[kypfer]

Most of my photography these days is wild birds, often from a public hide and inevitably often in flight. I achieve a reasonable degree of success, certainly for my purposes, but I'd like to reduce my failure rate, which are mostly "missed focus" of moving subjects.

I know my "birding" equipment (K-70 + Sigma 150-500mm f/6.3) is capable, when conditions are "right", but a few comparison tests recently suggest that a 400mm f/5.6 (actually a 70-200mm f/2.8 + 2x) might give a better "lock-on" rate on a moving subject. I should probably mention at this point that my "default" exposure is 1/1500 @ f/8 in TAv mode, which is chosen to cope with an amount of subject movement and give a degree of depth-of-field. My "default" autofocus mode is AF-S in spot mode. For in-flight shots, if there's time to make a choice, I try to use continuous autofocus, single point for a more distant subject and multi-point if the subject is reasonably close and against a plain sky. I have these options set up on the User modes so can switch from one to the other reasonably promptly

So, does half a stop really make that much difference to the autofocus accuracy and response, in which case I'll be considering saving (hard) for a faster lens and hope to get lucky second-hand, or is it simply that the smaller image in the viewfinder allows for better tracking on my part? Inevitably, no two "action" shots are exactly the same and the ergonomics of the shorter lens may play a part in all this.


I'll look forward to all the insights and experiences I'm sure I'll receive, the responses on this board have always been interesting and educational

---

Last edited by kypfer; 08-12-2019 at 01:35 AM. Reason: spelling

[beholder3]

For single center point AF all that happens when using a lens slower than the F5.6 that the AF is specified down to is that the line sensors are shaded partially and as a result the "cross" sensor gets smaller and since the algorithms have less pixels to rely on it works a little worse = less accurate.
The remaining sensor pixels get the exact same amount of light, so there should be no speed difference.

[RonHendriks1966]

Ad some images that show full sensor take of your birds. That helps giving advise.

I once ownd the Sigma 170-500mm, an even older version of your lens. Birds where not my thing, but people could do it with that lens.

Look at the dof calculator:
https://www.dofmaster.com/dofjs.html
You see that your win in dept of field with 500mm is very limited. Your lens needs a little stop down, but doing f7.1 is probably good enough and it won't get better.

Your K-70 is a little handicapt in the way that it has older and different parts inside then for instance the KP.
https://www.pentaxforums.com/forums/pentax-cameras-compared/
It has a slower processor and an older version SAFOX X af-module where KP and K-1 share an almost identical design af-module that dates back from the K-3.

[kypfer]

For single center point AF all that happens when using a lens slower than the F5.6 that the AF is specified down to is that the line sensors are shaded partially and as a result the "cross" sensor gets smaller and since the algorithms have less pixels to rely on it works a little worse = less accurate.
The remaining sensor pixels get the exact same amount of light, so there should be no speed difference.

OK, I think I understand this ... in layman's terms, anything slower than f/5.6 is non-optimal? If that's the case, it supports my suspicions and I need to think about "upgrading" ... or getting closer, somehow


Thank you

---------- Post added 08-12-19 at 10:39 AM ----------

Ad some images that show full sensor take of your birds. That helps giving advise.

I once ownd the Sigma 170-500mm, an even older version of your lens. Birds where not my thing, but people could do it with that lens.

Look at the dof calculator:
Online Depth of Field Calculator
You see that your win in dept of field with 500mm is very limited. Your lens needs a little stop down, but doing f7.1 is probably good enough and it won't get better.

Your K-70 is a little handicapt in the way that it has older and different parts inside then for instance the KP.
Pentax K-1 II vs. Pentax KP vs. Pentax K-70 - Pentax Camera Comparison - PentaxForums.com
It has a slower processor and an older version SAFOX X af-module where KP and K-1 share an almost identical design af-module that dates back from the K-3.

Thanks for the response. I'm aware of the depth-of-field limitations, hence my (arbitrary) use of f/8, primarily to improve the depth of field rather than for overall sharpness, which I'm not dissatisfied with. I do "open up" in low light to reduce grain/noise and will use slower shutter speeds when conditions allow.


My K-70 may be "handicapped", but it was all I could afford at the time and is a serious upgrade in capability for this kind of photography from my K-5, which still does an excellent job when I can get sufficient subject matter in the viewfinder. I'm just trying to maximise my success rate

[RonHendriks1966]

My K-70 may be "handicapped", but it was all I could afford at the time and is a serious upgrade in capability for this kind of photography from my K-5, which still does an excellent job when I can get sufficient subject matter in the viewfinder. I'm just trying to maximise my success rate

Sometimes that improved succes can come with more knowledge and sometimes it needs new hardware.

[jbinpg]

I stopped blaming my camera bodies and instead went to blaming the lens focusing internals after I bought the 55-300PLM lens. The simplest way to witness this increase in focusing speed would be to try the PLM to see if it makes a difference for you.

[beholder3]

OK, I think I understand this ... in layman's terms, anything slower than f/5.6 is non-optimal? If that's the case, it supports my suspicions and I need to think about "upgrading" ... or getting closer, somehow

All my Pentax cameras so far have actually worked down to F9 lenses (teleconverters used...), but the official specs of the AF sensors is F5.6 and F2.8 as far as I know. That is how far to the image edges their pair of line sensor look.

[kypfer]

All my Pentax cameras so far have actually worked down to F9 lenses (teleconverters used...), but the official specs of the AF sensors is F5.6 and F2.8 as far as I know. That is how far to the image edges their pair of line sensor look.

