My K-3 AF experiences

derekkite · 11-08-2013, 08:37 PM

A few observations using other lenses. The F 100mm 2.8 macro lens focuses accurately and quickly. There is an abundance of mushrooms this fall, and I've been taking photos of them. I'm not an experience macro photographer and am figuring out how to do it well. Last weekend I had my K-5 and macro lens. It worked reasonably well, but in many instances, the mushrooms were under a tree on that wet cloudy day, and it wouldn't focus at all. The K-3 at worst will hunt once before finding focus accurately.

The DA12-24 focuses quickly and accurately. Not a tough test; at f8 it is like a point and shoot. I did a night time shot, 30 second exposure of the bridge near here, and it focused quickly, no hunting. I think I remember it hunting on the K-5 in low light.

DAL 55-300 is much better on the K-3. I couldn't stand using it on my K-5, even though it was pretty nice at F8. It sounded like a rough dentist bit in my mouth, back and forth, back and forth. On the K-3 it is still noisy, but it gets to focus and stops. It is tolerable.

I'm getting to appreciate more and more the stunning IQ of the K-5. The K-3 is improved with no AA filter and more resolution, but not a revolutionary improvement like was the K-5.

Last edited by derekkite; 11-08-2013 at 08:42 PM.

Sandy Hancock · 11-09-2013, 02:40 AM

I did my first test of the K-3 to track moving things this afternoon. My two year old labrador pelting around the garden after a ball seemed a reasonable subject.

I uploaded the FULL SEQUENCE to flickr. This was shot at f/2.8, so while they are not all tack sharp I suspect it is more to do with where the camera focussed rather than its ability to track at high speed.

I was quite impressed. I don't think my K-5 IIs would have been close to this.

JinDesu · 11-09-2013, 05:55 AM

Originally posted by Sandy Hancock

I did my first test of the K-3 to track moving things this afternoon. My two year old labrador pelting around the garden after a ball seemed a reasonable subject.

I uploaded the FULL SEQUENCE to flickr. This was shot at f/2.8, so while they are not all tack sharp I suspect it is more to do with where the camera focussed rather than its ability to track at high speed.

I was quite impressed. I don't think my K-5 IIs would have been close to this.

Towards the end where it seems the camera caught the face and held on, that looked quite good.

audiobomber · 11-09-2013, 06:20 AM

Originally posted by JinDesu

Towards the end where it seems the camera caught the face and held on, that looked quite good.

Dog Face Detection?

JinDesu · 11-09-2013, 06:26 AM

Originally posted by audiobomber

Dog Face Detection?

'Twas the smile that did it.

derekkite · 11-10-2013, 06:24 PM

Often I see waterfowl on the nearby lake and take a photo to identify and record the species. These aren't usually shots of any merit except for the purpose of identification. The K-5 focus was a real challenge; it wouldn't focus on the birds but something else. The K-3 is much better. I saw some small brownish birds quite a ways off shore, and took some shots. The first one is not cropped.

I'm pleased. A longer lens and good light might make one of these printable, but that wasn't what I was after. I wanted a clear shot, and got one.

Dionisis_K · 11-11-2013, 05:26 AM

Hi. My biggest problem with K-5 is its problem to focus correctly in tungsten light.. Almost always it missfocusses.. Can you confirm that this is not a problem anymore with k-3?? Thanks in advance.

rawr · 11-11-2013, 06:10 AM

Originally posted by Dionisis_K

Can you confirm that this is not a problem anymore with k-3??

This problem was fixed from the K-30 onwards, inc K-5 II.
If it shows up in the K-3, I will be very surprised.

**Michaelina2** · 11-11-2013, 07:19 AM

Originally posted by frogoutofwater

I have learned the basics of Photoshop and Lightroom 5, but need to improve here as well - especially fine-tuning noise reduction and sharpening. I'm wondering if there is a better program out there for noise reduction.

Consider the NIK Collection for LR. While LR's noise tools have improved over time, I find NIK's Define to work quite well on noise in default mode. Define also allows localized noise suppression, so you can attack a small section of an image that's giving you trouble while leaving the rest alone. I find the ability to easily localize refinements a real plus. Finally, NIK has an extensive collection of excellent (and free) on-line tutorials that demonstrate how all of this works.

Cheers... M

RockvilleBob · 11-11-2013, 01:29 PM - **1 Like**

Originally posted by Michaelina2

Consider the NIK Collection for LR. While LR's noise tools have improved over time, I find NIK's Define to work quite well on noise in default mode. Define also allows localized noise suppression, so you can attack a small section of an image that's giving you trouble while leaving the rest alone. I find the ability to easily localize refinements a real plus. Finally, NIK has an extensive collection of excellent (and free) on-line tutorials that demonstrate how all of this works.

