Week 6 - Example 5 - Temple Bridge

This is the discussion board for the December 2010 CM 101 Class.
dkrassen
Posts: 106
Joined: Sat Dec 04, 2010 12:03 am

Postby dkrassen » Sun Jan 30, 2011 2:09 am

Hello.

I tried this image although I am not satisfied with the correction.  The fence appeared more yellow before the correction; how would I get the fence more yellow again?  Also, the statue lost some detail; how would I bring out the detail in the statue?  Overall, the foliage looks better.  What do you think of this image?

Thank you
David
Attachments
20110129_example5-templebridge-lab1_screenshot1-jpg
20110129_example5-templebridge-lab1_screenshot1-jpg (999 Bytes) Viewed 11744 times

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

Postby ggroess » Sun Jan 30, 2011 9:36 pm

David,
You have two things going on here that are trouble...

1) the shadow end of the L curve is not being set.  This will leave you with a false sense of the tonality of the overall image. 

2) the inverted B channel area that I circled.  This creates an area where the pixels are gray or worse inverted.  They would be more yellow than blue when they are are on the blue side of the curve. 
Lastly, remember that green leaves and grasses are closer to yellow than they are green. Let's try this again and keep the B channel out of the red curve guard area.

Greg
Attachments
screenshot1-jpg
screenshot1-jpg (288.71 KiB) Viewed 11744 times

dkrassen
Posts: 106
Joined: Sat Dec 04, 2010 12:03 am

Postby dkrassen » Sun Jan 30, 2011 10:01 pm

Hello.

You state green leaves and grasses should have a hue clock closer to Yellow than Green.  Is this a general rule of thumb. 

Also,  I know you answered this question before, please answer it again because I am still a bit confused;  A hue clock points to Yellow, however, the RGB value indicates green - explain the relationship between the hue clock and the RGB/Lab values of an actual point?

Thank you
David

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

Postby ggroess » Sun Jan 30, 2011 10:38 pm

The hue clock is not absolute.  As you well know there are thousands of variations that we would perceive as green.

Our perception of the color is a part of the problem. 

The measured values show us one thing but our eyes show us another.  There are many examples of this in the coarse.. 
The green leaves being more yellow is measurable.  Go looking for leaves you consider normal and measure them...in almost all cases you will find they are closer to the yellow mark than the green. 

We live in a world where green leaves become amber and red in the fall, where grass that dies and dries out turns yellow.... 
Yet we always question why something looks better when the yellows and reds are increased.  Ok..only up to a point...but please look at the two shots below.

I will tell you that these flowers were in my back yard and that during the time I was testing, both monitor and printing calibrations....the second version is much closer to the real plants in side by side comparison;  I literally took the prints to the plants and compared...

Shot 1 is right out of the camera RAW converter with all color settings zeroed out. 
Shot 2 is simply the adjustment shown to make the greens more yellow than Green...

Greg
Attachments
screenshot001-jpg-12
screenshot001-jpg-12 (165.78 KiB) Viewed 11744 times
screenshot002-jpg-9
screenshot002-jpg-9 (164.5 KiB) Viewed 11744 times

dkrassen
Posts: 106
Joined: Sat Dec 04, 2010 12:03 am

Postby dkrassen » Mon Jan 31, 2011 3:05 am

Hello.

Thank you for the explanation.  I am still looking for a quantifiable relationship between the Hue Clock and the RGB/Lab values.  Please explain this relationship in terms of CM.

Thank you
David

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

Postby ggroess » Mon Jan 31, 2011 4:50 am

David,
I am looking for a way to get you this information without writing a book.  Let's start here...

http://curvemeister.com/support/curvemeister3/help/Wizard/CSMatrix.htm

Remember that RGB, LAB and HSB are mathematical models for color. 
They can be converted between each other; but there are colors possible in LAB that RGB and HSB cannot create.  When this happens you are "Out of Gamut" and the color will be estimated to the closest allowed color for the space you are converting into. 

In Terms of "space" the largest is LAB followed by RGB and HSB being very close together.  CMYK is a much smaller space because it is limited to colored inks in whatever profile you are using to create the CMYK space.   