Oh yes, the AF on my K-70 will handle a 2x converter on the Sigma 150-500mm in good light on a contrasty subject, so that's effectively f/12.6, but it'll take it's time to lock on. With a 1.5x converter it'll work most of the time at effectively f/9+ if it's a stationary subject in good light, but I'm trying to establish what parameters are necessary for totally reliable operation in most circumstances and it would seem, as you point out, that f/5.6 is needed, especially for a moving subject!
So I'm scouring the second-hand listings and saving the pennies, even the cat has got to catch her own supper (just kidding


Thanks for the detail

[stevebrot]

but the official specs of the AF sensors is F5.6 and F2.8 as far as I know.

Those are the wide-side focus sensitivities. The narrow side for prism black-out is about f/8. FWIW, the f/2.8 sensors are paired with a set of f/5.6 sensors to allow for extended range down to f/8 before blackout. (The difference between the two is effective prism baseline/pitch.)


Steve

[clackers]

Hi Kypfer, the PDAF sensors of nearly all cameras are rated for f5.6. Slower than this, and all bets are off. That's why even kit lenses are typically made no worse than f5.6.

Your K-70 has a central cross sensor that can take advantage of an f2.8 lens for increased accuracy, but anything slower and it reverts to being an f5.6 vertical only sensor like the edge ones.


So you can see why sports and wildlife shooters choose their equipment as they do. A 400 f2.8 is a standard pro's sporting prime. Teleconverters can be added, but the game is to keep the aperture with increasing damage in focus accuracy to f2.8, f4, f5.6 and then f8 by the trick @stevebrot outlines above (central point only IIRC, to choose a tracking AF-C mode or whatever is being hopeful).


Putting a TC on a slow lens like the DA55-300 can work but the sensors are not rated for that, the light is coming in very tight, not covering much area of the sensors, and there will be guessing.

[beholder3]

Those are the wide-side focus sensitivities. The narrow side for prism black-out is about f/8. FWIW, the f/2.8 sensors are paired with a set of f/5.6 sensors to allow for extended range down to f/8 before blackout. (The difference between the two is effective prism baseline/pitch.)


Steve

There is no prism involved in the path of light to the AF sensors and there is no full "blackout" either obviously. It is incremental.

[kypfer]

Hi Kypfer, the PDAF sensors of nearly all cameras are rated for f5.6. Slower than this, and all bets are off. That's why even kit lenses are typically made no worse than f5.6.

So you can see why sports and wildlife shooters choose their equipment as they do. A 400 f2.8 is a standard pro's sporting prime. Teleconverters can be added, but the game is to keep the aperture with increasing damage in focus accuracy to f2.8, f4, f5.6 and then f8 by the trick @stevebrot outlines above (central point only IIRC, to choose a tracking AF-C mode or whatever is being hopeful).

Thanks ... unfortunately 400mm f/2.8's are thin on the ground in Pentax AF mount, new or second-hand, so I'll struggle on with what I have or can find at an affordable price.
I'm not dissatisfied with the Sigma, far from it, especially at the s/h price I paid for it ... just trying to get maximum "bang for my bucks"


Thanks for the input

[clackers]

Thanks ... unfortunately 400mm f/2.8's are thin on the ground in Pentax AF mount, new or second-hand, so I'll struggle on with what I have or can find at an affordable price.
I'm not dissatisfied with the Sigma, far from it, especially at the s/h price I paid for it ... just trying to get maximum "bang for my bucks"


Thanks for the input

Oh, I have that Sigma 140-500, too, Kypfer ... shot a football game with it just a couple of months ago.

But it's not the DFA*150-450. It's a consumer lens, just like the current Sigma and Tamron 150-600s. As you know, its performance drops off so much after 400mm you wonder whether you're better off shooting at a shorter length and then cropping.

A secondhand F*300, FA*300 or DA*300 f4 with the 1.4TC will get that magical number, f5.6, BTW, that means a lot to your camera's focus system.

[kypfer]

Oh, I have that Sigma 140-500, too, Kypfer ... shot a football game with it just a couple of months ago.

But it's not the DFA*150-450. It's a consumer lens, just like the current Sigma and Tamron 150-600s. As you know, its performance drops off so much after 400mm you wonder whether you're better off shooting at a shorter length and then cropping.

A secondhand F*300, FA*300 or DA*300 f4 with the 1.4TC will get that magical number, f5.6, BTW, that means a lot to your camera's focus system.

Thanks again


I must re-iterate at this point that I'm not dissatisfied with the Sigma 150-500mm lens optical performance at any focal length, especially now I've used the "AF Fine Adjustment" option on my K-70 to advantage. I can read the ID numbers on birds leg-rings at 100+ yards if there's not too much heat-haze, which is plenty good enough for me


The only problem I perceived was an occasional "reluctance" to autofocus accurately ... I think we've confirmed my suspicions that I need an extra half-stop or better to help that circumstance, so whilst waiting to win the lottery I'll be carrying my 70-200mm f/2.8 + matching 2x, which may become my "default" lens, fitting the 150-500mm when I need the extra reach for a stationary subject.

[stevebrot]

There is no prism involved in the path of light to the AF sensors and there is no full "blackout" either obviously. It is incremental.

It is technically true that there is no "prism", though the combination of microlens and aperture pairs accomplish the same task of acting as a beam splitter and both approaches work on the same principle. The notion of rangefinder base also applies to both as does my comment. Sorry for the imprecision of terms. As for the "blackout", true again; it is incremental, but the threshold for poor performance comes quickly. Sorry again for falling back on the common analogy to the split-image rangefinder historically common in many SLR cameras and still available for some high-end models for brands other than Pentax.

The analogy is useful because of the broad familiarity in photography circles with the spit-image and also because the blackout behavior is similar as the limits are tested. Also demonstrable is the how/why of AF fine adjustment works...too much stuff for discussion here.


Steve

(...currently uses a split-image screen in the K-3 that features a variable-cut prism to allow both excellent sensitivity at wide apertures and amazing black-out resistance...)