[reivax]

Recently got my hands on a Tokina 28-70 f2.8-4.3 and I had some questions about how to use it.

First thing I noticed is that when I turn on the camera, it will ask me for the focal length. I know this has been discussed in other threads, but there's never really been a consensus on what to enter. Some people recommend "go low", others recommend "go high", and then others recommend "shoot for the middle". Does it really matter? Is this just information that gets stored in the camera? Or is it something that can affect the quality of an image?

The next thing that I noticed is that the camera does not change its aperture when I zoom in (and I assume the aperture is supposed to change being that it's not a constant aperture lens). Is this something I'm supposed to do manually on older zoom lenses? If it is? How do I enter 4.3 on my camera?

Are these issues that come up with all older zooms? Do you have to enter focal length, etc. with A-Series lenses as well? Or is it because it is not a Pentax zoom? Or is it because there is something wrong with the lens itself? When I set the aperture and take a picture, it does seem to meter properly, I'm just really curious as to why it doesn't change on its own as it would with other zooms that I own (DA model zooms)

Any other tips on working with older zoom lenses (Pentax or non-Pentax) would be greatly appreciated. I'm interested in getting some more older zoom lenses (mainly for video use) but I want to make sure that I get what would be best to use. Might be a good time to ask for recommendations on older zoom lenses for video use as well.

Thanks everyone.

[GabrielFFontes]

The focal length shows on the EXIF and also tell the SR system how to act. I think the best it to set it to like 35mm, not sure tho...

[sterretje]

If in doubt, switch SR off.

I can't answer the question about the aperture. Maybe you can contact one of the members that have the lens (see the review that you linked to) to confirm the behavior. Curious what you see in the exif data (2.8 or 4.3).

[jimr-pdx]

If you're going to spend most of your time at one end or another of the zoom, I'd bias the SR that way. You can dive into the menus and reset the value if you find yourself at another range more often.

I don't do much video but I would think a one-touch style would serve better, so you don't need to regrip the lens to zoom vs. focus.. but that's just a thought, not something that I have real experience with. That idea would cause me to seek the SMC-A 35-135mm, SMC-M 75-150 or K 45-125. In non-Pentax I really enjoyed a Vivitar 28-105 f/3.5-4.5 one-touch (serial #77xxxx, Kobori make); other than immediate zoom creep to 105mm it was excellent optically, and had the A aperture setting. My current m-focus fave is the Rikenon XR 70-150/4 with close-focus skills and internal zoom, but it's a two-touch design.

[m42man]

Your camera needs to know the current focal length so that it can calculate how much to move the sensor in order to counter any shake that it detects. So, you'll find that you won't be prompted if you turn SR off.

However, it's a pity to lose one of the best features of a Pentax DSLR, so what would be the best value to enter (assuming you don't want to keep changing it as you change the zoom setting)? Well, I figured out a formula a while back which enables you to input a value which will give you the same (reduced of course) amount of SR at each end of the zoom range:

Value = 2*Max*Min/(Max+Min)

If you set this value (or nearest convenient), you'll end up with maximum SR at this focal length (obviously) and the effectiveness of the SR mechanism will drop off as you move the FL away from this, with minimum effectiveness at either end of the zoom range.

For your lens, you'd set 40mm, and you'll get over 2x shake reduction at each end of the zoom range (at 40mm FL youll get max SR, which, in theory, results in zero blur, but in practice will probably be more than 8x). I think that getting over 2x SR is well worthwhile - far better than turning SR off (though you should turn it off - with any lens - if you're panning to follow a fast-moving object, or if you're on a tripod).

Of course, if you envisage usage at either one end or the other of the zoom range, then you'll get best results by setting the actual FL that you're using. However, do be aware that if you forget yourself, and operate at 28mm when you've input 70mm, you'll get more blur than if you'd turned SR off! You have been warned!

There's more info here (and at the link I gave):

https://www.pentaxforums.com/forums/pentax-slr-lens-discussion/178019-tamron-...ml#post1859602

[reivax]

If you're going to spend most of your time at one end or another of the zoom, I'd bias the SR that way. You can dive into the menus and reset the value if you find yourself at another range more often.

