Recovery?

I want to talk about something I've learned recently and that's about the blinkeys, blown highlights or clipping, and the magical Recovery slider in Adobe Camera Raw (ACR). You know, the blinkeys. Those insipid pixels that blink on the back of your camera when you review an image and the camera wants to let you know that you've gone too far and clipping is occurring.

Up until now, I've thought of the blinkeys and blown highlights or clipping as being truly evil and something to avoid at all possible cost because its lost data. 255, 255, 255, pure white, no detail. Something your printer won't even print out because there's no white ink!

In order to combat this we are armed with the Recovery slider in ACR. With one quick movement to the right we can eliminate said blinkeys and blast those blown highlights back into the realm of subtle detail! But alas, sometimes the Recovery slider falsely leads us astray and alters our images with a flattened dynamic range and bleeds any pop from them. What's left is a distant memory of the zeal your original image contained.

All is not lost. I've come to realize that not all blinkeys and blown highlights are created equal. Some are not as evil as other would lead you to believe. Sometimes you don't need to touch that Recovery slider at all!

Above is an image where highlight clipping is obviously occurring. Thanks to ACR we can clearly see that and said areas are highlighted in a fantastic and not so subtle red. Usually I would be tempted to slide the Recovery slider to the right until I've eliminated all clipping. In many cases, you'll end up with an image now with a flattened dynamic range and a lot less pop in the highlights and bright areas of the image. Since there's a lot of sky in this photograph, using lots of Recovery has an adverse affect on the image. You can compromise some, leave some clipping but recover some of the blown out areas.

So what's going on here?


I started thinking about what these clipped highlights actually are and what effect do they have on my final product and image. I typically share my images on the Internet like many do and then I print the images that I like best.

So, for sharing on the Internet, do the clipped highlights really matter? I don't think they do.

What about printing? Hmm, well, yea, I think they do matter.

Why? Well first off like I said before, there's no white ink, and second, my printer has a limited color gamut so it's best to try to process my images so they match my printer's color gamut as much as possible so screen and print match as closely as possible.

With this new thinking, clipping matters to me when I'm printing. And what my printer can print is what I should really be concerned with. That lead me to my next question and discovery...

What colors are actually present in these clipped regions?

Lucky for us ACR has the Color Sampler Tool (S). Select this tool and then select a point in your image and ACR will tell you what color is there. So I did that and selected nine different points in the clipped regions.

Those nine different points revealed some very interesting information from them.

It would appear that many of these clipped regions are not actually pure white which I thought was the case. It appears that only some of the RGB values are 255. True, there were some that were pure white, 255, 255, 255. But others were really really light shades of pink and magenta.

So I know my printer can't print out pure white but maybe it can print out really really light shades of pink and magenta. To test that theory out I let go of all desire to slide Recovery to the right and just clicked Open Image with reckless abandon! Well, maybe I'm exaggerating just a bit.

In Photoshop, I turned on Proof Colors and loaded up a profile I would typically use to print out an image like this. I also turned on the Gamut Warning. The result on my screen was what you see below...

That's not so bad, right?


Seems like my printer can't handle a few regions of really really bright magenta and pink. The gray areas are where my printer doesn't have a color to match the color in those areas. Seems like my printer does a pretty good job of rendering the colors that ACR and the blinkeys normally would lead me to believe were evil!

In fact, if you overlay the ACR image and the soft-proofed Photoshop version, you can see how well the colors would hold up in a printed image.

The darkened red areas are the intersection of the areas that are out of gamut for my printer and what ACR says has highlight clipping. There are even regions where there are out of gamut colors but ACR didn't claim any highlight clipping there.

Now I have a lot of confidence in my printer and its ability to print this image and match colors so well that perhaps the subtle tone and color differences in these out of gamut areas won't even be noticeable in a print. I could find out by printing and reviewing. That sounds like a good test. I could also play the compromise game and go back to ACR and move the Recovery slider and then later play with Hue, Saturation, and Vibrance and use brushes or masks of some sort to try and bring these out of gamut colors back in gamut.

Sounds like a lot of work. Why not just print a test image and see what we see? Maybe things aren't as bad and as evil as we have been lead to believe. Perhaps the print will look great regardless of what all this technology says should not! Perhaps using our eyes and a little bit of faith is all we need instead of examining each and every pixel in our images and being fooled by all the alarms and warnings. Beauty is in the eye of the beholder, isn't it?

So, know what is in your image as far as colors. Know your final product and goal. Know what your printer can handle. Print out a test print. Make adjustments as needed and iterate. Don't just blindly move sliders and alter colors simply because the computer told you there may be a problem. In the end, you'll end up with fewer if any compromises and much better images.