LAB color and saturation

[edward_karaa1]

Does anyone recommend increasing the saturation in LAB mode? I tried it today

on some photos and it seems to give different results than in RGB mode. I'm

not sure which one I like more, but the LAB saturation seems more natural. Any

advice on the best way to do it?

Thanks.

[paul_sokal___dallas__tx]

You should read Scott Kelby's book on photoshop channels. It gives lots of info on this sort of thing. He describes a technique of Dan Margulis' to increase saturation using LAB color. You creat a duplicate layer and use the apply image command to basically multiply the a and b channels. Then you can adjust your final result by varying the opcaity of the duplicate layer.

[walterh]

Or read "Professional Photoshop" by Dan Margulis :-)

 

It is only 400 pages and with a little practise after digesting the essence from the book you will get by ok.

 

Edward unfortunately color correction or artistic use of color is complex and can not be answered in a few lines.

 

A short and stimulating approach would be to try different things and see how this works for you. The problem is that the people who used to do the color printing for you in commercial printing labs needed to get an education over years (at least the ones with "professional" output). Now we get a digital camera and think we can do it as well -hmm no - better in three days. Simply does not work. I am always amused to see how some professionals take a few quick twists on my "problem" images (the ones I gave up on after hours getting nowhere) and all the sudden it looks perfect and it seemed so easy ^^. A friend of mine used to say the first 10 years he had problems too.

[minicucci]

Edward: I agree with Walter's comments and would add that Margulis has a newer book out

that deals specifically with the LAB colorspace (Photoshop LAB Color).

 

Personally, for me the charm of LAB is the complete separation of luminosity/contrast from

color which cannot be replicated in RGB. That allows for some mighty powerful editing/

retouching but not without the cost of subjecting an image to a round trip of color mode

changes. Worth it if you need it but not an automatic, especially for a well shot image.

[blakley]

Get Dan Margulis' book "Photoshop LAB Color".

 

His basic technique for increasing saturation is to convert the image from RGB to LAB, then apply a curve like this to both the A channel and the B channel.<div></div>

[blakley]

Note that it's important that the center of the curve go through the exact center of the grid; this is what maintains neutrality of whites and grays. If you use this technique, you can turn a picture like this....<div></div>

[blakley]

... into one like this:<div></div>

[eugene_scherba]

Hmm, Bob, I think I prefer something in the middle.

[edward_karaa1]

Thank you all for your advice, especially to Bob for the short tutorial. This technique is definitely worth a try, especially that images have a different look than with boosting saturation with RGB. This is going to complicate my life for a while as I won't be able to decide which looks better ;-)

[blakley]

So do I, Eugene; I exaggerated the effect for illustrative purposes (and because I don't trust that everyone's screen is calibrated the same way mine is....).

[roger_kallet]

Here is the link to the "Margulis LAB Color Book-Reading Group." Plenty of LAB Color info there.

 

 

http://www.dgrin.com/showthread.php?t=18203

[edward_karaa1]

Roger, thanks a lot, that's a very useful link.

[glenn_chan]

One of the good things about Lab saturation adjustments is that it (more or less) maintains the same lightness*.

*Aside: Technically Lab is designed so that the Lightness component tries to be uniform in spacing, where spacing your ability to tell two shades of color apart. This is slightly different than your perception / your brain's interpretation of lightness.

Because Lab does this, you don't see the color shifts evident in the normal Hue/Saturation algorithm. e.g. if you just increase hue/saturation, the colors tend to turn neon very quickly.

 

2- One of the downsides to Lab saturation is that it does not maintain constant *perceived* hue and saturation. I demonstrate that in this white paper here:

 

http://colormancer.com/whitepapers/comparison/filter_comparison.html

Scroll down to "The problem with CIE L*a*b*"

 

Colors like blue can turn purple-ish since Lab is designed for perceptually uniform *spacing* (ability to discriminate between two colors), which is not the same as our perception/interpretation of color. In my tests, the Hue/Saturation algorithm in Photoshop does a lot better in that department. But this effect is fairly subtle, especially if you are not dealing with large patches of color, real-world colors (which tend not to be highly saturated), or you aren't making big changes in saturation.

 

3- To get back to the original topic a little more, a totally different way of increasing saturation would be to implement a s-shaped color curve on the RGB channels.... this boosts both contrast and saturation. Film stocks to some degree are designed to do this as it makes images look better.

 

You might already know that trick.

 

Glenn Chan

Colormancer