I don't do much video but I would think a one-touch style would serve better, so you don't need to regrip the lens to zoom vs. focus.. but that's just a thought, not something that I have real experience with.

What do you mean by one touch style?

[reivax]

Of course, if you envisage usage at either one end or the other of the zoom range, then you'll get best results by setting the actual FL that you're using. However, do be aware that if you forget yourself, and operate at 28mm when you've input 70mm, you'll get more blur than if you'd turned SR off! You have been warned![/url]

Thanks. This kind of stuff makes me wonder if older zooms are worth the trouble. It kind of seems to remove the convenience of a zoom.

Anyone know if if A series zooms require you to input focal length?

[northcoastgreg]

Anyone know if if A series zooms require you to input focal length?

Yes they do. Focal length information begins to be communicated to the camera with the F series zooms.

[reivax]

Yes they do. Focal length information begins to be communicated to the camera with the F series zooms.

That is very good to know. Thanks.

[newarts]

If you know about what focal length you'll actually use, enter it. Otherwise enter the average focal length for the range you might use; it can make a big difference in image quality.

Focal.length.SR.optimal = (max+min)/2

It is the theoretically optimal focal length for minimizing mean squared shake error for random shots between max and min.

Think about what happens to other formulas in the case of very large or very small focal lengths; using a formula often offered, 2MaxMin/(max+min), for large max results in 2min - which makes no sense at all, conversely, if min=0, 2MaxMin/(max+min) = 0 which also makes no sense.

Using average focal length for SR is better than not using SR for hand-held photos.

The next thing that I noticed is that the camera does not change its aperture when I zoom in (and I assume the aperture is supposed to change being that it's not a constant aperture lens). Is this something I'm supposed to do manually on older zoom lenses?

Most zoom lenses maintain their aperture setting during zoom if the setting is at or above the maximum. For example, if the aperture range for the lens is 3.5-5.6, any aperture 5.6 or greater will be maintained during zoom. If you set 3.5 at the lens' minimum focal length the aperture will increase gradually to 5.6 as you zoom out to the maximum focal length.

The Data pin on a Pentax mount lens transmits focal length info to the camera - no data pin, no focal length info. Data pin appears with the F series...



Dave

[sterretje]

What do you mean by one touch style?

There are zoom lenses with two rings (like the current DA kit lens); one ring for zoom and one ring for focus.

There are also zooms with only one ring; you push/pull to zoom and you rotate that ring to focus. I found this type very convenient as you don't have to move your hand that supports the lens while you're focusing and composing.

[m42man]

If you know about what focal length you'll actually use, enter it. Otherwise enter the average focal length for the range you might use; it can make a big difference in image quality.

Focal.length.SR.optimal = (max+min)/2

It is the theoretically optimal focal length for minimizing mean squared shake error for random shots between max and min.

Think about what happens to other formulas in the case of very large or very small focal lengths; using a formula often offered, 2MaxMin/(max+min), for large max results in 2min - which makes no sense at all, conversely, if min=0, 2MaxMin/(max+min) = 0 which also makes no sense.

Using average focal length for SR is better than not using SR for hand-held photos.

I'm afraid you're wrong!

If max is (very) "large", then the value approaches 2*min, and the net effect of the SR accordingly approaches zero (which is OK). If the "lens" is a flat plate of glass, then the FL will be infinite, and the result would indeed be 2*min, so you'd get zero net SR. But a flat plate of glass isn't actually a lens.

Likewise, a "lens" with a FL of zero isn't actually a lens, so that assertion is also meaningless.

However, for a real lens (i.e. with non-zero and non-infinite FL), the formula holds true.

As for your (max+min)/2 formula, you can easily see that a large value of max will result in a value approaching max/2, so for a zoom ratio of greater than 2x, you'd be setting a value more than 2*min, which means that the camera's SR would be magnifying any camera shake. You can demonstrate this crudely by mounting, say, a manual 50mm lens and setting a value of 200mm. Now take a few shots at, say, 1/60s and compare the results with shots taken with SR switched off.

