LAB Overlay Technique

[macaddict]

This is a storage house for the historical archive of the Chosun Dynasty from the 19th century. My wife and I stumbled upon it while hiking the mountains of Korea's east coast.

I found this photoshop tip at NAPP (http://www.photoshopuser.com). (NAPP is a great place to learn by the way.)

Open a flat color shot like this:
http://d200guy.smugmug.com/photos/49759774-M.jpg
Go Image>Mode>Lab Color. Not much happens, but then go Image>Apply Image. In this dialog box you change the blending mode to Overlay and lower the opacity to something that looks good. Also, choose the Lab channel that gives the most punch to the colors. In this case, I lowered the blending opacity to 35%:
http://d200guy.smugmug.com/photos/49759760-M-1.jpg
I love this technique!
Sorry if it is old-hand to you - I've been using Photoshop for years and learn something new almost everyday!

(BTW, I just got my Amazon shipping notice on "Photoshop LAB Color.")

Dave

[DavidTO]

Nice results with that technique. I'll have to try it out. It seems like it is generally related to the Man From Mars technique, in that it makes the colors way too much, and then you scale back by using the opacity slider. Unlike MFM, though, it seems that it would effect the L channel much more significantly. More after I play with it (and Rutt and Edgework chime in...)

Also, I split this post from the Portrait thread, which we want to keep on track with the techniques covered in Chap. 16....

[Tom K.]

I thank you so much for posting this superb technique. I have tried it on several images and it works quickly and with much drama added to the images. Great job and nice tutorial. :thumb

[chrisjlee]

Nice results with that technique. I'll have to try it out. It seems like it is generally related to the Man From Mars technique, in that it makes the colors way too much, and then you scale back by using the opacity slider. Unlike MFM, though, it seems that it would effect the L channel much more significantly. More after I play with it (and Rutt and Edgework chime in...)

Also, I split this post from the Portrait thread, which we want to keep on track with the techniques covered in Chap. 16....

Seems like an easier faster way to use the MoMars technique. Which I love to death by the way.

[Tom K.]

Does anyone have a technical explanation of why this technique work so well. I have been using it and it really can add serious punch to an image.......quickly.

[edgework]

Go Image>Mode>Lab Color. Not much happens, but then go Image>Apply Image. In this dialog box you change the blending mode to Overlay and lower the opacity to something that looks good. Also, choose the Lab channel that gives the most punch to the colors. In this case, I lowered the blending opacity to 35%:
Dave

Could you be more specific about:

• Which channel is active at the time of the operation? (I'm assuming the LAB channel, meaning all three at once.)

• Which channel you choose as the blend channel?

The color enhancement makes sense for the same reason that it works in Dan Margulis' portrait recipe. When A and B are overlayed into themselves, values close to neutral (50% gray in the channels) are affected least, everthing else gets pushed farther out along the color axis and so becomes more brilliant. Lighter values in the A and B channels become lighter still, darker values darker still, which means Reds become Redder, Blues bluer, etc.

It's the Lightness channel that has me flummoxed. I'm having trouble duplicating what seems to be an overall lightening in the sample image. An overlay operation that blends the lightness channel into itself will push contrast making shadows darker and highlights lighter. Is is my imagination, or is there MORE detail underneath the structure after the operation than before? When I overlay the entire image into itself, the shadows plug up unacceptably, certainly not like the results obtained here. That's why I wanted some clarification on the set-up.

[jfriend]

Could you be more specific about:

• Which channel is active at the time of the operation? (I'm assuming the LAB channel, meaning all three at once.)

• Which channel you choose as the blend channel?

The color enhancement makes sense for the same reason that it works in Dan Margulis' portrait recipe. When A and B are overlayed into themselves, values close to neutral (50% gray in the channels) are affected least, everthing else gets pushed farther out along the color axis and so becomes more brilliant. Lighter values in the A and B channels become lighter still, darker values darker still, which means Reds become Redder, Blues bluer, etc.

It's the Lightness channel that has me flummoxed. I'm having trouble duplicating what seems to be an overall lightening in the sample image. An overlay operation that blends the lightness channel into itself will push contrast making shadows darker and highlights lighter. Is is my imagination, or is there MORE detail underneath the structure after the operation than before? When I overlay the entire image into itself, the shadows plug up unacceptably, certainly not like the results obtained here. That's why I wanted some clarification on the set-up.

