[m42man]

I've gone through my workings again and realised the formula still isn't right! (Though my derivation for the formula was flawed, the end result unfortunately worked correctly for the 35-105 case - unlucky!)

The correct formula (for ALL zooms) is:

x = 2*H*L/(H+L)

and the min SR effectiveness is still x/H *100%

Note that in general you won't be able to set the exact value for x, but if you choose the nearest lower value, then the formula for min SR above still stands (and it applies to when you have the zoom set to max focal length). If however the value of x is significantly closer to the next value higher, then you might want to choose this instead, and the formula for min SR is as follows (and applies to when you have the zoom set to min):

min SR effectiveness = (2*L - x)/L *100%

BUT this particular expression fails for x > 2L, because you start to get shake magnification, so never set your camera to more than 2L! (This is especially true for lenses with large zoom ratios.)

Generally, you'd want to choose the lower value - it's safer.

[m42man]

I thought I'd do some calculations for real lenses:

28-70: theoretical setting: 40mm, actual setting: 40mm, min SR: 57%
28-80: theoretical setting: 41mm, actual setting: 40mm, min SR: 50%
28-90: theoretical setting: 43mm, actual setting: 40mm, min SR: 44%
28-105: theoretical setting: 44mm, actual setting: 45mm, min SR: 39%
35-70: theoretical setting: 47mm, actual setting: 45mm, min SR: 64%
35-105: theoretical setting: 52.5mm, actual setting: 50mm, min SR: 48%
35-135: theoretical setting: 56mm, actual setting: 55mm, min SR: 40%
70-150: theoretical setting: 95mm, actual setting: 100mm, min SR: 57%
70-150: theoretical setting: 95mm, actual setting: 85mm, min SR: 57% **
75-150: theoretical setting: 100mm, actual setting: 100mm, min SR: 67%
70-210: theoretical setting: 105mm, actual setting: 100mm, min SR: 48%
80-200: theoretical setting: 114mm, actual setting: 100mm, min SR: 48%
80-200: theoretical setting: 114mm, actual setting: 120mm, min SR: 50% ***
80-210: theoretical setting: 116mm, actual setting: 120mm, min SR: 50%
60-300: theoretical setting: 100mm, actual setting: 100mm, min SR: 33%

Note that I did the calculations for both "next lower" and "next higher" for the 70-150 and 80-200 lenses and for the 70-150 you can see that you can just as easily use the lower value, despite it being significantly further from the theoretical value.

[Douglas_of_Sweden]

I wouldn't follow this idea. You would still be missleading the camera to introduce unvanted blurr as soon as you shoot at a focal length below the setting you have entered. I'd recommend either turning the shake reduction off, or set it to the shortest focal length, unless you know that you are going to work for a while at a specific (and longer) focal length.

The A35-105 is a fine lens. Good for portraits.

No problems here since it was used on a Program A with BNC400 film.

[m42man]

I wouldn't follow this idea. You would still be missleading the camera to introduce unvanted blurr as soon as you shoot at a focal length below the setting you have entered. I'd recommend either turning the shake reduction off, or set it to the shortest focal length, unless you know that you are going to work for a while at a specific (and longer) focal length...

No, the camera wouldn't be introducing blur at FLs other than that which you input. The camera will still be moving the sensor to counteract the shake, it's just that the movement will be less than optimal for longer FLs, and more than optimal for shorter FLs.

Iy you want the best SR possible, then you have to enter the FL at which you're operating. If you don't want the hassle, then you should enter as I've described.

You'll notice that, most of the time, with lenses of moderate zoom ratios, you'll be getting about 50% of SR at the extremes of the zoom range. What this actually means is that the camera will be reducing shake by a factor of 2. This means you get a worst case of 1 stop of shake reduction. Which is pretty worthwhile, I reckon.

You get 100% SR when the FL is equal to the value input to the camera (obviously). If everything were perfect, the camera would exactly counteract the shake, so you'd end up with no blur at all - but of course everything isn't perfect, so you get around 2-3 stops worth of shake reduction (usually, depending upon shutter speed).

[m42man]

Talking about shutter speed has given me an idea: maybe we should skew the FL value so that the longer end of the zoom range is favoured. That would mean that we'd be getting less shake reduction at the short end, more at the long end. We'd end up with slightly (but only slightly) larger values for x. I'll have to have a think...

Meanwhile, here's an interesting link:

Falk Lumo: Pentax shake reduction revisited

[Douglas_of_Sweden]

No, the camera wouldn't be introducing blur at FLs other than that which you input. The camera will still be moving the sensor to counteract the shake, it's just that the movement will be less than optimal for longer FLs, and more than optimal for shorter FLs.

You contradict yourself. Obivously you introduce blurr if the sensor move "more than optimal for shorter FLs". It means that any motion you do while exposing is magnified by the SR system.

[m42man]

You contradict yourself. Obivously you introduce blurr if the sensor move "more than optimal for shorter FLs". It means that any motion you do while exposing is magnified by the SR system.

No, I'm afraid you don't understand. If the camera moves the sensor exactly to counteract the shake, then there will be zero blur. If it moves it by 50%, say, you'll get 50% of the blur you would have if SR was switched off. If it moves it by 150%, then you'll again get 50% of the blur you would have with SR off.

Moving the sensor too much therefore doesn't "introduce" blur - unless you move it by more than 200%, in which case you get what I've called "shake magnification". The whole point of my little expedition into maths was to find a value which keeps the sensor movement within bounds, even at minimum FL.

[Turbo_Spool900]

No, I'm afraid you don't understand. If the camera moves the sensor exactly to counteract the shake, then there will be zero blur. If it moves it by 50%, say, you'll get 50% of the blur you would have if SR was switched off. If it moves it by 150%, then you'll again get 50% of the blur you would have with SR off.

Moving the sensor too much therefore doesn't "introduce" blur - unless you move it by more than 200%, in which case you get what I've called "shake magnification". The whole point of my little expedition into maths was to find a value which keeps the sensor movement within bounds, even at minimum FL.

Took me a while, but I get it. Thanks for doing the math on this one. Much appreciated.

[m42man]

Took me a while, but I get it. Thanks for doing the math on this one. Much appreciated.

Thanks for getting it, and thanks for the thanks! And you're very welcome - I had a lot of fun figuring it out (something to keep the old grey matter ticking over!).

Ironically, I myself possess only one vintage zoom - a Pentax M 75-150mm f4, which I've had from new. It was used a lot in the 80s (on my MX), and was pleased with the results. I seem to remember it was measured to be pretty top-notch by the pro reviewers. However, I have to say I don't really use it now, because it's a bit bulky, the focusing is a little heavy for my taste - and because I won't always get full SR (ouch!).

[goldenarrow]

Thanks to both m42man and Douglas of Sweden for having two different approaches (and opinions). I appreciate this and will cogitate about both. Seems like Douglas of Sweden has an approach similar to mine, kind of like a "hippocratic oath version" (doctor, do no harm) where the SR is set at a more conservative level. m42man has a more "joystick control" approach similar to flying an airplane, keeping things within bounds of estimated extremes. You both get 50% blur in one scenario, but the blur ends up being in opposite directions (one is under-controlled (Douglas and me), the other method is over-controlled (m42man). But it is the same AMOUNT of blur. Interesting to think about.

Of course, like I said earlier (in a different thread), I would rather not have to think about it at all, Pentax limitations being what they are. Thanks to both of you......

GA