What I've been doing in PL recently

[Herbert123]

Here's my quick test run:

try1.jpg

[img=576x768]http://www.estructor.biz/testje/try1.jpg[/img]

Before you start manipulating any colours, make absolutely certain to colour correct - the original's histogram was all over the place; otherwise any "improvement" requires more adjustments to counter-act the incorrect shifts in colour.

It looks a bit painterly, but that's what you get when working in low-res photos. And the hand needs a lot of work (maybe even replacement), but I could not be bothered to do all that - I do not like the hand at all in this photo, though - the shape takes away from the model's face. Nails are still too red - but again: this is just a quick'n dirty correction, and needs much more work, though with a low-res image the point is moot.

I made all sorts of mistakes in my hurry (left cheek is too red, for example), and the eye was unsalvageable - I grabbed one from a googled black model. Interesting enough for me that this is my first model correction in Photoline - ordinarily I would have used Photoshop. I have to say I prefer Photoline's approach and control. Having the option to work in LAB mode at any time with curve adjustment layers is gold to me!

A lot of brighter/darker painting to increase visual interest/contrast, especially the hair. Skin looks a smite too smooth, but at this res it would be difficult to restore the needed skin detail, so I left it as it is.

Hope you like it a bit - I've included the (junky) pld for you to disect. The detail in her hair is awesome, btw.

PLD file: http://www.estructor.biz/testje/try1.pld (right-click to save as)

[Herbert123]

Oops, noticed I made her skin a bit too yellow.

Adjusted version:

try2.jpg

[greenmorpher]

Hiya Herbie

Interesting stuff. Lab mode, eh. Hmm -- I'll have to look at that. I worked exclusively in RGB.

You mention the nails being too red; they have ended up redder than the lips in your rendition, which was the opposite of actuality. How did that happen, can you say? My problem when doing this stuff is that I start manipulating and stuff gets out of balance. The next thing you know, everything turns to muddy brown!

You eye is much more naturalistic that I have done. I went to an extreme on that because it is for poster/packaging use but I freely acknowledge it is over the top. I would have liked something a touch more restrained -- something closer to what Hoogo did.

I'm learning lessons, guys. Many thanks.

Cheers, geoff

[Herbert123]

The main reason for colours to shift into unrealistic looking ones, or adding more to colour casts: always, always colour correct first. Meaning: whites should be whites, black should be blacks, greys should be greys (though this is not always the case with some pictures). Know your colour values for typical subjects. And as long as it is believable, and looks better, it will be fine.
After colour correction, add contrast, etc. And towards the end local selections and masks can be utilized for localized changes.

For example, I did not like the lips makeup colour in the original photo: too purplish red after colour correction. It did not suit the model's skin tone (in my opinion), so I masked the lips, and added a hue/saturation adjustment layer to get rid of some of the saturation and change the colour a towards a reddish tone. I also increased the overall saturation of the image in Lab mode (see below). I masked the background to remove the saturation, adding more focus on the facial tones and the hair.

Again, the eye is taken from another photo I found on the web. It's not her eye - perhaps cheating, and that is required and done (unfortunately) in the industry - no magazine cover model or advert model is unretouched these days. I know this, you know this, we all know this. I painted in shadows and highlights wherever I thought it would improve the contrast/look. Mind, I did it very quickly, made mistakes, and the overall look could be much improved.

To really understand colour correction / image enhancement, I recommend without question Dan Marguilis's books:

http://www.amazon.com/Photoshop-LAB-Col ... 0321356780
http://www.amazon.com/Professional-Phot ... 032144017X

German edition:
http://www.amazon.com/Photoshop-Farbkor ... 3827325463

He also made a bunch of video tutorials (google it).

Even though these are aimed at Photoshop, all the techniques and theory are readily applicable in Photoline as well - these books are aimed at understanding the theory, and much less about the actual program used to achieve it.

I'd even go as far as saying that Photoline achieves the same effects in an improved workflow compared to Photoshop: in Photoshop one is required to switch between image modes *all the time*, and this is then applied to the entire image. Convoluted "smart objects" in PH are only a patch to an old-fashioned workflow. In Photoline each layer can have its own mode, and the curves can be applied in LAB mode whenever you like: absolutely brilliant, and completely surpasses Photoshop's workflow. For an advanced user like myself this opens up whole new vistas!

One really nice trick in Lab mode: add a curves adjustment layer, switch to LAB, activate the a channel, and move the left bottom and right top points of the curve ~20%-30% inward, compressing the a values. Do the same identical move in b mode. All your colours will be much more saturated, but the contrast will not have changed one bit: courtesy of the luminance channel in Lab. Then adjust the layer opacity to reduce or (in Photoline) increase the effect. Awesome for those sunset photos

Another advantage when sharpening: only sharpen the L channel, which leaves the colour channels alone, allowing much stronger sharpening/unsharp mask moves with superior results. In RGB both contrast and colour would be affected, resulting in strong halos, while in Lab this will not happen (as quickly).

