4

I have a dog that is notoriously difficult to get good photographs of. This is because he has a dark facial mask, some lighter tones on his face and particularly sides, and nearly white legs. So essentially regardless of how I meter, some part of him in the picture ends up either over- or under-exposed. The camera is a Canon EOS 50D.

I know that I can use the flash on low power (either using the external flash with the main reflector pointed in some oddball direction, and the secondary reflector set to low power; or the built-in flash setting the flash exposure compensation negative), but short of that, does anyone have any advice for this type of photography? I'm shooting exclusively in raw format and most often handheld in outdoor settings with little or no time to prepare a shot.

What I do is try to expose towards the high end of the histogram, and then later on adjust the exposure and curves to try to bring out some detail in the darker areas during postprocessing. However, this is notoriously difficult to get right, particularly while preserving natural colors, and still depends on a decent original shot. I am aware that high-contrast photography puts most cameras to the test, but what can I do to at least increase my chances of success in such situations?

For some examples, see the last picture here (this particular one was taken with another camera - a Canon PowerShot S50 - but still illustrates the general problem fairly well), or the third picture from the top here. The ultimate goal would be to have at least some definition throughout the subject, in both low-key and high-key areas.

user
  • 1,099
  • 1
  • 14
  • 22
  • I don't see how incorporation of a flash can reduce the dynamic range of the subject itself - unless you have an array of narrow beam sources pointed specifically at the dark areas of the dog. – ysap Jul 27 '11 at 12:33
  • It doesn't affect the dynamic range of the subject itself, obviously, but illuminating the subject does have an effect on the apparent dynamic range of the shot in many (though not all) lighting conditions. – user Jul 27 '11 at 12:52
  • Michael - you described the problem as the subject being a black/white object. It does not matter how much (uniform) light you shed on it, its dynamic range will be the same. If it is greater that the camera's capture ability, then you have a problem no matter what. – ysap Jul 27 '11 at 14:26
  • We had a discussion about the subject in the comments to this question: http://photo.stackexchange.com/questions/13878/how-can-i-handle-two-extremely-different-brightnesses-in-the-same-picture/13879#13879 – ysap Jul 27 '11 at 14:34
  • That "third picture from top" is actually a good example of what not to do - you seem to capture the dog exactly when there is a spot of light on his white patch, while the rest is in the shadow... – ysap Jul 27 '11 at 14:37
  • @ysap, That is an interesting observation, thank you. Unfortunately, at that particular moment, to get anything even resembling that shot, I was pretty much stuck with that angle relative to the sun. That's no reason to not try to keep it in mind in the future, though! (Incidentally, that picture is also one example of where I think using a flash at low power might have helped.) – user Jul 28 '11 at 07:50
  • By all means, I did not mean to say that the shot shouldn't have been taken. Just an inverse example of the principle of selectively using light to fill-up the dark areas. That said, unless you use a flash with narrow beam pointed at your dog's head (for example) then the flash will increase lighting of the dark and the bright areas - thus preserving the total dynamic range (difference) of the dog. – ysap Jul 28 '11 at 08:57
  • The truth is that this particular dog does not seem to be that big of a problem. Exposing for the head (with little underexposure), there are relatively small bright areas that may get into saturation. But most of it will be fine. When first reading your question I had the image in my mind of a coal-black dog with big white spots, which will be much more difficult. – ysap Jul 28 '11 at 08:58
  • @ysap, the way I figure it is that the low-power fill flash will add only a little illumination compared to the already-lit (by the sun) parts of the subject, but in shaded areas the amount of additional light will be relatively substantial relative to the total amount of light in those areas. That should lower the subject's apparent dynamic range and help bring out details in the darker areas. This might warrant a related question of its own, but am I completely mistaken in my reasoning? – user Jul 28 '11 at 10:59
  • You actually make my doubt my reasoning... The point is that the light is additive. If you add a stop of light to the dark area, it also adds a stop to the bright area, and the total difference remains the same. The one thing that I am not 100% sure now is whether the addition can be done in the log space (stop) or linear space (and then your assumption may be right, but it does not matter if the fill light is weak or strong). – ysap Jul 28 '11 at 12:25
  • I posted another question on this discussion; see http://photo.stackexchange.com/questions/14424 – user Jul 28 '11 at 14:28
  • @ysap addition should always be done in the linear (non log) space, otherwise it's multiplication, not addition! For the less mathematically inclined, adding a stop means multiplying by two, so adding a constant amount of light does not add a stop to both the highlights and shadows, as adding a stop to the highlights (doubling them) would require more extra light than adding a stop to the shadows. – Matt Grum Jul 28 '11 at 16:51
  • @Matt - you are right, of course, that adding in the log space is equivalent to multiplying in the linear space. However, the way this matter is perceived in photography is that everything is done in log space, hence we can shift exposure by one stop... which is actually doubling the brightness. When you have one light source, then things are pretty obvious. If the subject has two areas with different reflectance, Ra, Rb, and the source has intensity S, then the recorder brightness will be: [...] – ysap Jul 28 '11 at 16:59
  • [...]: SRa, SRb, and the DR is Ra/Rb. If we double the light, then again the DR is Ra/Rb. That was my point. Now in the OP's case, the source is actually a superposition of two lights Sa, Sb. Now, the brightness recorded is (Sa+Sb)Ra, (Sa+Sb)Rb and if we double Sb we get (Sa+2Sb)Ra, (Sa+2Sb)Rb. Now the DR is changed (which is what made me doubt). So, I guess the correct reasoning is to say that in order to be effective, the fill flash has to be strong enough in order to decrease the brightness gap (and the DR). – ysap Jul 28 '11 at 17:04

