wild moves to get things to press.

Later an additional important reason emerged - incredibly poor colour profiles embedded. We are to strip them the frst thing in our workflows and go Lab, using one of our artificial profiles as source. More then often it was some wide-gamut profile with gamma chosen for visually/histogram-based best tonal range spread. It cold be anything from gamma = 0.7 to gamma =3

One of strengths of Lab is when dealing with dull images, as Lab allows to separete colours which are too close by numbers in any other working space, thus extracting slightest colour variations present in the original. The other one is that while editing L curve one does not cause colour changes. The third one is that sharpening L channel does not cause colour artifacts. The forth one is that "a" and "b" channels are representing very useful masks and sources for cannel blendings. The list goes; but of course Lab is not the silver bullet, and not is going to change anybody's life.

 

[cwilt]

I purchased this book today and so far have read the first 4 chapters. So far I have understood everything but I'm sure that will change. Dan has a good writing style.

 

[cwilt]

Iliah,

I am now many chapters into the LAB book and have a question for you. Since current RAW converters do not render an editable file in LAB profile what is the best profile to convert from? I assume you would want the largest possible but with so many conversions, RAW to BetaRGB to LAB to sRGB, wouldn't there be an increasing odds of rounding errors.

 

[Iliah]

While we stay in 16 bit, rounding errors are not so important. As RML uses floating point, precision is very good. Most of Phase II is done in Lab, and we are going to add explicit Lab tools to it; as well as direct output in Lab. We have tremendous speed problems using Lab in Phase I; still we are trying our best to convert curve tool to Lab - to avoid colour cross-overs and other implications of adjusting curves in RGB spaces.

 

[cwilt]

Thanks for the quick reply Iliah. I figured there was some use of LAB in RML but what about profile choice for NC? Does the same 16bit rule apply to it?

 

[Iliah]

Dear Charles,

From NC you are getting normal 16-bit RGB files, and conversion to Lab is handled by PS, right? I do not see why that should pose a problem.

And BTW NC uses Lab too as a stepping stone for LCH.

 

[cwilt]

Thanks again.

I will leave you alone now. :wink:

 

[Iliah]

Thanks again.

I will leave you alone now. :wink:

No way. we are celebrating our middle daughter's birthday

 

[cwilt]

Tell her Happy B-day from us! :biggrin:

Break out the single malt later. :wink: :wink:

 

[David Fink]

Darn it!

I am REALLY disappointed.

I ordered the Margulis book a few days ago.....it should arrive on Tuesday or Wednesday.

Now I learn from this thread that it will NOT change my life after all?? Whatever must I try next?

Seriously though, it should at least be an interesting read, and by all accounts Lab space offers some interesting (and relatively nondestructive) opportunities for tweaking colors and tonality.

Best wishes,

David

 

[Rich Gibson]

I am REALLY disappointed.

I ordered the Margulis book a few days ago.....it should arrive on Tuesday or Wednesday.

Now I learn from this thread that it will NOT change my life after all?? Whatever must I try next?

Seriously though, it should at least be an interesting read, and by all accounts Lab space offers some interesting (and relatively nondestructive) opportunities for tweaking colors and tonality.

Best wishes,

David

I got it yesterday and am still on the first chapter. It looks good so far. Right away, though I noticed I can lift more weight on my bench press. It's also noticeably improving my....ahem...never mind!

:wink:

Rich