Color Channel Clipping


Today’s Question: I read your response to the question of “clipping” [in the September 15, 2017 edition of the Ask Tim Grey eNewsletter] and again you outdid yourself in the depth of the explanation. The only thing that I did not understand and you only touched on peripherally is the question of clipping on the color channels.

You used the red rose as an example but did not go into how to address clipping of a color channel, only blacks and whites. I have had some specific examples of a photo of a red rose that I cannot get the red color right and I think this may be the issue. Can you address this question as well in the topic of clipping?

Tim’s Quick Answer: Clipping of an individual color channel in a photo can indeed be a problem, primarily from the standpoint of a loss of texture and detail in areas of the image that are dominated by a single color.

More Detail: When we talk about “clipping” in the context of a photographic image, we are typically referring to the white point and the black point. If the whites are clipped than some of the bright areas of the image will become pure white, losing all detail. Similarly, if the blacks are clipped then the dark shadow areas of the image may lose all detail. This is obviously an important thing to keep in mind when adjusting the white and black point for an image, so that detail is not sacrificed in areas where you want to retain that detail.

Similarly, there is an issue as it relates to individual color channels. A red rose is a good example. If we assume a photo that is a close-up of a red rose, in theory the image might be comprised of information that is only present on the red channel, with very little information on the green or blue channels. In actual fact the distribution of values is not quite so extreme, but I think it can be helpful to consider this theoretical example.

Remember that clipping of the white point requires that all three color channels reach the maximum brightness level for specific pixel values. In other words, you can’t produce a pure white unless all three color channels (red, green, and blue) have a white value for a given pixel.

Of course, if you think about a red rose, you can imagine that there would not be any pixels that are pure white within the image. That might lead you to assume that it is impossible to clip any detail, since you can’t easily produce a pure white value. You might therefore assume that you will retain texture in all areas of the image. But this is not necessarily true.

Just as you can over-expose an image so much that you lose information on all three channels, so too can you over-expose (or over-adjust) so that detail is lost on one or two (but not all three) channels.

What this over-exposure (or over-adjustment) would translate to is not so much an issue of accuracy of color, but rather of texture in colored areas of the photo. This is why I highly recommend evaluating a histogram on your camera that includes the individual color channels, if your camera offers this feature (and most cameras now do include this option). Furthermore, when applying adjustments to overall brightness levels (such as with Exposure, Whites, and Blacks adjustments) it is important to evaluate all of the channels, rather than only the areas that might be clipping to pure white or black.

In other words, when using the clipping preview display available with many adjustments (such as Exposures, Whites, and Blacks in the context of Adobe Camera Raw or Adobe Lightroom), you’ll want to try to avoid clipping of individual color channels if detail is important in the applicable areas of the photo. So when holding the Alt/Option key while applying an adjustment that includes support for the clipping preview display, you’ll want to watch for individual colors that indicate clipping for one or two channels, in addition to black or white areas that indicate clipping to pure black or white.