Cheers... M

Another vote for NIK software. I do all my edits with Adobe Camera Raw and then NIK Software using the Photoshop Plugins. The only reason Photoshop gets started is to run Bridge, Camera Raw and the NIK plug In.
I have been using NIK for years - very user friendly meaning I can use them without much training and remembering a zillion short cuts. NIK was acquired by Google a year or so ago. I sent an email asking about future upgrades and enhancements to NIK. The NIK software team is in Germany and they replied they are at work on new versions/products. I thought NIK was a bargain at $400 or more for the complete set, for the price Google charges I feel it is a must have addition if you do post processing.

derekkite · 11-11-2013, 09:43 PM

An issue with the K-5 was the slow response in live view making it difficult to focus manually. It was so slow that as you turned the focus ring you overshot. I mounted an old manual 300 2.8 on the K-3, put it on a tripod. It responds quickly with no appreciable lag, and the focus peaking makes it clear when you are in focus.

This also makes it easy to check and do fine autofocus adjustments to lenses. I put my DA*300 on the tripod, put a focus chart on the floor and had it at 45 degrees. Focus, then hit live view, and the focus peaking shows exactly how much it is out. The DA*300 was right on, the 150-500 was front focusing considerably, and the F 100 2.8 macro was back focusing a bit.

bigted · 11-11-2013, 10:01 PM

Thanks so much for doing this. Like others have said it puts the K-3 in real situations. Many of your birding frustrations have been mine , but I use a K-r. I also appreciate the suggestion of spending some time with the manual. I think I'll be joining JP and doing some reading while I get the finds together to pick up a K-3.

derekkite · 11-13-2013, 09:41 PM

I had the opportunity to shoot some birds in flight today. The light wasn't exceptional; cloudy and low. There were some gulls floating on the wind over an area they were interested in, so they stayed within range for the 300mm lens. I took 34 shots in three sequences, and 3 were soft due to not sharp focus. One was way out, the camera lost it and focused on the mountains in the background. The other two were just off. Most shot are not worth looking at twice due to motion blur; I was at 1/500 to keep the iso down, 1000-1250, f4 on the lens. In this light if you speed up the shutter noise gets nasty and detail is lost. Hold status medium, sel 27 for the af.c mode. Metering is center weighted. I found spot to be very precise to the point of causing problems; my white and black dog would throw the exposure off to either extreme depending on which was read by the spot. I was shooting a mallard, reasonably close in, and as it moved the dark or light parts of it's body set the metering. This is something that can be used, and I will use it, but it is a bit extreme for general shooting. I have the three focus settings to focus priority.

It will take practice to get good at this. The initial focus has to be right on with the center point. Spot af is very difficult for following a bird in flight because you have to keep it precisely in the small center point area, else it focuses on something else. The extended area helps considerably, but you need to acquire focus at the start. Once done, the body seems to keep focus reasonably well. It isn't omniscient, but it doesn't fight you either.

These shots are from the third sequence. The bird was hovering on the gusty wind, eventually landing on the water. Not an extreme example, but all were in focus. I didn't crop any of the shots, and this is the series. Noise reduction and shadows and highlight correction applied. Handheld, with SR on, Tv mode 1/500.
I have had few opportunities to shoot a bird coming towards me or away from me, so I can't comment on how that will turn out.

dosdan · 11-13-2013, 10:30 PM

Originally posted by derekkite

. I was shooting a mallard, reasonably close in, and as it moved the dark or light parts of it's body set the metering

It sounds like AE is linked the the AF point. You should be able to decouple this:

Menu C1
5. Link AE and AF Point
Sets whether to link the exposure and AF point in the focusing area.
Default Setting: Off

Dan.

derekkite · 11-13-2013, 10:48 PM

I do have that set. I'll try turning it off to see it if changes anything. Would it change the size of the spot that the metering uses? I set the exposure mode to center weighting, and it seems to be more consistent.

I was experimenting with the flash, specifically triggering a slave with the built-in flash. It worked quite well and consistent, except when the exposure mode was set to spot. I would point it at a white part of the dog, and it would underexpose. Pointed at the black part, it would overexpose. Quite dramatically. The spot seems as small as the af point. I'm certain the K-5 wasn't as dramatic as this.

I can see situations where that would be very useful, being able to control highlights in some circumstances, or get the exact exposure for a specific spot. It will also cause issues for some people if they are not aware of it. But for most of the time, the center weighting seems to work.

