Hey, Rutt

DavidTO

So I've started reading Margulis, but haven't gotten far enough for it to really make a difference yet.

But I have been using the lightness channel in Lab mode quite a bit.

And I've got a question for you.

I took this shot (Processed RAW file only, click on image for EXIF):



I then opened it in PS and did some curves work in Lab using lightness, and ended up with this:




My question is, the curves I did to get that were pretty radical, I thought. What do you think? Is this crazy?

rutt

What's the goal here? I suppose we want to increase the detail and hence contrast on the mountain? What about the grass in the foreground? Usually I find that I need very specic goals when writing a new L curve, in particular there is some detail I want to enhance.

I took stab at guessing what you wanted and wrote a set of LAB curves. Here is the result:






My goal was to bring up the contrast in the sky and in the distant mountain. Basically, this image lives in the highlights (sky) and dark midtones (grass, distant mountain, hamlet). I bought steepness throughout the L curve by lopping off the ends, blowing out the sky and darkening the dark grass. Then I flattened the curve a bit in the midtones where the image really has no interesting detail. There result is better definition of the grass and clouds and details of the distant mountain side.

Also, whenever I see vegatation, I think about steepening the A and B curve to make it pop. In this case it also brought out the color in the sky. This technique only works where there already is color, but here there was plenty.

Was this what you had in mind? If not, what is the goal?

rutt

BTW, I uploaded the wrong version of the mountain image, so if you looked before now, you might want to look again at my reply.

DavidTO

rutt wrote:

BTW, I uploaded the wrong version of the mountain image, so if you looked before now, you might want to look again at my reply.

Nope, just saw it.

Thanks for the info, I'll try incorporating your ideas.

The initial question is about the highlights on the grass. The curves that I drew for the image were intended to bring out the highlights of the grass, which are more dominant in my curves than yours. I like the way the highights pop in my version, but was wondering if I was committing some sort of LAB fauxpas.

Which brings to mind one of my favorite cover band names. The Fab Faux.

rutt

BTW, after this kind of radical LAB surgery, it's a very good idea to make the return to RGB through CMYK in order to get everthing back in gamut.

DavidTO

rutt wrote:

BTW, after this kind of radical LAB surgery, it's a very good idea to make the return to RGB through CMYK in order to get everthing back in gamut.

Now that's completely counter-intuitive. Guess I'll learn more as I read.

Andy

i viewed david's edit
and rutt's edit

and i'm dipped if i see a big difference... rutt, where will the diffs show, in a larger print? i'm only looking at the -L sizes of both shots.

DavidTO

andy wrote:

i viewed david's edit
and rutt's edit

and i'm dipped if i see a big difference... rutt, where will the diffs show, in a larger print? i'm only looking at the -L sizes of both shots.

Andy, Rutt, et. al.

The big difference to me is in the highlights of the grass, which gives me more of a sense of contour and depth. Waves of grain, kinda thing. That's what I was going for, and I think the difference is highighted in the section of the photo below. Rutt's on the left, I'm on the right. I think it also gives a sense of sharpness without me having added any. Capture One adds a little, but that's there in both images. I didn't add any extra to my favored image.

Just a little worried that what I did is destructive in a negative way.

rutt

DavidTO wrote:

Andy, Rutt, et. al.

The big difference to me is in the highlights of the grass, which gives me more of a sense of contour and depth. Waves of grain, kinda thing. That's what I was going for, and I think the difference is highighted in the section of the photo below. Rutt's on the left, I'm on the right. I think it also gives a sense of sharpness without me having added any. Capture One adds a little, but that's there in both images. I didn't add any extra to my favored image.

Just a little worried that what I did is destructive in a negative way.

Yea, I didn't undestand what David was after. I sort of liked the "Irish" lush green that I got, but I can understand what he wanted here.

In general it's good to achieve your result with the least radical curve you can. Lots of stuff won't show up on the monitor (or on your monitor) but will spoil prints. So know what you are trying to accomplish and try to do so with moderate changes.

DavidTO

Well, I wasn't planning on ordering a print of this, but now I'm going to have to order two: the two versions in my first post, to see if the curved version looks seriously degraged from the flat version.

rutt

DavidTO wrote:

Now that's completely counter-intuitive. Guess I'll learn more as I read.

This might be more than you really wanted to know, but here goes. Conversion back and forth from LAB and RGB is essentially lossless (well almost, LAB has a wider gamut than any flavor of RGB, see below), but conversion into CMYK is essentially not lossless. When PS converts to CMYK, it always does what it has to in order to make the image fit into the gamut of the flavor of CMYK. This is because traditionally the prepress people have used CMYK to get images ready for particular presses and they want to work within the gamut of the target device.

There is a also a theoretical reason. There is only one way to represent a particular color in RGB. But the existence of black ink gives us an infinite number of ways to represent most colors in CMYK. The simplest example is black, which can be represented either with black ink or equal amounts of C,M,&Y inks. Or you can use 1/2 as much black and 1/2 as much of the colored inks. So conversion to CMYK essentially involves decisions. The rules PS uses are designed to result in in-gamut images.

Try View->Gamut Warning with your image. Then convert into CMYK and back to RGB. You'll see.

Having your images in gamut is goodness and light. Out of gamut images have a tendency to look different not only on different devices but also on different display software.

OK, I also promised to explain why LAB->RGB conversion is sometiems not lossless. This is because LAB is so powerful it can actually represent imaginary colors. For example with L = 100, A=100, B=0 you have a green that is as bright as white (L = 100, A = 0, B = 0). But that isn't really a possible color. In RGB you could make it brighter by adding more red and blue. In CMYK you could make it brighter by adding less ink and letting the underlying white reflect more light.

DavidTO

rutt wrote:

Try View->Gamut Warning with your image. Then convert into CMYK and back to RGB. You'll see.

Sounds like I should do this on everything, to bring it all within range. I mean if I can't rely on an image to provide a reliable print, it's not much good.

I guess I've got to read more...