Originally posted by Donald Hello. I have done this alot with Canon (don't remember model but new last year), Nikon (D70s), Leica (Rsomething) and Pentax (DS2).
All cameras had variable frame rate, Canon and Pentax the worst, depending on storage media. With an extreme-something number card frame rate seemed continuous, with some cheap SD cards I bought in emergency the frame rate was very inconsistent, for all of the above camera makes.
All of them hunted for focus so much that I abandoned and went to manual focus. The more I try autofocus the more dissatisfied I am with it: very self critical I believe I've never got a sharp photo using autofocus, and also now believe any body arguing autofocus speed performance just truly doesn't know what they're talking about producing real photos for a real purpose.
Try a D200, 5D or even the 30D, then you'll see the difference.
My experience is that the D200 has the highest frame rate but the focusing accuracy is close but not absolute. This applies to the tracking ability when the subject moves across the frame too. The 5D has a lower frame rate of 3 fps (vs the 5fps of D200 or the 30D) but the focusing accuracy is obviously better. Both the D200 and the 5D keep the frame rate in a good way.
As for the K10D, it tries to adjust the focus before it "advances" the frame. If it fails to track or measure the focus, it will then "releases" the frame and then the procedures will repeat. My impression is that it cannot do both thing at the same time or simply that the AF system is not fast enough so that the camera tries to delay the frame advance to give it a trial.
The final results is that neither the frame rate can be kept constantly nor the correct focus hit rate is high for the K10D and actually the out of focus photos is in the highest number and that the amount of OOF is very large (The D200 ones are slightly off only)
K100D performs even worse owing to the very shallow buffer and the fact that its AF is even slightly slower and less responsive than the K10.
Nonetheless, I do agree to certain extent that the camera does not matter but provided that for certain applications it meets our shooting requirements. It actually is not a brand issue but a camera issue afterall, but it happens Pentax still don't have any up-market model which can perform in such maybe more "adverse" shooting conditions with more demanding requirements.
And yes, we can use MF if the AF fails but I just wonder if we can do it better than the camera in such cases. Also, why not let the camera do it if there really exists some cameras can do it?
For normal and typical shooting under daylight, I do believe the AF system of the K10D and K100D is Okay. At least the achieved focus has been more precise than pervious AF system of Pentax *ist Dx DSLRs already (The K cameras are the first Pentax DSLRs which can focus accurately for the "problematic" lens FA*85/1.4 at 1.4). For low light, yellow light and continuous moving object tracking, indeed Pentax is lagging behind and I do hope the coming higher end model DSLR of Pentax, if any, should be able to catch up in the game.
I really hope that what Roland Mabo told us will be true, i.e., K-1D (or whatever it is named) will come in autumn next year and this camera will perform up to the pro class DSLRs of C and N, with newly designed and highly improved metering and AF system in it.