Performing a Basic Color Makeover (continued)

8 Using the Color Sampler tool, click the white stucco section of the image — the section with RGB values of R=241, G=253, B=251.

The Color Sampler tool records these values in the lower portion of the Info palette (the one showing the label #1).

9 Again using the Color Sampler tool, click the (black) pavement section of the image.

The driveway is another neutral or near-neutral area of the image.

Note that the RGB values saved here as Sample #2 are R=52, G=78, B=85. Once again, you see that a supposedly neutral area of your image is indeed not neutral.

You now have two sample points recorded for you in the Info palette — one in the highlight and one near the three-quarter tone. Time to put the info gleaned from these two sample points to good use.

0

0

JjNdHflalor^ Infc

histogram 11

K

241

C :

4%

Jf G

253

4 M

■!%

t

251

Y :

K:

UK

S-bil

8-bit

1 X

2.8M)

— W :

1, y

1.41 5

P H:

#1 R

241

-n

253

251

One: J Q9M/3.22M

a

0

0

1 NavHjiLtw ^ hfc

Histogram

H

50

L

so %

H c

76

4 M

58%

t

83

Y

52%

K

35S

6-bit

8-bit

1 X

2.213

— W

1, T

2-8W

JT H

#1 R

243

#2 R

52

254

78

252

85

Oct.: ItWHJJ 22W

Photoshop Confidential

Neutralization

In my experience, the key to most color-correction decisions and techniques is to understand and work with neutrals— shades of gray. In RGB terms, neutral and gray are identical values. So an image area that measures 128R, 128G, 128B is a neutral-gray midtone value (Note: 128 is midway between 0 and 255). Similarly, an image area measuring 242R, 242G, 242B represents a neutral 5% highlight (0.95 x 255 = 242). And any other shade of gray in an image should also be neutral — that is, have equal RGB values. And since all RGB color images are really constructed as distinct sandwiches of three grayscale channels, all neutral or gray areas of such a color image should have the same grayscale values on all three channels.

Your basic color-correction techniques for neutralization take advantage of this sandwich effect. Suppose, for example, you measure an area that's supposed to be neutral gray (equal RGB values) but it isn't — that is, it has unequal RGB values. Then you know the color is incorrect because it isn't really gray — the image has a color cast. Now, if you make an image-wide correction so that all neutral areas have equal RGB values, then all the other (non-neutral) color areas should be corrected as well. That's easier to get a handle on as you gain more experience using the Info and color-correction tools . . . just remember: Think grayscale!

0 Choose ImageoAdjustmentso Curves to call up the Curves dialog box.

The Curves dialog box duly appears.

! Choose Blue from the Channel drop-down menu in the Curves dialog box.

You want to isolate the Blue channel because your Color Sampler tool has told you that the Blue value needs to be lowered.

@ In the Curves dialog box, click and drag the Highlight (right) end of the curve downward until the Blue value reads approximately 242 in the Sample #1 section of the Info palette.

You are adjusting the Highlight end of the curve because that is the area of the image you are measuring and correcting:

The idea here is to drag the curve downward while monitoring the changes to Sample #1 in the Info palette — so be sure to keep the Info palette open and on-screen.

Taz's Take: The number 242 didn't come out of the blue, as it were. You'll remember that Sample #1 dealt with the white walls — a diffuse highlight that should read approximately 242 in RGB values. Also keep in mind that, if your sample point is slightly different than your initial measuring point, your RGB values may vary slightly.

# Choose Green from the Channel drop-down menu in the Curves dialog box.

RGB stands for Red Green Blue, so you're going to have make sure all three colors are set at approximately 242.

0 0

Post a comment