Another question arose. Since the tracking function uses the metering array, does the metering mode have any effect on the af tracking performance?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K-5 with a Metz 44 AF-1 ... any experiences?	Jean Poitiers	Pentax K-5 & K-5 II	28	08-15-2014 03:45 AM
K-3 AF-C Test Picts	Winder	Pentax K-3 & K-3 II	64	11-05-2013 11:09 PM
Wedding shooter thinking about upgrading my K-5 to a K-3 but...	enoeske	Pentax K-3 & K-3 II	18	11-03-2013 09:07 AM
Newbie K-x Owner Looking for Opinions about Tamron AF 18-200mm f/3.5-6.3 XR Di II LD	iankh	Troubleshooting and Beginner Help	2	02-08-2010 04:57 AM
My London experiences(lots of pics)	vdubbin099	Post Your Photos!	8	10-09-2007 11:25 AM

11-08-2013, 08:37 PM	#31
derekkite Pentaxian Join Date: Sep 2011 Location: Nelson B.C. Posts: 3,782 Original Poster	A few observations using other lenses. The F 100mm 2.8 macro lens focuses accurately and quickly. There is an abundance of mushrooms this fall, and I've been taking photos of them. I'm not an experience macro photographer and am figuring out how to do it well. Last weekend I had my K-5 and macro lens. It worked reasonably well, but in many instances, the mushrooms were under a tree on that wet cloudy day, and it wouldn't focus at all. The K-3 at worst will hunt once before finding focus accurately. The DA12-24 focuses quickly and accurately. Not a tough test; at f8 it is like a point and shoot. I did a night time shot, 30 second exposure of the bridge near here, and it focused quickly, no hunting. I think I remember it hunting on the K-5 in low light. DAL 55-300 is much better on the K-3. I couldn't stand using it on my K-5, even though it was pretty nice at F8. It sounded like a rough dentist bit in my mouth, back and forth, back and forth. On the K-3 it is still noisy, but it gets to focus and stops. It is tolerable. I'm getting to appreciate more and more the stunning IQ of the K-5. The K-3 is improved with no AA filter and more resolution, but not a revolutionary improvement like was the K-5. Last edited by derekkite; 11-08-2013 at 08:42 PM.

11-09-2013, 02:40 AM	#32
Sandy Hancock Moderator Site Supporter Join Date: Feb 2012 Location: Adelaide Hills, South Australia Photos: Gallery \| Albums Posts: 11,275	I did my first test of the K-3 to track moving things this afternoon. My two year old labrador pelting around the garden after a ball seemed a reasonable subject. I uploaded the FULL SEQUENCE to flickr. This was shot at f/2.8, so while they are not all tack sharp I suspect it is more to do with where the camera focussed rather than its ability to track at high speed. I was quite impressed. I don't think my K-5 IIs would have been close to this.

11-09-2013, 05:55 AM	#33
JinDesu Veteran Member Join Date: Jun 2011 Location: New York City Photos: Gallery Posts: 5,638	Originally posted by Sandy Hancock I did my first test of the K-3 to track moving things this afternoon. My two year old labrador pelting around the garden after a ball seemed a reasonable subject. I uploaded the FULL SEQUENCE to flickr. This was shot at f/2.8, so while they are not all tack sharp I suspect it is more to do with where the camera focussed rather than its ability to track at high speed. I was quite impressed. I don't think my K-5 IIs would have been close to this. Towards the end where it seems the camera caught the face and held on, that looked quite good.

11-09-2013, 06:20 AM	#34
audiobomber Veteran Member Join Date: May 2008 Location: Sudbury, Ontario Photos: Albums Posts: 6,806	Originally posted by JinDesu Towards the end where it seems the camera caught the face and held on, that looked quite good. Dog Face Detection?

11-09-2013, 06:26 AM	#35
JinDesu Veteran Member Join Date: Jun 2011 Location: New York City Photos: Gallery Posts: 5,638	Originally posted by audiobomber Dog Face Detection? 'Twas the smile that did it.

11-10-2013, 06:24 PM	#36
derekkite Pentaxian Join Date: Sep 2011 Location: Nelson B.C. Posts: 3,782 Original Poster	Often I see waterfowl on the nearby lake and take a photo to identify and record the species. These aren't usually shots of any merit except for the purpose of identification. The K-5 focus was a real challenge; it wouldn't focus on the birds but something else. The K-3 is much better. I saw some small brownish birds quite a ways off shore, and took some shots. The first one is not cropped. I'm pleased. A longer lens and good light might make one of these printable, but that wasn't what I was after. I wanted a clear shot, and got one.

11-11-2013, 05:26 AM	#37
Dionisis_K Senior Member Join Date: Nov 2010 Location: Athens - Greece Photos: Albums Posts: 249	Hi. My biggest problem with K-5 is its problem to focus correctly in tungsten light.. Almost always it missfocusses.. Can you confirm that this is not a problem anymore with k-3?? Thanks in advance.

11-11-2013, 06:10 AM	#38
rawr Veteran Member Join Date: Jun 2009 Posts: 11,913	Originally posted by Dionisis_K Can you confirm that this is not a problem anymore with k-3?? This problem was fixed from the K-30 onwards, inc K-5 II. If it shows up in the K-3, I will be very surprised.

