This week I had the honor to be the photo judge for a photo blog called I Heart Faces (http://iheartfaces.blogspot.com/), a blog dedicated to portraiture. This week’s theme was “Wonder” and after sifting through nearly 400 photos in both an adult and kid’s category, the top ten in each category was selected and the winners will be announced on Saturday.
But the blog has a cool feature called “Fix it Friday” where the blog owners post a photo and challenge members to fix it. This week I decided that the “Fix It” would make a good post, so here goes.
Step 1. Convert to sRGB
The photographer was working in Adobe RGB color space. Most professionals prefer to work in Adobe RGB since it supports a wider range of colors (gamut), but if your primarily working on photos for the web or your likely to print your photos at a local mini lab and don’t want to mess with color Management, an sRGB workflow (sRGB is the color space for web browsers) while not allowing the maximum colors, keeps things simple and easy. Think of it in music terms. Adobe RGB is like Vinyl LP albums and sRGB is like CD’s. The keen ear, not offended by the scratches and hiss of an LP will find a fuller sound from the LP’s and the CD’s are clipped a little. But most people prefer to not have the scratches and hiss in favor of the clean sound.
My photoshop setup detected the Adobe RGB file and gave me the option to convert it when I loaded it. If I left it Adobe RGB and you looked at it in your browser the colors would look muted. Since this is heading to the web, sRGB conversion was a no-brainer.
Step 2. Crop to 8×10
You can see in the original, that the photo is centered horizontally, and there is a little space above the subject’s head. Not caring for the empty or negative space on either side of the subject, I decided that a vertical crop was in order. I couldn’t reasonably get the photo to a rule of thirds crop and keep it horizontal and I’m not sure I would have liked it anyway since it would have just shifted the negative space from one side to the other. I also like to crop to print sizes, 4×6, 5×7 or 8×10 since I have to assume at some point I will want to print the photo. For this one, an 8×10 crop let me crop the photo fairly tightly. This put the subjects eyes on a rule-of-thirds line.
Step 3. Blemish Removal
Using the Rubber Stamp or Clone tool, I cleaned up a couple of blemishes and cleaned up some other skin coloration issues.
Step 4. Bright Eyes – Dodged whites, overlay layer – iris lightening 50%
Next, I attacked the dark eyes. Using the Dodge tool (12% midtones, soft brush a little larger than the area to lighten) I lightened the whites a bit. Then it was time to tackle the irises. To solve this, I add a new layer above the background, set the blend mode to “Overlay” and check the “Fill with 50% gray” check box. Then pick a paint brush a little smaller than the iris and with the color white, paint on the overlay layer around the iris’s making sure to stay away from the outer and inner edges of the iris. This will typically be too bright and sometimes can have an odd color shift. So I typically set the overlay layer to 50% opacity to not go overboard with the eyes. You can adjust to taste.
Step 5. Flatten
Then flatten the image. I find that the overlay layer gets in the way for future actions.
Step 6. Equalized (3 channel white balance)
Looking at the Levels tool, the histogram shows that the tones don’t make it to black, nor do they make it to white. An easy fix, called equalization stretches the tones to range from black to white giving the image a little more contrast. But in this case, I took a peak at the histogram for each color channel, Red, Green and Blue (you can pick the color channel from a drop down on the layers tool (CTRL-L or Apple-L) or you can use the short cuts CTRL-1 for the Red channel, CTRL-2 for the Green and CTRL-3 for the Blue channel. CTRL-~ gets you back to the combined RGB view.
Well the Red and Green channels needed more stretching than the Blue channel, so using the White point dropper I visited each channel, dragging the dropper from the right side of the histogram box until it touched where the histogram bottomed out on the light side. This is a very quick color correcting white balance method, presuming there are whites in the photo! I then went back to the combined channel view and slid the black point dropper in until it met the histogram. This did wonders for the color.
Now I have a good photo, suitable for printing or uploading to the web. But lets take it further and see what we can make out of it.
Step 7. Screen layer 50%
I was in a particularly high-key mood today, so I decided to lighten the image. After playing with some contrast curves and not liking what I saw, I felt it was time to try this tried and true method. Duplicate your background layer and set the blend mode on the new layer to Screen. This is like over exposing the photo by a full stop and it doesn’t blow out highlights. In this case, it was still a bit too much. I used the opacity to adjust to taste, and as it worked out 50% made me happy.
Step 8. Alabaster Portrait Action from Action Central.
Next I tried a couple of canned actions that I have. First up was Urban acid, so working on a duplicate copy of the image, the Urban acid script ran and the results were pretty good, but I still wasn’t feeling it. I went to another script I have called “Alabaster Portrait” that I got from Action Central (http://www.atncentral.com). And after running that script, I was very happy with my results.