Where does the ¹/shutter speed = focal length rule for hand shake come from?

Question

The generally accepted rule of thumb is that the shutter speed must be the same or larger than the inverse of the focal length.

As is, it seems that it makes no sense as is:

1. On a 24 Mpixels full-frame camera, at 100%, the blur from camera movement will be more visible than on a 10 Mpixels full-frame camera.

2. A photo intended to be printed small can have slight blur at 100%: nobody will see it when scaled down for printing. When doing a high-quality large print, even a small blur will be noticeable.

3. Image stabilization (vibration reduction) affects the blur when shooting handheld.

4. The blur will not be the same on a cropped vs. full-frame sensor.

I imagine that the rule of thumb appeared first when there was no DSLRs yet, and photographers were talking about SLRs with 35mm film. Is this that fact that makes the three of four points irrelevant? If yes, what about the second point? If not, what is the origin of this rule?

Answer 1

I did some quick Google Books searches, and while I can't pinpoint the origin, there are a number of references to it as a rule of thumb or general guideline in the early 1970s, and none that I can find before that. There are plenty of earlier references to the idea that a longer focal length requires a faster shutter but they're all general advice.

The first reference I find is from Popular Photography in 1972:

A rule that will help you determine the slowest hand-held shutter speed to use is: place the number one over the focal length of the lens (in millimeters). For example, with a 100-mm lens, one over 100 is ¹⁄₁₀₀ (¹⁄₁₂₅ would be the closest speed to set); with a 250-mm lens, the rule gives ¹⁄₂₅₀ sec. Use this rule as a guide. You may be able to hold for somewhat slower speeds if you're steady and your camera holding technique is good. If you're shaky, you may have to shoot at a faster speed than the rule indicates. Experience will tell this. If in doubt, use a tripod or other firm support and a cable release, when possible.

A year or so later, I found this

You can minimize or completely eliminate camera movement if you remember this rule: For hand-held shooting, don't use a shutter speed any slower than the focal length of the lens. The normal lens on a 35mm camera is is 50 to 55mm. When using this lens, set the shutter at ¹⁄₆₀th second. ... — Walter Chandoha, How to Photograph Cats, Dogs, and Other Animals, Crown Publishers, 1973

I doubt that either of these is the first occurance, though. There's a whole bunch of examples from around the same time, like this:

A rule of thumb is to use a shutter speed at least as high as the focal length of the lens: a 60th for the 50mm, 125th for the 105mm, 250th for the 200mm, and so on. But experience may show you are steadier or shakier than this rule assumes. — Robert Foothorap and Vickie Golden, Independent Photography: a biased guide to 35mm technique and equipment for the beginner, the student, and the artist, Simon and Schuster, 1975

So, I don't know exactly where it came from, but it's definitely an idea for 35mm film, and it's clear that in its early form, it was seen as a general guide, not a law.

Answer 2

It is indeed a rule that comes from film cameras.

On point 4 the answer is simple:

Multiply the focal length with the crop factor of your sensor. Because the sensor is smaller than a full frame sensor, it will not cover the full image circle, cropping out a smaller image. This has the effect of looking like a longer focal length.

E.g. on Canon, a 50mm lens from full frame has a rule of thumb of 1/50s. On a Canon crop sensor, with a crop factor of 1.6, this reduces to 50*1.6 = 80 hence a recommended shutterspeed of 1/80s.

Point 3 is bit more tricky:

I hope you are familiar with the concept of stops. Double or half the exposure = a difference in one stop. IS is described in stops. e.g. IS that gives you two stops should allow you to handhold your lens for an amount of time 4 times as long as without IS. Of course it requires you to be reasonably steady, as IS can only do "so much".

Point 2: Yes, but you shouldn't aim for blurry photos anyway.

Point 1: Yes and no. A higher resolution sensor will resolve the blur more clearly, but it isn't blurred more. More pixels cover the same area, hence, when viewing an image at 100%, you will get the impression of more blur, though the blur is identical. The resolution of the "other sensor" was just too low to resolve it.

To give you an absurd example:

If you had a camera with 1 pixel, it would never show any blur - because it cannot resolve it.