Knowing your Lab is essential to improve the quality of your work - not knowing or wanting to know Lab is... well, it's just not too smart That's why I will not ever touch any image editing app that cannot work in Lab mode - hence the reason why I have started to enjoy working in Photoline so much, because Photoline's Lab workflow is just far superior to Photoshop.

And ps: why on earth wouldn't you want to work in RAW mode? In the case of this badly shot photo you would have had much more information to play with in the highlights and shadows, which is totally lost in rgb - and what's more, depending on the camera jpg format may save your colour information not even at 1:1:1 colour information (meaning your colour pixels will have less resolution than luminance pixel resolution!), resulting in even worse colour control, and causing potential issues with masking jobs! More detail in the hair, more detail in the eyes, the shadows, the highlights... jpg rgb really limits your options with a bad shot - often turning it into an unusable one. Anyway, that's my two cents.

[Herbert123]

One more thing: I take it you do hardware-calibrate your screen(s)? Without a Spyder, Hue, or something similar colour correction becomes a bit of a pain, to say the *very* least.

[bkh]

And ps: why on earth wouldn't you want to work in RAW mode? In the case of this badly shot photo you would have had much more information to play with in the highlights and shadows, which is totally lost in rgb -

What's so wrong with this photo? In fact, as far as this photo is concerned, I don't see much of an advantage in using Raw. There are no lost highlights in the photo (except the red cnannel in the reflection on the fingernails, but I don't think this needs to be fixed). JPEG w/AdobeRGB retains all the shadow details even of a 14 bit Raw (could even handle 16 bit), and sRGB retains all of 12 bit raw.

and what's more, depending on the camera jpg format may save your colour information not even at 1:1:1 colour information (meaning your colour pixels will have less resolution than luminance pixel resolution!), resulting in even worse colour control, and causing potential issues with masking jobs!

Assuming that you are using a Bayer sensor, the sensor itself captures colour only at half the resolution (and the same is true for raw files). So you don't lose information in your jpeg that way. In addition, the colour resolution of the human eye is also lower than the luminance resolution, so there is nothing wrong with jpegs with lower colour resolution.

Cheers

Burkhard.

[Martin Stricker]

Here is my approach of correcting the image:

retusche.jpg

First of all I adjusted the contrast. Then I removed the skin blemishes. I created a mask to soften the background. Retouching the eyes is really hard, because the picture is so small. On the skin colors I made only small adjustments.

Martin

[Twindaddy]

Here is my approach of correcting the image:

retusche.jpg

First of all I adjusted the contrast. Then I removed the skin blemishes. I created a mask to soften the background. Retouching the eyes is really hard, because the picture is so small. On the skin colors I made only small adjustments.

Martin

*hats off* Great effort. Definitely / easily the best of the shown images (especially regarding the small image to work with)

[greenmorpher]

Hugely informative; thanks all. Herbie, thanks for all the detail on Lab mode; and I might even experiment a little with RAW. Just that I have found the Oly JPEG results generally to be excellent but I have had some problems with color shifts in the hair salon pix I've been doing. At least part of it is lack of control of light; it is all very ad hoc. Also, iPhoto doesn't handle the Oly E-PL3 RAW so I have to import them separately and keep them in a separate catalog. A damned nuisance.

Your comments on the ease PL workflow should be front and center on the PL website and should be raised in every forum. I despair of PL's marketing.

Burkhardt, thanks for your qualifying comments.

Martin that is an excellent result. When you speak of "retouching the eyes", what did you do? Could you explain what you did? I would be most grateful. You’ve got very close to what I would have liked to have, but couldn’t get, so I settled for faking the eyes. True, they work well for posters and a little less well for packaging, but nevertheless, they could be a little more restrained! Like the eyes you produced.

Cheers, geoff

[Martin Stricker]

When you speak of "retouching the eyes", what did you do? Could you explain what you did?

I think it is an well known trick copying parts of the better looking eye to restore the worse one and so did I. The copied part of the right eye where scaled and rotated to fit into the left eye. The eyes where separately adjusted in contrast. I also removed the reddish color. At least I wanted to get some more light into the eyes. So I created a vector layer with a circular gradient from white to transparent. The vector layer was converted to a pixel layer. This way I could transform the circle to an ellipse. The effect was much to harsh, so I reduced the opacity of the layer.

Martin

[greenmorpher]

Excellent, Martin. Many thanks.

Cheers, geoff