[JamieP]

The crop factor / resoulution thread in this forum and the handheld prowess thread over in the General forum got me thinking:

How does the crop factor affect hand-held shots? When using the (1 / focal length) minimum shutter speed rule of thumb, do we need to apply the crop factor and use the effective focal length?

For the sake of simplicity, let's assume no shake reduction used.

[Marc Sabatella]

How does the crop factor affect hand-held shots? When using the (1 / focal length) minimum shutter speed rule of thumb, do we need to apply the crop factor and use the effective focal length?

Yes. The effect of camera shake is dependent on angle of view, not the actual focal length - focal length is just a convenient stand-in for angle of view in the calculation. And really, it's just a rough approximation. But again, yes, when using any format other than 35mm you need to include the crop factor so that we are at least approximating the same angle of view.

[Aegon]

I agree with Marc. However, in minutes a half dozen people will agree with us and another half dozen will disagree.

But here's how I "prove" it to myself. Digital cameras with tiny sensors an 12x zooms may get out to an effective 400mm with an actual 55mm (or whatever), and good luck holding that steady without a tripod. It is all about angle of view.

[noblepa]

With 35mm cameras, the ROT was to not attempt to handhold a shot at a shutter speed slower than 1/focal_length.

On page 62 of the manual for my K10D, it says "Although there are individual differences among photographers, the shutter speed for a handheld camera is generally 1/(focal length * 1.5). For example, it is 1/75 for a 50mm and 1/150 of a second for 100mm. Use a tripod or the Shake Reduction function (p.67) when using a lower shutter speed."

So, Pentax seems to agree that its the angle of view that governs the slowest handheld shutter speed.

SR gives you, conservatively, two stops of leeway in your shutter speed. When using SR, I revert to the 35mm ROT, or 1/focal_length, with the knowledge that I can fudge it a little more, if necessary. This makes the mental arithmetic easier and gives me a good safety margin.

I also keep in mind that, shooting in Raw mode, I can underexpose by two stops and still recover a good image in PP. That means that if the meter tells me that I should be shooting at 1/4 sec at f/3.5 (fastest aperture on my kit lens), I can shoot at 1/16 and SR will probably give me a good image. I don't think that SR can save a shot at 1/4 sec. At 1/16 with SR, assuming a short focal length, I stand a chance of handholding the shot. At 1/4, even with SR, no way.

Paul Noble

[newarts]

It depends on what display criteria one uses to decide. Lets say we consider a particular scene in the viewfinder of shakey cameras.

Each point on the scene sweeps across the sensor as a camera is rotated about its lens. The time this takes is very close to:

time=Sensor_width/((Focal_length)(camera_rotation_rate))

Now if you are going to make the same sized, uncropped print from each camera, you probably don't want the point to have moved more than some small fraction of the print size during the exposure (like .01"/10"=.001) therefore the exposure time must be:

Exposure_time=.001*sensor_size/((focal_length)(rotation _rate).)

Notice the appearance of sensor_size in the above. That's where crop factor enters. The sensor_size is equivalent to 36mm/(crop_factor) so the above can also be written as:

Exposure_time=.036/((crop_factor)(focal_length)(rotation _rate).)

On the other hand, we could do some pixel peeking & print the photo at its maximum resolution. In that case we wouldn't have wanted the image to have shifted more than one sensor pixel during the exposure. The exposure time requirement then is:

Exposure_time=Pixel_pitch/((focal_length)(rotation _rate))

Now the sensor size is gone and is replaced with the pixel pitch! In this case the crop factor doesn't matter - what matters is the pixel spacing.

Dave