Answer 3

the 1/focal length rule is based on the idea that degree to which detail is spread over the film plane is proportional to the focal length (when the focal length doubles, the blur doubles, any camera motion is effectively magnified), and also proportional to the shutter time (when the time the shutter is open is doulbed, the blur doubles, as twice much camera motion takes place during the exposure). Setting the shutter speed to the inverse (one over) of the focal length causes both effects to cancel out [in theory] and thus delivers a constant amount of blur.

Now like the depth of field formulas, this formulation relies on the final output size being held constant. When viewing images at 100% the rule falls down as the appearance of motion blur is then dependant on pixel size, not just the absolute size of the blur on the sensor (if the blur distance is smaller than a pixel the blur wont be visible).

IMO everything should be compared at the same final output size so the pixel density issue is not that important but it's worth knowing that if you apply the rule blindly you might not get any advantage of your extra pixels, but you wont lose quality compared to a camera with lower pixel density.

Answer 4

There's lots of good information in the other answers which we'll not repeat here, but one important thing that has been hinted at hasn't been explicitly mentioned: Pretty much all of the "rules of thumb" from the film era are based upon a presumption that the image will be printed no larger than 8x10 inches and viewed from a distance of about 10-12 inches by a person with 20/20 vision.

If one is concerned about perceptible blur when viewing a high resolution digital image on a large monitor, all of those rules of thumb that are affected by magnification go out the window. This would include any depth of field calculations based on the 8x10 assumption as well as any expectation of the impact of motion blur on a photograph.

Why? Because when you magnify an image more, you also magnify the blur in that image. This also applies to different sensor sizes when the results from them are viewed at the same display size. An image from an APS-C sensor must be magnified 1.5-1.6 times more than an image from a full frame camera or 36x24mm 135 film negative. An image from a µ4/3 camera must be magnified 2X that of an image from a 35mm camera to view at the same size.

If one is viewing a 24MP image at 100% (one pixel in the image = one RGB pixel group on the monitor) on a 23" HD (1920x1080) monitor, one is looking at a portion of the equivalent of a 60x40 inch print! Of course rules of thumb about shutter speed/camera motion and DoF calculations intended for much lower magnification will not be correct in such a situation.

Answer 5

The 1/shutter speed rule is intended as a guideline or rule of thumb, and shouldn't be taken as a perfectly precise number. Likewise, the transition from "sharp" to "not sharp" isn't absolute and sudden; the sharpness will degrade gradually as you go to longer exposure times. Also, different people of different ages holding different cameras and lenses at different times, with different levels of fatigue and strength will all vary in their ability to hold a camera steady.

The intent of the rule is to give you an image that's "acceptably sharp." That doesn't mean perfectly sharp. A photograph taken at a 1/FL shutter speed will almost certainly be less sharp, to a visible degree, than a photograph taken on a tripod, or at a much faster shutter speed.

And you are certainly correct that the rule predates digital photography. I learned it as a child from my father.

So with those expectations set, let's look at your concerns:

1. This would be a big concern if the rule were intended to produce perfectly sharp images. But you need very careful technique to produce images that are sharp enough to show the difference between 10 Mp and 24 Mp. Shooting hand-held at 1/FL is not "careful technique." My expectation is that at 1/FL, there will be blur due to shake of well over 1 pixel on either sensor.
2. This is true, you can tolerate less sharpness in a photograph that will be printed small. If I had to guess, I would expect a rule of thumb to target "acceptably sharp" at something like 8x10. But, again, that's "acceptably sharp," which is a vague notion, and falls short of "perfectly sharp."
3. This is absolutely true, and is of course the whole reason for image stabilization. Usually the manufacturers will claim a handholding improvement as a certain number of stops. So if the rule-of-thumb (or your experience) tells you that you can take a certain shot at 1/250 of a second with acceptable sharpness without stabilization, and your stabilization system is supposed to give 3 stops of improvement, then you should expect to get similar quality with a 1/30 second exposure and stabilization active.
4. Yes, assuming that you're enlarging the photographs to the same degree as full frame, this is correct. Thus I've often seen the recommendation to use the "35mm equivalent focal length" in the rule of thumb, instead of the actual focal length.

Over and above all, it's probably best to learn how well you can handhold your equipment, taking into account your innate ability to hold steady, the ergonomics of your gear, and the effectiveness of its stabilization. The rule of thumb, with adjustments for stabilization and sensor size, is a good starting point, but it's only a starting point. And if you need absolute sharpness, either aim for at least a couple of stops faster than the rule says, or use a tripod!