It's not your imagination edgework. I can't reproduce the Apply Image effect Dave shows either. I've tried all combinations of using Apply Image to blend one channel into either the composite or another channel, including all the [X] Invert possibilities. Nothing produces close to what what he posted. I can get closer if I first do a pretty strong shadow/highlights action to raise the shadows or a major curve move, but still can't reproduce what he has. It appears that some steps must be missing from the recipe.

[DoctorIt]

this is super easy and does some nice work! ready to learn more :ear

[bham]

Hey macaddict I am trying to replicate your results and while I am getting the idea, the images are not coming out with the same result. Did you do something else to the image to get the result.

[SCS_Photo]

I just did a Gimp translation of this technique here: http://www.dslreports.com/forum/remark,15173919

Tom K. (Edit This @ DSLR) posted the PS translation here, and DSLR members provided quite a few conversions: http://www.dslreports.com/forum/remark,15129269

Thanks for reminding me of this technique, I tried it when Tom first linked to Dgrin.

-Sam

[macaddict]

Image mode changed to Lab Color, then Image>Apply Image. Change blend mode to Overlay. Change channel to B. This is at 100% opacity:
http://static.flickr.com/42/82486704_1f20aa02b6.jpg (http://static.flickr.com/42/82486704_1f20aa02b6_b.jpg)

Change channel to A. This is at 100% opacity:
http://static.flickr.com/40/82486731_d3a5c97e57.jpg (http://static.flickr.com/40/82486731_d3a5c97e57_b.jpg)

Change channel to Lab. This is at 100% opacity:
http://static.flickr.com/37/82486752_bf5d7e19b3.jpg (http://static.flickr.com/37/82486752_bf5d7e19b3_b.jpg)

Original image:
http://static.flickr.com/36/82486954_7eacd987f9_o.jpg

The best punch, in my opinion, is B. So, I lowered the Overlay to 30% opacity and here's the final image:
http://static.flickr.com/38/82486851_120f000717_o.jpg

That's all there is to it - a quick and easy punch.

I will say, I'm now reading Chapter 3 of Margulis's Photoshop LAB Color - I'm amazed at the results using his techniques!

[bham]

Thanks for doing that. I have tried about everything to get the exact match on the first two pics and couldn't and still can't but could on the second example. I was doing the right thing but for some reason the images don't end up same.

Thanks again. Great technique.

[jfriend]

Image mode changed to Lab Color, then Image>Apply Image. Change blend mode to Overlay. Change channel to B. This is at 100% opacity:


Change channel to A. This is at 100% opacity:


Change channel to Lab. This is at 100% opacity:


Original image:


The best punch, in my opinion, is B. So, I lowered the Overlay to 30% opacity and here's the final image:


That's all there is to it - a quick and easy punch.

I will say, I'm now reading Chapter 3 of Margulis's Photoshop LAB Color - I'm amazed at the results using his techniques!

OK, this is fairly easy to understand. There must have been some other significant things you did to the original image you started this thread with because a strict overlay doesn't produce the result you ended up with. No worry.

Anyway, overlaying a LAB color channel with itself is very much like steepening the curves on the same LAB channel in chapter 1 of Margulis' book. If the image doesn't have any brilliant colors (which is true of most images), then it's nearly identical.

For reference, Rutt posted here (http://www.dgrin.com/showpost.php?p=202259&postcount=71) how you can get the exact same result as Apply Image/overlay with a curve (it's just a large S-curve that does the same thing as the overlay blend mode).

I, myself, prefer using the curve over Apply Image because it's 200% more flexible. Just switch to LAB, create a curve adjustment layer, switch to the A or B channel and steepen the curves. You can either push in the ends (a la chapter 1 in Margulis' book or you can make a big S-curve that approximates the overlay mode (see Rutt's posting referenced above).

I prefer this way because not only can you vary the steepness of the curve and the opacity to control how much of an effect you get, but it goes in an adjustment layer so you can easily turn it on/off to gauge the effect and even dial it back after-the-fact. Further, you have the abilty to manipulate the shape of the curve which can apply more of a change to some colors than others and you can use the BlendIf settings to dial-back the effect from brilliant colors if you need to exclude them. And, lastly, you can also adjust any color cast with the curve if any appear when you enhance the colors (which happens to me about 1/3 of the time).

So, feel free to use Apply Image/Overlay if you want, but I think it gives you a lot more control and flexibility to use a curve adjustment layer. Thanks for sharing.