[dosdan]

No. I have a 150-500 sigma, but didn't use it. A faster focusing lens may help closer in, but I suspect that this is the limits of the body. The second to last shot was back focused meaning that the dog was running faster than the focus logic could adjust.

I doubt this is the body. Lens moving mass/inertia, AF (in the body or in the camera) motor torque & total focal movement distance are probably more significant for these bigger zoom lenses than the body AF processing. AF predication isn't exactly rocket science. There should be enough subject-to-background discrimination with the new high-res AE sensor and the denser, high-sensitivity AF sensor array to provide sufficient input data for the AF processing. I would expect firmware revisions to polish the tracking ability further.

My experience with AF performance photographing Long Jump with the K-5. 3 different lenses:

Pentax DA 55-300/F4-5.8
https://www.pentaxforums.com/forums/115-pentax-k-5/115277-k-5-sports-photogra...ml#post2409556

Pentax DA* 60-250/F4
https://www.pentaxforums.com/forums/115-pentax-k-5/115277-k-5-sports-photogra...ml#post2410088

Sigma 70-200/F2.8
https://www.pentaxforums.com/forums/115-pentax-k-5/115277-k-5-sports-photogra...ml#post2410938


Because I couldn't trust the K-5's Auto AF performance here, I usually used centre-point focus. Sometimes the jumper moved above or to the side of the centre and AF went OOF because of the fixed centre-point. It would be good to try this with more effective AF.C tracking.

However, I think other focus problems are the result of the lens not keeping up. I plan to get the K-3 sooner or later, so it will be interesting to repeat these tests, in a more standardised and careful way with all 3 lenses, to see if the lens is the limiting factor in this running-towards-the-camera situation.

Dan.

---

Last edited by dosdan; 11-18-2013 at 03:47 PM.

[jpzk]

The toughest test of the responsiveness of the autofocus system that I've run across is to shoot a dog running towards you. These are with the DA*300, 1/400 f4. The whole series. The camera was working to get focus, and managed to get some. With the K-5 the focus was on the dogs shoulder with the face out of focus, meaning that the acquire focus to expose time was long. The same thing is apparent but less so in these shots. Just to be clear, I never got any shots on my K-5 that were in focus when I did this.

Yes, my dog is nuts. I had to interrupt composing this message because she was tearing around the house. She is a 7 year old Portuguese water dog.

Is the K-3 perfect? No. Can it be better? Yes. I have no idea how these shots would turn out on other bodies, or even other lenses. I did this series of comments to describe the changes and improvements from the K-5. For me it is a remarkable upgrade, fixing and improving things substantially. Others may have different experiences. But most of the problems that I had getting focus have been fixed. Pentax will now need to work on improving their lenses.

The shots from this thread are at https://plus.google.com/u/3/photos/101688500536284805756/albums/5942599665693651185?sort=1

Very timely set and thanks for sharing this, and previous tests you did: very useful for someone like me (and others I presume) who will soon get the K3.

As a side note: I too have the DA*300/4 and used to own the Siggy 150-500 -- which I bought, sold and re-bought recently.

It will be interesting to see how "slow" the Siggy 150-500 will be vs. the DA*300/4 on the K3.

Cheers!

JP

[derekkite]

A bit of calculation.

Assuming the dog is running 25 mph, or 11.176 m/sec, or 11.1 mm/ms.
Assuming that the time from focus acquisition to shutter opening is 100 ms. I'm basing that on the Imaging Resource discussion of the AA filter mechanism. They have a chart plotting the sound of the mirror and shutter over time. In AFC mode it may be different, but for the sake of argument.
The dog covers 1.1m during that time.
The depth of field for the 300mm lens at three meters, f4 is 1.44 cm. At 10 meters the dof is 17.25cm.

Is this right?

Its amazing that it gets any shots at all. In these circumstances the K-5 didn't.

[MRCDH]

This is my experience, not some lab report. I shoot mainly long lenses, and have been shooting with the K-5 with a variety of manual focus and autofocus lenses for a couple of years. I have been using the DA*300 for the most part, excepting the bright summer days when I use a Sigma 150-500 OS. I have had great success with these combinations, but have also run into situations where the limits of the equipment either cause frustration, make it difficult to get a sharp photo, or don't even bother because I don't want to waste time. I can get birds in flight with my manual focus lens, but not very many. The conditions that I shoot in range from perfect to utterly miserable. If I wait for the perfect day, I would shoot maybe half a dozen days a year. This isn't lab or studio conditions, but real conditions that I shoot in, and if I may, get some very nice shots from time to time.

