>OK. So what I am hearing is that it all comes down to the >issue of subject motion. I understand what you are all saying. >I wish I could think of a way to set up a test of motion in >low light and see the difference.
Yup, exactly their point! VR isn't going to do anything for this, as a slower shutter speed is a slower shutter speed is a slower shutter speed, is a slower shutter speed...and will always be a slower shutter speed, therefore showing more subject motion blur...VR or no VR...no difference at all if VR is used!
VR works on the WHOLE IMAGE...the projected image, not on individual parts of the image.
ISO, of course, will allow the use of a faster shutter speed, but VR won't. Next camera generation, you WILL be able to use an f/4 lens instead of the f/2.8 lens (because ISO 25,600 will have the image quality of today's ISO 12,800). But until then, the f/2.8 lens is needed.
And that faster shutter speed WILL affect the blurred (fast-moving parts) of the image more than it will affect the slower (less blurred) parts of the image. Therefore, shutter speed has a selective effect on object motion in the frame, based on object "speed" (or motion blur, which is a representation of speed in the image). And ISO is the only thing, other than lens aperture, that allows one to achieve a faster shutter speed in an ambient light situation (short of adding more light via flash).
Bottom line: VR = tripod subsitute High ISO = lens aperture substitue
As said, the only other way around the dilemma (other than ISO) is increasing the light level (using flash)...but that isn't allowed in some situations, so it's not even an option.
But I fully expect the ISO/image quality ratio to improve yet some more as time goes on. Next generation, you'll be able to use an f/4 zoom lens where you formerly used an f/2.8 zoom lens. And an f/2.8 zoom where only an f/2.0 prime used to work. This sort of improvement has happened in the past, and will keep happening.
Bottom line: VR compensates nicely for a wider lens aperture insofar as handholding is concerned, but does nothing to compensate for subject motion. Only higher ISO can do that.
That said, I do see your point...f/2.8 and f/4.0 are only one f/stop apart from each other. Sometimes, that one-stop improvment yields no practical improvment in shutter speed/handholdability/motion blur of fast-moving objects (= subject sharpness)--because the light is so dim. But other times, it's the difference threshold between a failed image and a successful one.
Often, people make a huge deal out of the difference between f/2.8 and f/4, and describe it like it's some sort of magic difference that will always solve the blur issue. It's no such thing! It's a one-stop difference just like any other one-stop interval. As such, for some photographers in some situations, an f/4 zoom lens coupled with one or more f/2 or faster prime lenses may actually be a better option than a single f/2.8 zoom. And for some photographers, even an f/1.2 lens (VR or not) wouldn't provide the shutter speed they need to avoid subject blur!
Addendum: Panning is another, often rarely understood/under-appreciated way to solve (or at least improve) the subject motion blur issue, but it's a separate topic for another thread.