Challenging Conventional Wisdom/Rules of Thumb

[William W]

Unfortunately Thom Hogan seems to be sadly mistaken . . . [etc]

Thank you Manfred.
I’ve had similar conversations before and did not really want to debate the issue here, so that’s why I merely wrote:

. . . perhaps the mathematics of Thom’s predications and the resultant flow of his “logic” should be questioned? (I think this point has strong legs.)
There are quite a few (Nikon users) who disagree with Thom’s hypothesis.

WW

[Adrian]

This is one of the best and most intelligent threads I have read, anywhere, in a long time, about photography.

[Manfred M]

Thank you Manfred.
I’ve had similar conversations before and did not really want to debate the issue here, so that’s why I merely wrote:



WW

Thank you Manfred.
I’ve had similar conversations before and did not really want to debate the issue here, so that’s why I merely wrote:
WW

My main issue with Thom's article is that is widely quoted and people believe it without question because of some of the techno-babble in his writing. Unfortunately, he is not an engineer and does not understand control systems.

That all being said; VR / IS is not going to give you as sharp a result as with it turned off under certain shooting conditions. I suspect that is where the 1/500th issue really comes from, but then we have to remember that he wrote the article several years ago, and technology does march on. I have no doubt that with a high enough shutter speed, on a given lens, having IS / VR turned of is going to give the shooter better results than shooting with it turned on.

While the measurement / predictive side does work, the real world mechanics to that makes the IS / VR adjustments are not perfect. The response times and component mass of the corrective mechanism (i.e. inertia) as it will tend to overshoot or undershoot the “correct” amount of correction required , if only ever so slightly. If you ever watch these types of systems in action, there is a tiny bit of “flutter” that is damped out over time. Overdamp the system, and the response time is too slow. Underdamp it, then the system does not settle down to a stable state quickly enough. Pop the lens on a tripod and lock down the floating element, and I would expect superior results. Likewise, shoot at 1/8000th of a second, I would expect to see little or no motion blur, with the IS / VR turned off.

Each lens and perhaps camera / lens combination is going to behave differently. A more modern lens should be far better at accurately compensating for camera movement than an older one. My 80-400mm lens is the first VR lens that Nikon introduced. When I shoot with it, I can’t get as good a result as with a newer design like my 70 200mm at similar focal length / shooting conditions. This is hardly surprising.


My main issue with Thom's article is that is widely quoted and people believe it without question because of some of the techno-babble in his writing. Unfortunately, he is not an engineer and does not understand control systems.

[NorthernFocus]

I believe Thom's premise regarding shutter speed is also based on a rule of thumb for control systems which may also now be outmoded with modern processors. As I (re)read his article that is what he suggests. He claims to have info. from Nikon regarding the sample rate of 1000hz then applying a rule of thumb that a control system can not (or should not) respond any quicker than 1/2 sample rate. I have done a good bit of work on control systems myself and have never heard said rule of thumb but that's how I'm interpreting his article on the subject.

All that said, forgetting Thom for a while, unless the shutter release is linked to the circuit that moves the IS/VR element in the lens, I have never understood the argument. The IS/VR is making corrections to keep the transmitted light in a stable position where it strikes the imaging sensor. If it is doing so independently, then it is hard for me to understand how a high shutter speed could be detrimental. Considering that what causes motion blur on the camera end is tha the image is transitting across the sensor while the shutter is open. Logically if the IS/VR function is able to limit said motion to acceptable levels at low shutter speeds, then by definition it should do so even better at higher ss. It only stands to reason that for a given transitional velocity of the image across the sensor, the distance travelled will be less as higher ss therefore less motion induced blur in the image.

Interestingly, until now I've never worked this out in my head (then again I have trouble tying my shoes in the morning). There was enough noise on the topic from various sources that I simply reasoned that we got by without VR for years, so if in doubt simply turn it off. And that worked for me for several years. However, and this is the point of starting this thread, with the addition of longer lenses and higher resolution sensors to my gear inventory, and having begun to print large metal prints, previously unimportant details are now becoming relevant.

At any rate, I shot a couple thousand frames over the past few days, mostly with my 500mm VR lens. Did some shooting with VR both on and off. Though not scientific experimentation, as I process through my files I'll try to do some informal statistical analysis on keepers vs blur related culls. But more importantly, I hope to have some decent images to post soon

List of species photographed in the past 10 days:

• sandhill crane
• horned grebe
• red necked grebe
• arctic tern
• muskrat
• barrow's golden eye
• harlequin duck
• northern sea otter
• bank swallow
• red necked phalarope
• german shorthair pointer

We're having an unbelievable run of fair weather for Alaska so one has to make hay....

[Cantab]

Although I have nothing to contribute to this discussion, I've found it very informative.

Dan, I'll look forward to seeing your recent shots (and any informal statistical analysis you do).