Week 2: Exercise 2

[mikemeister_admin]

I tried this image two different ways; firstly, a Lab correction of the original; secondly, before opening CurveMeister, I applied a false profile (gamma 1.4) to the image, converted it into Lab, duplicated the background layer, and set the duplicate to 'Multiply'* - and then corrected this in CM (again, in Lab).

The second approach is something Dan Margulis has been doing in the videos I have been watching, and I wanted to compare this approach with the results I might get from going directly to CM. I have uploaded comparison images, so you can judge for yourself which version is better. The Lab curves were not identical, so it is perhaps an unfair comparison, but in both sets of curves I tried to do the best I could for the image.

I generally preferred the look of the premultiplied version - but the blues in a few of the children's drawings looked unrealistic (I've cropped a bit of the photo for comparison). I fixed this by converting the image into Lab, and using the 'Blend If: B' slider to revert to the blues from the normal Lab correction.

This is more colourful than I usually make my images - perhaps watching Dan M has affected my colour judgement!

Cheers,
Lee.

*I took all of the new colour from this layer, but only half of the luminosity...

[mikemeister_admin]

Whoops - forgot to attach the comparison image (that I mentioned in relation to the Blend If move).

Here it is:

[ggroess]

Lee you might be over thinking this one...

The CM Lab correction you posted is a touch hot in the highlights but otherwise it appears to be pretty good.
I'd leave the flash flare alone...It is a good test of your skills but the type and color is more important in this image...

Again it is the "significant" shadows and highlights that we are after...if there is no tonality in the highlights it can be a problem but for this image we are after getting the type clear and the colors bright...

Greg

[leeharper_admin]

Thanks again for the advice and the curves  ;)

I just loaded your curves and compared the results to my versions. You're right - I should have been working on making the text more legible (something that your version does, whereas all of my corrections obliterate the text!)

I think that watching those Dan Margulis videos has had the effect of making me over analyse everything! Also, Margulis goes crazy with his colour boosting - making his corrections much more saturated than I usually think looks right, but I attempted this image after watching him correct files and that tendency has obviously messed up my judgement! It's really useful having these things pointed out as they are easy for me to miss.

Thanks again,
Lee.

[ggroess]

One of the things that many people miss about Dan M's work is that most of it is geared toward final printed output. 

Dan really wants you to have good saturated colors but he really does not define that...he cannot...it is completely subjective.  As time goes on and more of the images go on-line rather than printed Dan has had to adjust his saturation "ideals" just like the rest of us...most people watch his videos and look for the "magic" answer.."he must have some trick I can steal to make my images look good".  He really does not.he is patient highly skilled and very experienced. he leverages everything he knows about color and imaging every time he makes a correction.  His experience pool is deep and the "how does he know that" type answers come from the very deep end of that pool. ..after you use CM a while you will begin to see that some of the things he does, you can do much faster and with fewer segments because you can switch color modes in CM without changing the profile of the image back and forth to get at the curves.

Remember that the end result should drive the correction decisions.  I would make a different correction for printing that I make for on-line viewing...The screen is backlit and the print is light reflected from a piece paper with pigments on it...how white is the paper??  How bright is the monitor??

Lastly your screen cannot possible show you all the information in an image file.  The screen is at about 75 to150 dots...depending on the quality of the screen; your printer can do about 300 dots...that is why the threshold tools are so important...visually the screen clips the highs and lows...the threshold tools show you where the real values are and if they are out of Gamut for the color space you are using.

We need to have a pint and a good hearty discussion...
Greg