A friend recently asked me about the minimum handheld shutter speed for cropped sensor cameras; the old adage for 35mm cameras was known as ‘1/focal length’ and he was wondering if he should apply a 1.6 crop factor for a cropped sensor. I saved my reply and thought I’d post it here…
The correct answer is… There is no correct answer. Let’s start with the obvious, which is that a camera mounted on a sturdy tripod, with the mirror locked up and a shutter delay of several seconds after the mirror has locked up will achieve as close to zero camera movement as possible. Having said that…
The’inverse’ rule has been around for a long time but it’s no more a rule than the ‘sunny f/16’ rule for exposure. In his book, ‘The Camera’, Ansel Adams found in his test that using a 35mm camera and a 50mm lens, when making photographs of leafless trees he could see a difference between 1/125th second and 1/250th second. His ‘rule’ then was 5 times focal length. In the film days we generally had 1 size of film for 35mm, 1 size for medium format (6 cm x ___cm) and a few sizes for large format. Those days are gone. For example, in his recent field review for the Nikon D800/D800E, Michael Reichmann recommended 2x – 3x the 1/focal length rule as a minimum handheld shutter speed.
There are a few different issues, all of which come into play. One is the sensor size. Another is the pixel density. Another is the field of view of the lens which relates to focal length in comparison to image size. Then there’s the quality of the glass and the resolving power of the lens. Then there’s the f/stop, which impacts not only depth of field and shutter speed but also diffraction at smaller f/stops. Don’t forget sensor noise, introduced at higher ISO values. Then there’s the difference between mirrored DSLRs and mirrorless cameras in what is often called ‘mirror slap’ – the movement of the viewfinder mirror up out of the way to expose the image, followed by it crashing back down into place. We haven’t touched on image stabilization – either in-camera or in-lens and last but not least the user – who contributes time, experience, health, technique, stance, etc.
Take the Leica V-Lux III, a mirrorless 12.1mp (2011) camera with a 25-600mm ‘equivalent’ lens and compare it to a Canon 1Ds MII, a 16.7mp (2005) camera with a Canon 600mm lens f/4.0 lens and tell me which one will yield a better image handheld. Even though they offer ‘equivalent’ focal lengths in relation to sensor size, to start with one weighs 15-20 lb more than the other. Make one image with each and carry them down a trail for four hours and make one image with each.
We haven’t discussed CCD vs. CMOS sensors, the absence of the antialiasing filter on the Nikon D800E vs. the Nikon D800 for example, the definitition of ‘sharpness’ in terms of line pairs/mm vs. local contrast and that just gets us to the input resolution side of things. Now we have to consider the logic engine in the sensor, whether or not it’s using pixel binning to lower the pixel count but increase resolution, whether we’re comparing mathematical MTF charts or visual acuity, what is an ‘observable difference’ in terms of CoC, at what output resolution, on what device (monitors only display about 100ppi for example) and at what distance… The list goes on.
One technique often described is to take a laser pointer and attach it to the barrel of the lens with rubber band or something and point the light at a wall say 10 m away. Make a series of exposures at different f/stop, shutter speed and ISO combinations, handheld and tripod-mounted, with and without mirror lockup if you’re using a mirrored camera. Now decide how you’re going to compare the results – on screen, on screen at 100% magnification, in a 4×6 print at 180dpi, a 20×30 print at 720 dpi or on a billboard from 300 feet away.
One series of articles that’s very good is: DSLR SENSOR SIZE AND PIXEL DENSITY
Aren’t you glad you asked?
Now go out and make some photographs!!
P.S. Here’s a blog post from Jim Kasson: Rules of thumb for handheld shutter speed