The image above has a near-perfect exposure, with its dynamic range all inside the zones I to IX of Ansel Adams’ zone system. Although the ideal is to get perfect exposure straight-of-of-camera (SOOC), sometimes it’s just plain impossible due the very high dynamic range of a scene and the limited dynamic range our cameras’ sensors or film capture (which is the case for the image above – you can see it larger here).
To give you an idea, a very good camera like the EOS 1D Mark III has a dynamic range of 2048:1, while the human eye varies between about 1000:1 up to 16000:1. So a camera sensor alone will NEVER be able to properly capture and reproduce reality SOOC if the dynamic range of the scene is too wide. There are some techniques to get around this issue and I’ll cover them soon. But first, a few concepts.
* Note: the following text requires knowledge of Ansel Adams’ zone system for proper understanding.
What is dynamic range?
In photography, dynamic range is simply the ratio between the largest and the smallest possible values for the luminance (which is, in short terms, the amount of light per unit of area). In practice, we measure it using EV differences (stops) between the brightest (“whitest”) and darkest parts of the image that have some visible detail. In case you don’t know, going up 1EV means that the light amount is doubled – equations here, if you’re interested.
Please pay attention to the underlined text. That means, in the zone system, the dynamic range covers zones from I to IX. And within that, textures are recognizable in zones II to VIII (it’s called the textural range). Zone 0 is pure black and zone X is pure white, which should only cover, respectively, pure darkness and light sources/specular reflections in an image.
For nature/outdoor photographers it’s very important to know these concepts and to be able to control the dynamic range to properly reproduce nature on photos. That can be done in the act (i.e. filters, camera settings, fill flash) or in post-processing.
Getting proper dynamic range – i.e. reproducing nature
The filters we usually use to get the right light are graduated NDs. I already covered those in this post, so I won’t be talking about them. Plus, they can be simulated easily in Photoshop using two layers and a gradient-filled layer mask. FYI, I don’t carry filters anymore, except for those which live on the lenses (UVs for wides and normal lenses, 81A for teles). I simply hate carrying too much garbage around and the digital era means we can get rid of some weight.
For camera settings, most DSLRs today have some type of adaptive dynamic range setting to proper balance exposure. I suggest you leave that on.
As for fill flash, although it’s advisable to use it (i.e. wildlife shots), it isn’t entirely necessary. Sometimes you can just shoot an underexposed RAW at low ISO and the push it up one or two stops in Camera RAW (the same way negatives were pushed up in the past). It may produce a grey-ish image that will require a lot of color processing, it’s true, but it’s better than loosing the shot if you don’t have a flash ready at the right time. I highlighted the word ‘low’ because lower ISOs have higher dynamic ranges, in an analog manner as low ASA films.
From now on, I’ll assume you’re shooting RAW.
Post-processing means HDRI/Tone Mapping and/or different-exposed layers and masking. Your choice here will depend on the dynamic range of the image. If your image has a lot of spots within zone 0, like this, then you will want to skip any HDRI/Tone Mapping processing because you’ll end up with an ugly image with lots of black clipping. In those cases, it’s better to work with multiple layers (usually 2 exposures will do) in Photoshop. A good rule of thumb is to avoid HDRI/Tone Mapping from a single RAW whenever your base EV0 frame purposely doesn’t fit the dynamic range of the sensor (i.e. many areas in zones 0, I and X, like in silhouettes and against the light shots in general). In these cases, you can easily derive one or two under/overexposed frames and work with the ‘fake grad-ND’ I mentioned earlier to get the proper dynamic range. You may also use more selective masking in more complex scenes, if necessary – like this one, for instance.
Images with intended great dynamic range will require multiple exposures. And then, you can either use a HDRI/Tone Mapping software to create your image, or stack and carefully mask your layers to get the correct dynamic range (which is very tedious). Using a HDRI software like Photomatix alone will suck. The results will be unnatural and that’s not the goal with nature/outdoor photography. So what I will recommend is a technique which is a hybrid of the two.
First, try to get a tone mapped image that looks as natural as possible. Always check the histogram to ensure there isn’t any clipping or compression. Once you have the image, open it in Photoshop and also open your EV0 base image as a bottom-level layer. Then, control the opacity slide on the top layer until you get a mix of proper dynamic light and realism. Flatten the image and then finish your processing. You may need additional masking to get the exact dynamic range you’re looking for. I’ll illustrate this with an example.
Wrapping-up: processing 7 exposures into one
Ok, for this example I’ll use this photo of the U.S. Capitol:
That’s my EV0 image, with light metering done on the Capitol. As you can see, I’m loosing a lot of dynamic range on the clouds (zone X) and on the trees in front of the building (zones 0 and I). I’m also losing texture details on the water in the foreground and in the trees on the far left (zone II). In order to correct that, I’m going to use 7 images, spaced by 1/3 EV up to +/- 1EV. In order words, my new dynamic range will extend a full stop up and down in order to correct exposure in those zones 0, I and X spots. Creating the image in Photomatix, I got this unrealistic frame:
Although it looks like a cartoon, the dynamic range is now correct, mostly between zones II and VIII. So it’s a matter of properly balancing between this frame and the original shot. So, I’ll create two layers in Photoshop: the top one will be the tone mapped image, and the bottom one will be my EV0 frame. Then I’ll set the opacity of the top layer until I get the image I want. As a guideline, the optimal opacity is always between 50-70%. In this case, I used 60%. So, here’s the result:
Now that the image is properly exposed, I can work with the filters I would have used if I bothered to carry them to the location. First, I’d have used a grad-ND. You can see that the water is still dark (zone II) in the foreground, due to that shadow/reflection. A 0.6 grad-ND would correct that. So I’ll create another layer, increase the exposure and then use a gradient fill mask to simulate the filter:
And there you go! The image is properly exposed, with most of it within zones II-VIII and bits of it in zones I and IX. And looking natural. You can still further color-process it to pump it up, but that’s the topic of my next post. BTW, the finished image, after color processing, is here.
References and recommended reading
- Ansel Adams, Robert Baker, “1. The Camera”, Little Brown, 1995;
- Ansel Adams, Robert Baker, “2. The Negative”, Little Brown, 1995;
- Ansel Adams, Robert Baker, “3. The Print”, Little Brown, 1995;
- Galen Rowell, “Inner Game of Outdoor Photography”, Norton, 2010;
- Outdoor Photographer magazine, February 2011.