Having said that, a great many vintage zooms have a zoom ratio of 3x or less, so using the formula (max+min)/2 wouldn't be the end of the world - either you end up with a value that's close to the "optimum" value anyway, or you just get zero net SR when you set the zoom to minimum FL (if the zoom ratio is exactly 3x). But this formula does fail for zoom ratios greater than 3x.

The golden rule to remember is: you get shake magnification if you input a focal length value greater than 2x the actual focal length that you're going to be shooting!

[newarts]

I'm afraid you're wrong!

If max is (very) "large", then the value approaches 2*min, and the net effect of the SR accordingly approaches zero (which is OK). If the "lens" is a flat plate of glass, then the FL will be infinite, and the result would indeed be 2*min, so you'd get zero net SR. But a flat plate of glass isn't actually a lens.

Likewise, a "lens" with a FL of zero isn't actually a lens, so that assertion is also meaningless.

You are right ... it approaches but is never equal. That doesn't invalidate the conclusion. Surely as a zoom lens' max becomes greater the SR parameter should also increase - but your formula shows a very weak increase!

Your formula is true only when max=min. It leans towards the min side when max is bigger then min. Your formula can be rearranged (by dividing both top and bottom by max) to read

SR =2min/(1+min/max) As max increases, min/max decreases, and SR smoothly approaches 2min.

This says a 30-300mm lens SR should be set at 55mm and a 30-600mm lens SR set at 57mm! How can this be reconciled with the fact that for a non-zoom 300mm lens the SR setting is 300, etc?

As for your (max+min)/2 formula, you can easily see that a large value of max will result in a value approaching max/2, so for a zoom ratio of greater than 2x, you'd be setting a value more than 2*min, which means that the camera's SR would be magnifying any camera shake. You can demonstrate this crudely by mounting, say, a manual 50mm lens and setting a value of 200mm. Now take a few shots at, say, 1/60s and compare the results with shots taken with SR switched off.

If you already know what focal length you'll be using, that's the one to enter.

In your example of a zoom ratio of 2, the correct formula gives an SR focal length of 3/2 min, right in the center of the max-min range. It is a well established statistical and mathematical truth that this choice, the mean, on average, gives the minimum error.

Having said that, a great many vintage zooms have a zoom ratio of 3x or less, so using the formula (max+min)/2 wouldn't be the end of the world - either you end up with a value that's close to the "optimum" value anyway, or you just get zero net SR when you set the zoom to minimum FL (if the zoom ratio is exactly 3x). But this formula does fail for zoom ratios greater than 3x.

The golden rule to remember is: you get shake magnification if you input a focal length value greater than 2x the actual focal length that you're going to be shooting!

It does not fail for long zoom ranges. Consider a zoom range of 5min. The mean (expected value to a statistician) is min(1+5)/2 = 3min. If you give a zoom lens to an army of monkeys to fool with, the most probable focal length setting will be (min+max)/2. That's why the rule is correct.

The rule is you get shake reduction error any time the actual focal length differs from the SR focal length. Furthermore, the shake error is proportional to the difference between actual and SR focal lengths. This is true for both positive and negative errors. It is this fact and the expectation that the actual focal length is equally probable anywhere between min and max that leads to the identification of the mean as the best SR choice (in the absence of more knowledge).

Dave

[reivax]

I think the engineers are here. Hahaha. Great information.

[PPPPPP42]

Umm, ya we have way too many people that are way too smart on these boards and this happens every time, someone always breaks out the chalk board.

Aside from all that I use mostly all MF A series lenses, and I started skimming when this thread got long winded so this mighta been said already, but if your lens has the A aperture setting leave it on that and set the aperture with the dials, if not go to the beginner Q&A forum here because I think how to use manual only aperture lenses is part of a sticky thread there.

And I always set my SR to the highest a zoom goes to when I turn it on because that's where it needs the most shake reduction since every twitch is magnified. At the wide end movement isn't magnified so much so the SR doesn't really do much at all. Seems to produce good pics all across the range set this way.