The hue clock is pointing to the Hue of the color you are moused over.  In terms of HSB this would be the H value.  It is measured in degrees from 0 to 360.  Both of which are red BTW. This hue value is independent of the saturation or lightness of the color.  The same hue in RGB would have 1 discrete value.  It will be the ratio between the R and G and B values.  So for instance Red having a hue of 0 would be 255,0,0 in RGB.  If we add the Brightness and Sat  from HSB to the equation then we can have a different value in RGB.  So (Hue 0 brightness 50 Sat 100) would be 128, 0,0 in RGB.  In LAB we would have about L 26, a48, b39  Both A and B being Positive makes the color warm...

You can do these yourself by opening a blank document in PS and filling it with pure red.  Then use CM to place hue clocks on the image and change the hue clock displays to read the different color spaces.  See shot 1.  Click on the configuration button on the hue clock to get the fly out menu.

In shot2 you see all the different hue clock displays CM can provide for a color....

I'm not sure this will get the Idea across to you but...Let's keep digging, there is a ton of information on the subject of Hue and RGB.

Greg
Attachments
screenshot001-jpg-11
screenshot001-jpg-11 (73.42 KiB) Viewed 11744 times
screenshot002-jpg-8
screenshot002-jpg-8 (50.32 KiB) Viewed 11744 times

dkrassen
Posts: 106
Joined: Sat Dec 04, 2010 12:03 am

Postby dkrassen » Mon Jan 31, 2011 4:59 pm

Hello.

Thank you for the response.  If I understand correctly, the Hue Clocks color is based on the Hue in the HSB model where as, for example, the RGB color is based on the RGB model.  Now that I understand this much, I will look for additional details on your website and throughout the web.   

Thanks again.
David

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

Postby ggroess » Mon Jan 31, 2011 8:23 pm

You can configure CM to show you the color space you want...you can have it display LAB numbers in RGB mode etc.  It is up to you....

Greg

dkrassen
Posts: 106
Joined: Sat Dec 04, 2010 12:03 am

Postby dkrassen » Mon Jan 31, 2011 11:56 pm

Hello.

I definitely have a better idea of Hue Clocks, but I am still curious about some details.  In particular, I believe what I am interested in is the math behind the Hue Clocks and the conversion of the other color modes to Hue?  If my questions are to specific perhaps you can point me in the direction to find the answers on the web or in a book?

I understand that you stated that the ratio of RGB is the Hue.  I assume you mean  ( R /  G) / B ; Is this correct?  If I understand correctly the H under HSB is the Hue; please explain the relationship between H and color  (What I am asking is for the details for the HSB color model?)  Also, what is the conversion for Lab to Hue?  Given these required conversions, is there a table that matches caversions (ratios in case of RGB ),  to the Hue?    Also,  does such a table and conversions exist for the other color modes?

Under the CM Settings I can set the Hue Clock for Lab (there does not seem to be an option for setting the Hue Clock to RGB, etc).  The Lab setting has  an offset setting; exactly what does this mean?    Additionally, what does Shadow Target and Highlight Target mean?  Explain the four option under Advanced in CM Settings - Preview; (1) use off screen buffer, (2) update during drag, (3) support soft proof and (4) hbit sample conversion.

Thank you
David

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

Postby ggroess » Tue Feb 01, 2011 1:25 am

Well this is the very deep end of the pool.

try this article...
http://en.wikipedia.org/wiki/HSL_color_space

I believe that CM calculates conversion values on the fly; I am not aware of any tables that would be of use.
here are some thoughts from Mike about HSB and the future...I believe the offset is described in this one...
http://www.curvemeister.com/forum/index.php/topic,3097.0.html

Shadow and highlight targets keep you from creating a file with values less than the shadow and greater than the highlight.  in effect you limkit the RGB output to something less than 0 to 255.  Mine ar eset to 5,5,5 and 250,250,250.  This helps me to not clip the highlights or lock the shadows.

Preview using off screen buffer uses the RAM on your PC differently and can speed up or slow down the display.  This was much more important when PC's had RAM issues 5 or so years ago...update during drag is also related to the refresh rate of images in the CM program. 

Support for soft proof is a setting that helps you to see what the print will look like as a proof; it is a great way to see if you are out of gamut when you make color increases.   not sure what the hbit simple conversion does...that might be a better question for Mike.


Return to “Curvemeister 101 - December 2010”

Who is online

Users browsing this forum: No registered users and 8 guests