Week 1: A Dark Day in Duck-Land

This is the forum for posting to the June 2010 CM 101 Class
mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Mon Jun 14, 2010 11:18 am

Hopefully this is where Boyd and I are supposed to be posting our assignments - if not, please let me know  ;)

As suggested, I attempted this correction in both Lab and RGB. In this post I'll attach my final versions in each Color Mode. I will post a follow-up message with smaller examples, to show where each decision I made worked well - or poorly.

Of the two (final) versions attached to this post, I strongly prefer the RGB version (in terms of both colour and detailing). I found that the shadow, highlight, and neutral steps were more effective (in this instance) in RGB (again, I'll illustrate why in the next post).

My basic process (for my prefered outcome) was to use the wizard; apply CM a second time* for contrast (applied via the RGB master curve); a third application*, to get at the Lab saturation slider; I finished with an application of Shadow/Highlight, and some sharpening...

More to come - Lee.  :)

*These adjustments were applied through a mask (a curved B channel).
Attachments
duck_rgb_lh-jpg
duck_rgb_lh-jpg (999 Bytes) Viewed 2129 times
duck_lab_lh-jpg
duck_lab_lh-jpg (999 Bytes) Viewed 2129 times

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Mon Jun 14, 2010 1:48 pm

So, I began my correction of this image by working entirely in Lab - so the screenshots attached show my Lab corrections. However, the decisions I made about how I would approach my adjustments were the same when I corrected the image in RGB.

My initial decision (having set endpoints) was where should I set the neutral point. Lab_NeutralPinLocation.jpg demonstrates the effect of moving it around. The image on the left used the metal ring around the duck's foot; the image on the right used the feathers on the underside of the duck. I felt that using the metal ring made the image too warm (and I did not trust it to be completely neutral, because I thought that it could be contaminated by the sky reflecting from the water).

In my Lab correction, I found that my neutral pin had made the dark feathers slightly green (-3/-4 in the A channel). I moved the endpoints of the A and B channels to try to correct this. Unmasked, this adjustment made the whole image too red (Lab_FixFeathers.jpg).

The final adjustment that I want to mention in this post relates to adjusments I made to improve contrast and colour saturation. I did not want these adjustments to be applied as strongly to the greenery as I did to the duck, so I used a curved A channel mask to protect the greenery from the adjustments. MaskComparison.jpg shows the difference between generating the mask in CurveMeister versus generating it in Photoshop. The effect of the CM mask is shown on the left; the effect of the PS mask is on the right. As you can see, the mask generated in Photoshop is much smoother - thanks to the improved blurring options in PS.

In my next post I will compare my Lab adjustments with the corrections I made in RGB.

Cheers,
Lee.
Attachments
lab_neutralpinlocation-jpg
lab_neutralpinlocation-jpg (999 Bytes) Viewed 2129 times
lab_fixfeathers-jpg
lab_fixfeathers-jpg (999 Bytes) Viewed 2129 times
maskcomparison-jpg
maskcomparison-jpg (999 Bytes) Viewed 2129 times

leeharper_admin
Posts: 263
Joined: Tue Aug 20, 2013 2:39 pm

Postby leeharper_admin » Mon Jun 14, 2010 1:55 pm

I should probably attach my two final Lab versions  ::) The difference between the two neutral pin locations is more obvious here...

Lee.
Attachments
lab_finalversions-jpg
lab_finalversions-jpg (999 Bytes) Viewed 2129 times

leeharper_admin
Posts: 263
Joined: Tue Aug 20, 2013 2:39 pm

Postby leeharper_admin » Mon Jun 14, 2010 2:15 pm

Having decided that the Shadow, Highlight, and Neutral pins were more effective in RGB, I then needed to adjust the contrast in the image and boost the saturation.

(The neutral point adjustment worked so well in RGB that the duck's dark feathers did not require further adjustment).

I tried two approaches (see ContrastSaturation.jpg, attached): firstly, a L channel curve followed by a move of the saturation slider; secondly, a RGB master curve adjustment, followed separately by a Lab saturation slider adjustment.

Overall, I prefer the second approach. Most of the duck's feathers look better with the L channel curve, but the image looks more three-dimensional in the RGB version - and consequently, there is a greater feeling of openness in the image. I also prefer the colour (although I may have simply moved the saturation slider a bit further on the second attempt).

Now then, on to the other images  ;)

Cheers,
Lee.
Attachments
contrastsaturation-jpg
contrastsaturation-jpg (999 Bytes) Viewed 2129 times

ggroess
Posts: 5342
Joined: Wed May 24, 2006 2:15 am
Contact:

Postby ggroess » Mon Jun 14, 2010 4:43 pm

This is indeed the way to post the images.

Both of your corrections are really well done.  One of the topics I touch on often in the class is color saturation.  One of the trade offs for LAB is a loss of saturation when you curve the L channel.  This is the primary reason for the Saturation slider.  The secondary reason for the slider is that in PS when you boost the A and B curves you have to fiddle around a bit to make sure they are equal.  In CM the saturation slider moves both curve together the exact same amount; you can then boost the separate channels as needed.

As you have worked through the neutral problem; I see that you are doing a solid comparison between LAB and RGB; in theory they should be the same and based on the same neutral selection.    In fact RGB usually wins for color until you add back some saturation in LAB.  It looked to me that you were comparing the images by color space and not neutral point.

I'll post a link to a short video regarding the subjective choice of neutral.  It's not much but it makes my point.

Really nice work...
Greg

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Mon Jun 14, 2010 4:51 pm

This was background layer set to screen,reeled in with the acv file.I've always had trouble with a neutral on the metal band so I used a rock over at 3 oclock with a=-0,b=-0 in Lab as a neutral.GregM
Attachments
2007-01-23_145313-duck-copy-jpg
2007-01-23_145313-duck-copy-jpg (999 Bytes) Viewed 2129 times
2007-01-23_145313-duck-jpg-lab-acv
(78 Bytes) Downloaded 73 times


Return to “Curvemeister 101 June 2010”

Who is online

Users browsing this forum: No registered users and 1 guest