Photography - hand-holding question.

Factual Questions
1

Long ago, when I was learning photography (back when Ansel Adams and George Eastman still trodded the earth), I learned a rule of thumb about hand-holding a shot: You wanted to make sure your shutter speed was as least as fast as the reciprocal of your focal length. That is, if you were using a 50mm lens, you wanted to use a shutter speed of 1/50 of a second or quicker. For a 300 mm long lens, you wanted 1/300 of a second. And so on.

Nowadays, however, many digital cameras use much smaller sensors. Which means that a “35mm equivalent” focal length is actually a much shorter physical focal length. For example, I just bought a camera with a 4.3 to 129 mm zoom lens. But with the smaller image sensor, that is “equivalent” to a 24 to 720 mm focal length in 35mm format.

So, my question is, does the “one over the focal length” rule apply to the ACTUAL focal length, or does it apply to the 35mm equivalent focal length? Or, to put it another way, at my maximum zoom, should I use a shutter speed of faster than 1/129 of a second? Or faster than 1/720 of a second? Obviously that makes a big difference.

(I could make matters even more complicated by bringing in changes to the technology, like the fact that modern cameras are smaller, lighter, and have all sorts of modern image stabilization mechanisms. But I’m trying to ignore those kinds of differences here, and focus on the physics and optics related to the smaller image sensor, for now. But feel free to enlighten me on these other issues as well if you can!)

Thanks!

2

I honestly never considered how the smaller sensor affected that rule of thumb. Combined with the image stabilization and the fact that my digital cameras can have some outrageously high ISO settings, it just isn’t something I usually find myself thinking about most of the time.

3

This page says it applies to the post-crop-factor focal length, so the 35mm equivalent one.

I don’t think it has to do with sensor size per se but rather its effect of field on field of view – at longer focal lengths, even a tiny bit of hand shaking can shift the resulting view quite a lot.

In my own shooting, I find that if image stability is all you care about (it rarely is, but let’s just assume that for this case), putting the camera on aperture priority and opening wide up will give you, in the camera’s judgment, the shutter speed it thinks it can get away with. Look at the picture and see if it’s blurry after the shot. If it is, ramp up the ISO 200-400 at a time.

I think that’s more helpful than old rule of thumb because 1) lightning conditions are so variable and 2) digital review means it’s easier to just tinker with the settings and seeing how they look, zoomed in, rather than trying to calculate ratios in your head.

4

The 1/focal_length rule can be though of like this:

For an average camera mass and human holding it, the distance that the image on the film can move during the exposure time should be less than (or at least of the same order as) the grain size. So, the rule is also dependant upon the grain size. When shooting silly fine grain film (Tech Pan, Kodachrome 25) you really needed a more extreme rule. When shooting a fast grainy film you could relax it. So the rule really worked for “average” film and grain. Expressed like this you can see that the size of the film makes no difference. You could be shooting on 125 or 35mm or 6cm film, and the rule stays the same. The rule is also influenced by the final size of the printed image. Don’t expect the rule to work if you print A3 or bigger pictures.

So, for image sensors the same idea holds. You don’t want the image on the sensor to move more than something of the order of the pixel size. For image sensors this does get a little more messy. The Bayer interleaving of individual sensors means there is no easily defined pixel size, and the effective grain is a function of the noise in the sensor - which is a function of the gain - which is set by the ISO setting. There is a neat duality between film grain and its ISO rating and image sensors and their noise and ISO rating. In the end it all works out very similar in terms of information capture. That said, the quantum conversion efficiency of modern backlit sensors is remarkable, and in general you can say that digital sensors outperform film in information capture for a given area.

Different cameras have different sensor densities. There was an unfortunate marketing driven race to deliver more and more pixels in cameras, ending up with cameras where the high pixel density was only meaningful under very stringent conditions, and where most pictures are noise limited long before being resolution limited. You can see pro-level full frame cameras with the same (or even fewer) pixels than cheap point and shoot cameras with relatively tiny sensors. The larger sensor has vastly better low light and noise performance.

So, the rule of thumb. In principle, a sensor with higher density should have a more stringent rule than a sensor with a lower density, but the increased noise of the smaller sensor may tend to make it perform little to no better than the low density sensor in most conditions. But when the light is good, the denser sensor will deliver, and a more stringent rule might be useful.

In general, the ISO setting does provide a guide to the noise, in the same way as it does for grain. So, the rule of thumb can be - for “ordinary” ISO values, and for “average” sized printed results, 1/focal_length < shutter, speed is still probably a good start. A more stringent rule is justified in when the ISO is low, and less when the ISO is high. This is independent of the sensor size. So the focal length is the real, not effective relative to 35mm format length.

5

I agree that viewing actual results is better than following various rules of thumb, but…

They are still useful, especially when you don’t have time to do all of the checking. Yesterday I was shooting with a 90mm (equiv) lens, trying to capture certain moments as people were moving around. I took a shot, noticed that the shutter seemed a bit sluggish, and then glanced at the shutter speed the camera had chosen: around 1/20 sec. Rather than waste the shots wondering if I could hold the camera steady enough, I bumped the ISO up a few notches and all was good.

In my opinion, the focal-length rule is most useful when using a more traditional camera with prime lenses and no image stabilization. I have three good primes, so it’s easy for me to remember that with my 21mm, 50mm, and 90mm lenses I can get away with 1/20s, 1/50s, and 1/90s respectively, and if I see the number dip below those I had better be extra careful with my camera holding technique.

If I were to switch to a zoom lens and then throw in image stabilization the rule becomes a bit less convenient to work out, I probably would just look at a few shots.

Unfortunately, I have never had a camera that had a quick “show 1:1 magnification” button in order to verify focus. For some reason camera manufacturers don’t think that’s important, so to check for camera shake I have to view the photo and either click “+/-” buttons or scroll a wheel every single time in order to get the photo to max magnification.