Setting Neutrals

[mikemeister_admin]

Hi Greg,

I've just started reading through the solutions for week 2, and in the exercise 2 discussion (about the children's drawings) the following has me a little confused:

Because it's generally more useful to have the shadow near the midtone, we generally start off by looking for something that we know has no color, and it about midway between the shadow and highlight. But all we have is white paper, black ink, and crayons, none of which we can count on as 50 percent gray.

This, and the subsequent text suggest that - given the opportunity - neutrals should be set on areas with a luminosity value of 50. The engineers of Adobe Camera Raw recommend setting neutrals with a brighter sample (the second lightest neutral square on a Gretag Macbeth chart). They say that raw files contain more information at this point, so their recommendation seems to be specific to raw files (so I suppose that there is no reason to suppose that the same would hold true in CurveMeister), but it did make me wonder.

If I am reading the solution text correctly, and 50% gray is preferable for defining neutral points, why is this the case?

Lee  :)

[ggroess]

The neutrals issue always seems to confuse people....
A neutral is not 50% gray. 

A neutral is equal parts of all the colors used in the color space.  for RGB it is 0,0,0 or  255,255,255 or even 100,100,100  in LAB it is always XXX,0,0 it matters very little what the actual values are.  When we get to color by the numbers you will see that you need all of those examples to make it work.  You need a highlight , shadow and mid-tone marker to correct any given image. 

"They" suggest 50% gray so that you have pixel values to work with; so that you are not working with a blown out highlight or a locked up shadow.  It is also the easiest to illustrate; it's Gray....The Sample size is also important at this point.  If your sample is 3X3 you are really only looking at 9 pixels to set the neutral for the entire image. 50% would have a higher than average number of neutral pixels.  The only way of having a truly Gray value in your image is to shoot a known gray card under the lighting you are using.  If the lighting changes so does the neutral. 

So what do we do??  We approximate.  We use parts of the image that we know intuitively are supposed to be neutral...with simultaneous contrast we can get away with a lot...if you go completely numeric you get locked into a big struggle with the values...too much time and effort from my position.  You can get all calibrationist if you need to... but I simply calibrate my monitor every 2 weeks or so with a I-one2 and try to not over stress it too much.

It is always "preferable" to have a 50% gray just as it is always preferable to have a $100 over $50.  The reality is a 50% gray is a good find...but not a show stopper...one of the best uses of CM is to shoot a gray card if you are doing a batch of images in a set location. 

Let's say I am shooting a studio portrait.  I would shoot a gray card with the model as my first image.  This gives me a known Gray in an image under the lighting I am working with...1000%.  I correct the image to the gray card, bump the saturation as needed, Save the ACV and apply that ACV file to every image I shoot in that session; and I quickly correct every image in the set.  I'm consistent, fast and accurate...Life is good....

In your quote you are forgetting White and Black can be neutrals...I'll have to go watch the solution again...sometimes Mike takes the long way to the description....

I hope I answered you r question...I feel like I just stepped off the soap box....
Greg

[leeharper_admin]

It is always "preferable" to have a 50% gray just as it is always preferable to have a $100 over $50.

I almost fell out of my chair when I read that  ;D

I think perhaps I didn't articulate my question very well - I know that blacks, whites and grays are all neutrals. It's just that the video suggested there was something special about finding a region of the image - that should be pinned to neutral - with an L value of 50 in the hue clock. My understanding of RGB and Lab would make be believe that trying to find something of that particular brightness level would not contribute anything in terms of quality over say an area with an L value of 25. I'm just wondering if I'm missing something  ???

Lee.

[ggroess]

I don't think you are missing anything..these are great questions and I am serious about the answers.. ;D  50% gray is a great find...but without knowing what that is you are guessing...experience and common sense are your best tools...

For instance...there is an image coming up with a street scene and drummers...there are lot's of "target" grays in the image but which one is going to give you the best solution??  As I am sure you know 50 on the LAB scale is great..but then find an object actually at 0,0 from there...

Personally I don't like magic...there's always a trick and it is rarely real...

[leeharper_admin]

Hi Greg,

I'm working on my fashion shoot images today, and am finding your SHN Skin Tone pins absolutely invaluable. It has got me thinking though, and I wonder if you could advise me. The CM wizard does not let users set both a skin tone pin and a neutral. Why is this? And by extension, while I'm working on these files (I'm working in Lab) is it advisable for me to not set a neutral?
If so, does this apply in all color modes - or should I be working on these files in a different color mode?

Thanks for all the great help you're giving me - I really appreciate it  :)
Lee.

[ggroess]

Lab does not allow more than 1 neutral. 

A skin tone pin in LAB is treated by the software the same as  a neutral.  The reason is that neutral in lab is LAB  In RGB you can have multiple neutrals based on brightness or multiple pins based on brightness.  Do the heavy lifting in LAB and fine tune the skin in RGB if possible.

Greg