There are some situations where I've run into problems

1. Low light focus. It was a regular occurrence, even during the summer, to have the K-5 and DA*300 not focus at all. There wasn't enough light, enough contrast, and it would either try and fail, or not even try. In good light with sharp contrast it worked fine. Some particular situations where this was an issue was when trying to focus on wildlife in the trees or bushes.

2. Another situation where it was difficult to acquire focus was when shooting a small subject on a horizontal plane. For example, last week I ran across an American Pipit that sat obligingly in the middle of a gravel road. The K-5/DA*300 would either focus on a detail on the gravel, or the grass and bushes behind. The same thing happens when shooting water birds at a distance. The contrast of choppy water or the far shore would get focussed, and not the desired subject. I believe the cause was a combination of the low light inadequacy and the large focus points.

3. A third situation is what I call threading the needle. This is where a subject is in a busy scene with many contrasty elements in the fore and back ground. Not an uncommon situation in many contexts; for me it was a bird in the bushes. The large focus points and the inconsistency or unpredictable focus logic made it an exercise in frustration to try to thread the focus point in between obstacles to find and home in on what I wanted. A longer focal length helps, magnifying the subject so that the camera has a better chance of focusing on it, but that wasn't the only problem.

4. Moving subjects were very difficult to focus on. Many things conspire to make this difficult. A large focus point, the inability to pick out low contrast subjects from a busy background, as well as the speed of the focus-mirror-open shutter timing gives you lots of out of focus shots. There are quite a few discrete challenges here. For example, a bird flying against the sky is not difficult to focus on, but the speed of the image capture train will cause a problem if the direction of movement is towards out of focus. If the bird is flying against a tree or mountain background, then the low contrast as well as large focus point makes it probable that the background will be in focus. If the subject is moving towards the camera the speed of the adjustments is critical. The K-5 is a challenge to focus properly in these circumstances.

5. Using live view had its own challenges. Focusing my DA*300 mounted on a tripod using the CDAF in Live view was unsatisfactory, most times unsuccessful. Using a manual focus lens, the magnified Liveview screen was helpful, but the delay in updating the screen meant that it was easy to overshoot the focus adjustments.

First impressions? The limits of focus are the obstreperous subject or inept operator. The DA*300 felt slow and clumsy on the K-5. It feels lightning fast and accurate on the K-3. The shutter noise is different, not louder. The buttons feel more positive than the mushy K-5 feel.

The center focus point is small. I took these shots with spot focus, afc. I followed the bird around, and when I was able to get the point on the bird, it focused. It was a technical challenge, but a positive one; if I point it at something, it will focus quickly. Here are some shots. They are processed in Darktable, compiled from the development branch. There is no profiling of the K-3 coded as yet, but I had no problem reading the files or doing corrections. They are cropped, some substantially to frame what I wanted to shoot, often the limit of cropping before losing quality. Some exposure and contrast adjustments were made, as well as noise reduction.

This is the fourth shot I took with the camera. 1/400, iso 200 f4.


This is what it was doing. Note that the focus is on the background. 1/800 iso 400 f4


This was a long way away. I've shot this site numerous times but the K-3 was able to focus on what I wanted. iso 1600 1/2000 f4. Cropped.


The focus point on this shot was about the size of it's head. Cropped, 1/320 iso 160 f4.


Easy shot, especially when the focus mechanism ignores the intervening brush. 1/640 iso 1000 f4, cropped. Nice and sharp.


This is where I was impressed. Not a great shot, but illustrative. There were a bunch of chickadees, brown creepers and rather frenetic yellow crowned kinglets working their way through the bush. I shot this guy for quite a few minutes as he went from tree to tree. I had it on 7 point with the center as main. I was able to easily shoot through intervening brush to focus. The challenge was low light, low contrast subject and movement, focus was simply just working. I didn't bother doing this type of shooting with the K-5, it didn't work and was very frustrating. I will do much more, I enjoy this.


So far, low light is not an issue at all. If there is a surface to focus on, it will. The size of the focus point is very nice for what I shoot. If it is there, it will focus on it. My lens feels brand new, fast and confident. So far I'm very pleased. Now to test the more powerful and complex capabilities.

Thanks for starting this thread and I enjoyed reading the feedback that is contained and will check periodically for updates.