How can I handle two extremely different brightnesses in the same picture? [duplicate]

Question

I am the fresh owner of my first DSLR (a Canon EOS 600D, as far as I know called Rebel T3i in the USA). I'm a total newbie, and the camera is a lot smarter than me most of the time. But I'm trying to learn how to choose the settings for certain situations myself (in particular when it comes to lighting) instead of leaving stuff to the electronics.

Now, I understand the theoretical basics of the influence of exposure time, ISO, and aperture size on the resulting image, but knowing the theoretical basics of a combustion engine doesn't enable me to drive in the Indy 500.

So I'm trying to take pictures here and there and looking at results etc., and now I have a situation where I have no idea what to do. The example is probably silly; this is more a "I'd love to know how it works" than a "this is an actual problem I need solved".

I have a ceiling lamp (essentially a tinted glass ball) that has a very nice structure to it and gives the (white) ceiling above it a nice orangy-red glow. Capturing the structure is doable:

(f/5.6, 1/500s, ISO100)

Getting the nice glow is probably possible as well, by finding something between these two and adjusting the white balance:

(1: f/5.6, 1/125s, ISO 800; 2: f/5.6, 1/60s, ISO 3200)

(yes, the aperture is identical among all three images; I also tried slower ones, but didn't get any noticably different results).

So usually I'd say "well yeah, it's impossible; you have a glass structure with a bright light bulb behind it, and a dimly lit ceiling; no way you're getting those two nicely into a single image". But over the past few days since I started trying to learn about photography, I've read up on so many ways to play with how light reaches the sensor, that I wouldn't be too surpised to learn I can actually achieve both.

Can I?

Note that I'm interested in doing this in one shot; I can obviously combine two different images during post-processing, but I'm interested to learn "tricks" to achieve this in a single shot.

Answer 1

Well it depends on how big the difference is between the dark and light areas. Every sensor as a certain dynamic range that it can capture - right now most DSLRs are in the 10-14'ish EV range. Your particular camera can capture 11.5 EV in a single exposure. This is the range you can capture in a single go. This doesn't mean this is the dynamic range of the picture you're seeing rendered in a JPG.

In a high dynamic range scene like a bright light source and dark walls or such, you're very unlikely going to be able to get in all in one shot. You'll have to make a subjective call about what parts of the scene are most important to you creatively and expose properly for those.

Because of the wide DR of many DSLR's you may also be able to take one exposure, digitally push and pull it to create separate shots and combine to form a single shot HDR or Exposure fusion.

Often times in a situation like the hanging light, if you want to expose properly for the light - you'd spot meter off the light and then exposure compensate a little to get the rest of the scene like you like.

Answer 2

@rfusca offered up a superb answer, so I'll only try to add to what he has stated. Given that cameras have limited dynamic range (DR) in relation to the human eye, if you wish to capture all the surfaces in a scene, you will have to either compromise (i.e. expose for one thing and ignore the effect to another), bracket (capture multiple shots at different exposure levels with the intent of merging to HDR), or...add more light.

One thing about light sources is they tend to be considerably more intense than the objects they illuminate. That often causes problems with the limited DR of a camera. One way to compensate is to add more light. This can be done a variety of ways...turning on other lights, enabling the cameras flash, or setting up your own lighting (flash or continuous) that can be used to precisely control the lighting and shading of a scene.

In the case of your example scene, you might simply try using the cameras flash as a fill flash. Enabling flash should illuminate the very dark background enough to capture it with a proper level of detail, while allowing you to reduce exposure enough to capture the detail of the lamp itself. You may need to experiment with the flash power (often called flash exposure compensation) to get the right level of illumination. On-camera flash often creates harsh, dark, and undesirable shadows. You can compensate for that either with an off-camera flash, or some kind of diffusion or bounce modifier. Diffusion is usually done by attaching something semi-translucent to the flash (i.e. a piece of paper, or an actual flash diffuser that you can buy). You can also create bounce modifiers, which are opaque but reflective attachments that redirect the flash light to another surface, which in turn bounces and scatters light to illuminate your scene from a different angle.

When it comes to dynamic range, sometimes all you need is to illuminate the darker parts of a scene and bring them back within range of your camera. It should also be noted that the unattenuated dynamic range of a camera...that 10-14 stops, is not necessarily what you can actually utilize in a final scene. Even when shooting RAW, the safe approach is to assume your camera is only capable of about two stops (two EV) less than its official rating. For a 14 EV camera, you would assume it can capture about 12 attenuated stops (attenuation, or the application of a tone curve, is pretty much always necessary to make a photo look proper...it brings in a necessary factor of contrast). For an 11 EV camera, you should assume that you have about 9 EV worth of usable DR. When shooting RAW, the extra headroom above your assumptions will give you room to work within, minimizing the chances that you blow out highlights or block shadows.

Answer 3

As jrista hinted, saving images as RAW will help because RAW has a greater dynamic range than a compressed jpeg. Using a RAW editor (such as lightroom) you will be able to darken the highlights and boost the lowlights considerably.

It is also possible to use a filter to compress the dynamic range, although I don't know if anyone makes a filter that suits your application.

Answer 4

In addition to using flash/extra light, to balance out the exposures you could also try using lens filters to reduce the power of the light from one part of the image. This is a common problem in situations other than the OP, e.g. dark landscape and bright washed out skys. In this case I'd use a grey graduated filter on the lens to darken the sky to something near the exposure for the land. In the OP's case, perhaps a dark grey spot filter would dim the light enough to balance things.

Answer 5

You could give a try to Technicolor Cinestyle for Canon.

The Technicolor CineStyle provides better dynamic range of the captured content. This allows for greater artistic freedom during shooting (as the look does not have to locked in during shooting) as well as during the color correction process in postproduction. When the Technicolor CineStyle is selected in the camera [...] still images are converted into the same log color space.

What this gives you is a gray-ish desaturated image, that looks flat, and helps you keep detail in the shadows and highlights, like this image from here:

All you have to do is post-process the image and select the range you want to boost and increase the contrast.

To reverse the flatness you have to pass it through a S-curve like this.

To sum it up:

1. Load the profile in your camera using the EOS utility.
2. Use it to shoot high-contrast (wide dynamic range) pictures
3. Achieve the desired effect in post-processing.