Model Railroad Photography III - Cameras


Once upon a time, I thought I understood layout photography. Throw a bunch of light, point the camera, and take a picture. I seem to know less now than I did then. And while that’s probably a good sign that I’m learning, I do feel like I’m going backwards.

My current concern is depth of field. Without getting into the technical definition, that’s the extent of the region in a photograph that appears to be acceptably in focus. While it would be nice to have the entire image in focus, typically either the nearest or furthest-away portions will be somewhat out of focus. In the photo above, the wooden ruler close to the camera is out of focus, as is the far end of the red ruler, so my depth of field here is somewhat less than 12” (30 cm). And that’s viewed from a distance as a ~660 pixel-wide image. Seen in larger form (e.g., the 800 pixel versions I post in my photo album), the depth of field should appear even more shallow.

The usual method for added depth-of-field is to increase the f-stop. As you do, the size of the opening in the camera’s lens gets smaller, and less light is allowed in, which means that the required exposure time goes up. Typically, layout photography requires long exposures and use of a tripod. I’ve always cheated on that, accepting shallow depth of field in exchange for hand-holding the camera. Although I do use a tripod sometimes, I still operate at exposures where I don’t need to be concerned with vibration even then, typically 1/15 of a second or less.

But about two months ago, I bought a new camera, and depth of field has been more of a problem since then. I was somewhat aware of this, and was aware of the cause, but hadn’t really spent much time thinking on it. The problem is that my old camera, a Canon 40D, uses an “APS-C” sensor that’s about 15 x 22 mm in size. The new camera, a Canon 5D mark III, uses a “full frame” sensor that’s 24 x 36 mm in size. I bought the camera for outdoor use (landscape photography is another hobby of mine) but had also planned to use it on the layout since it has better low-light characteristics than the old one. But while that’s true, it’s largely offset by the need to increase the f-stop to regain the lost depth of field.

I wanted to quantify just what the extent of the problem was, and what that meant for me. First I went off and read about depth of field, and played around with the math trying to recreate the results described there, and having done that applied the results to some typical camera sensors (you can read the summary of that on my Depth of Field page). Then I took some photos to explore this some more.

The photo above is one of these, showing the effect of my new camera focused about 20” away from the sensor on the near corner of the house to the right of the red ruler, about in the middle of the ruler. As you can see, both ends of the ruler are slightly out of focus.

This image was taken with my lens at a focal length of 50mm, at f/11, which is about the largest f-stop you can set before image quality begins to degrade due to diffraction effects. Now in practice, those won’t be all that obvious, and I could probably push this to f/13 or even f/16. But at f/11 I’m also at the point where I need to set the camera to ISO 1600 to get a reasonable shutter speed without using a more stable support (I was hand-holding the camera braced atop one of my bridges to take these shots). Going to a higher ISO will add visible noise, so I need to either add more light or go to a slower shutter speed. And I already have a fairly large amount of light on the layout, although this area doesn’t have the new lights yet, so I will eventually add more. Still, I’m pretty close to the limit.

What this tells me is that either I need to change my methodology, to make use of some kind of mini tripod (which I’ve tried before and didn’t like) and long exposures, or I need to use a camera with a smaller sensor. My old 40D had about twice the depth of field, around a foot (30cm) rather than six inches (15 cm). That’s assuming I’ve done the math right, and it seems to roughly agree with some test shots I took.

Stepping down to a smaller camera, such as one of the higher-end “point and shoot” cameras that can offer manual control over focus and exposure, plus RAW image data export, might be an alternative. I’d been thinking of getting some kind of point and shoot for outdoor/travel photography where the DSLR was too much of a nuisance to cart around. I used to have a small digital camera, and although I retired it when I bought the 40D several years ago, I did find it more convenient for some things. Its lack of manual controls however was a serious problem, so I don’t want to get a cheap point-and-shoot. And that means spending some significant cash. I’m not sure I want to do that.

One benefit of a small camera would be taking photos “inside the layout”. The DSLR is really too large to fit most places. It only fits here because it’s sitting on top of the bridge, and the angle is really too high. A photo should look like it was taken from somewhere a human could be standing (or from an airplane / tall building), and this one is just a little too high up for a ground-level look.

So, I’m going to think on this some more. But it’s pretty clear that I need to pay more attention to depth of field than I had in the past. And that light-gathering ability isn’t the only important criteria for layout photography.