|
When Red Flowers Turn Into Red Blobs
|
A red rose complete with burned out petal tips (more obvious at full size). The meter said the exposure was fine.
|
|
If anything, the composite (luminosity) histogram looks somewhat underexposed
|
|
The red histogram shows the problem
|
|
The green histogram looks severly underexposed. Very little green in a red rose.
|
|
The blue histogram seems underexposed too. This makes sense since the composite histogram is an average of the overexposed red together with the green and blue.
|
Most camera owners will at some point take pictures of flowers. After all, they make a very photogenic subject. Pictures of red flowers though sometimes don't come out very well, looking somewhat oversaturated with little detail in some areas. Understanding why can help you make better photos in the future. It's all about how a digital camera works.
Whether you shoot in raw mode or not, and even if your camera doesn't claim to support raw capture, all digital camera images are shot in raw mode. That's all any digital camera is capable of. What happens next depends on the settings made on your camera. If you have a choice of shooting raw and have the camera set that way, your raw image files will get saved to the camera memory card so you can upload them to your computer later. If you are not saving the raw files, your camera will internally convert each raw capture to jpeg format and save that instead. Either way though, you always start with raw.
A raw file consists of data captured by millions of individual red, green and blue photosites (pixels). By interpolating and blending these later, the camera creates the appearance of the full spectrum of colors. Your computer monitor performs a similar trick or creating millions of colors from just red, green and blue pixels. The RGB color model commonly used in Photoshop works along the same lines too, separating colors into separate red, green and blue channels. Different hues and shades contain different amounts of these three constituent primary colors. Neutral tones such as gray or white contain equal amounts of red, green and blue, but most colors have are more lopsided in their makeup.
The lowest value possible for any red, green or blue pixel is of course zero, a situation that occurs if no light of that color falls on that photosite when the image was captured. And just as there's a minimum, there's also a maximum possible value, beyond which values get clipped. In standard 8-bits color, you probably already know that this maximum is 255. If your camera tries to record anything about this, you'll end up with 255 instead. If enough pixels in any part of an image end up clipped at 255, you'll see a burned out, overexposed area there with little detail.
All this probably makes perfect sense, but I can here you saying to yourself that you didn't overexpose that picture of a red flower, so that can't possibly be the reason it came out looking more like a red blob than a flower.
Ahh, but it's not the overall exposure that matters, it's the exposure of each individual channel. If you take a picture of a red rose as I have done here, the raw data will obviously contain relatively high values for each red pixel, and relatively low values for green and blue, colors on the opposite side of the color wheel from red. If you average the red, green and blue values you may think your image is perfectly exposed while all the time your red channel is at least partly burned out and clipped. Most images have enough variety of different colors to make this not a problem, but an image that is strongly one color can give you a misleading reading. The camera's meter is telling the truth, but it's the exposure of the individual channels that is important for digital photography.
In addition to the meter, most digital cameras also allow you to see the histogram for an image to check your exposure, but unless it lets you see the individual red, green and blue channel histograms, you're not seeing the whole picture. When you average out the red, green and blue channels to form a composite "luminance channel" histogram, you end up with the same problem that the camera meter suffers from. It's the individual channels that matter, and if you burn out the red channel it doesn't matter if you nailed the exposure when averaging in the other two channels as well.
If you can set your camera to show you the histograms for the red, green and blue channel separately, this is your best solution. You'll be able to tell if you've clipped any of the channels by checking if any of their histograms is bunched up against the right hand end of the graph. Histograms that are clipped on the left hand end may be perfectly acceptable as long you expect to see shadows, but a histogram that is clipped on the highlights is pretty much never a good thing. Even if your meter says an image is OK, a clipped histogram for any channel still means it's overexposed. To fix things, you'll need to underexpose enough to avoid the problem.
If your camera doesn't let you see individual channel histograms, you'll still need to be careful with your exposure to avoid clipping a channel, even if you have to do so by guessing. There aren't that many situations where this will be a problem, and you will soon enough learn those subjects that cause your camera difficulties. The most likely case is images such as this rose that are strongly red. Highly saturated green and blue subjects are surprisingly rare in nature. Green grass and blue skies might be a problem, but these generally aren't as vibrantly saturated as your typical red flower is.
|