2 Answers2

2

If you don't have time to take readings and set manual exposure maybe you can do something with exposure compensation.

Take multiple pictures of the dog with the settings (practically metering mode) you normally use but with different exposure compensation values (+1/3,-1/3,+2/3,-2/3 ...) - the composition is not important but it should be similar enough to the pictures you usually take so the camera metering is the same.

Load all those pictures into your computer and see what EC value gives you the best result, always use this when taking pictures of the dog

I don't believe this will give you the optimal results - but it should be close and it's fast and doesn't require taking readings and thinking about camera settings.

Nir
  • 20,825
  • 4
  • 38
  • 74
0

Personally, I'd take separate readings from both the light and dark patches of his coat, and normalise them so they are equally distant from the middle of your exposure meter. For example, if the camera says his white patches are 4 stops lighter than his dark ones, then meter so that the white patches show overexposure by 2 stops, and and the dark ones under by 2 stops. The closer the areas are to the middle of the scale, the more detail will be recorded.

Use spot metering on your camera and take readings from the part of your dog that are lit by your primary light source. That way, if you're shooting in manual mode, once you've set the exposure it will continue to work as long as the lighting remains consistent.

  • This assumes two things. One, it seems to me, that I don't shoot much else during a session (true at times, utterly wrong at others). Two, and more importantly, that I have the time to take detailed measurements, which "most often ... with little or no time to prepare a shot" I rarely do. – user Jul 27 '11 at 12:12
  • 1
    If your priority is to get a good shot of your dog it may be worth taking time to prepare; I don't think a couple of spot meter readings would take that long. However if you are desperately short on time, using grass as a reference point to set your exposure (grass usually equates to 18% grey. i.e. the middle of your meter) will help to keep the range of shades on his coat to a manageable range. –  Jul 27 '11 at 12:22
  • 2
    Michael - It also assumes a third thing, that the subject's own dynamic range is within the available range of the camera. This may not be true for black/white subjects. – ysap Jul 27 '11 at 12:31
  • Excellent point, ysap. Michael - what model of camera do you use? May help if we understand the dynamic range it can cope with. –  Jul 27 '11 at 12:47
  • @Nick, ysap: Canon EOS 50D. Edited the question to incorporate that. – user Jul 27 '11 at 12:53
  • Thanks. That's a pretty new model so should have a decent DR. I can't really suggest anything other than what's already been said, I'm afraid. –  Jul 27 '11 at 12:56
  • @Nick, I edited the question again to point to a few examples showing what I'm talking about. Maybe that will help. – user Jul 27 '11 at 13:02