11-11-2013, 07:19 AM	#39
Michaelina2 Site Supporter Join Date: Jan 2008 Location: SE Michigan USA Photos: Gallery Posts: 1,302	Originally posted by frogoutofwater I have learned the basics of Photoshop and Lightroom 5, but need to improve here as well - especially fine-tuning noise reduction and sharpening. I'm wondering if there is a better program out there for noise reduction. Consider the NIK Collection for LR. While LR's noise tools have improved over time, I find NIK's Define to work quite well on noise in default mode. Define also allows localized noise suppression, so you can attack a small section of an image that's giving you trouble while leaving the rest alone. I find the ability to easily localize refinements a real plus. Finally, NIK has an extensive collection of excellent (and free) on-line tutorials that demonstrate how all of this works. Cheers... M

11-11-2013, 01:29 PM - 1 Like	#40
RockvilleBob Veteran Member Join Date: Dec 2011 Location: Lewes DE USA Photos: Albums Posts: 1,780	Originally posted by Michaelina2 Consider the NIK Collection for LR. While LR's noise tools have improved over time, I find NIK's Define to work quite well on noise in default mode. Define also allows localized noise suppression, so you can attack a small section of an image that's giving you trouble while leaving the rest alone. I find the ability to easily localize refinements a real plus. Finally, NIK has an extensive collection of excellent (and free) on-line tutorials that demonstrate how all of this works. Cheers... M Another vote for NIK software. I do all my edits with Adobe Camera Raw and then NIK Software using the Photoshop Plugins. The only reason Photoshop gets started is to run Bridge, Camera Raw and the NIK plug In. I have been using NIK for years - very user friendly meaning I can use them without much training and remembering a zillion short cuts. NIK was acquired by Google a year or so ago. I sent an email asking about future upgrades and enhancements to NIK. The NIK software team is in Germany and they replied they are at work on new versions/products. I thought NIK was a bargain at $400 or more for the complete set, for the price Google charges I feel it is a must have addition if you do post processing.

11-11-2013, 09:43 PM	#41
derekkite Pentaxian Join Date: Sep 2011 Location: Nelson B.C. Posts: 3,782 Original Poster	An issue with the K-5 was the slow response in live view making it difficult to focus manually. It was so slow that as you turned the focus ring you overshot. I mounted an old manual 300 2.8 on the K-3, put it on a tripod. It responds quickly with no appreciable lag, and the focus peaking makes it clear when you are in focus. This also makes it easy to check and do fine autofocus adjustments to lenses. I put my DA300 on the tripod, put a focus chart on the floor and had it at 45 degrees. Focus, then hit live view, and the focus peaking shows exactly how much it is out. The DA300 was right on, the 150-500 was front focusing considerably, and the F 100 2.8 macro was back focusing a bit.

11-11-2013, 10:01 PM	#42
bigted Moderator Join Date: Oct 2011 Location: Jessica Lake, Whiteshell Provincial Park, or Winnipeg Manitoba, Canada Photos: Gallery \| Albums Posts: 1,594	Thanks so much for doing this. Like others have said it puts the K-3 in real situations. Many of your birding frustrations have been mine , but I use a K-r. I also appreciate the suggestion of spending some time with the manual. I think I'll be joining JP and doing some reading while I get the finds together to pick up a K-3.

11-13-2013, 09:41 PM	#43
derekkite Pentaxian Join Date: Sep 2011 Location: Nelson B.C. Posts: 3,782 Original Poster	I had the opportunity to shoot some birds in flight today. The light wasn't exceptional; cloudy and low. There were some gulls floating on the wind over an area they were interested in, so they stayed within range for the 300mm lens. I took 34 shots in three sequences, and 3 were soft due to not sharp focus. One was way out, the camera lost it and focused on the mountains in the background. The other two were just off. Most shot are not worth looking at twice due to motion blur; I was at 1/500 to keep the iso down, 1000-1250, f4 on the lens. In this light if you speed up the shutter noise gets nasty and detail is lost. Hold status medium, sel 27 for the af.c mode. Metering is center weighted. I found spot to be very precise to the point of causing problems; my white and black dog would throw the exposure off to either extreme depending on which was read by the spot. I was shooting a mallard, reasonably close in, and as it moved the dark or light parts of it's body set the metering. This is something that can be used, and I will use it, but it is a bit extreme for general shooting. I have the three focus settings to focus priority. It will take practice to get good at this. The initial focus has to be right on with the center point. Spot af is very difficult for following a bird in flight because you have to keep it precisely in the small center point area, else it focuses on something else. The extended area helps considerably, but you need to acquire focus at the start. Once done, the body seems to keep focus reasonably well. It isn't omniscient, but it doesn't fight you either. These shots are from the third sequence. The bird was hovering on the gusty wind, eventually landing on the water. Not an extreme example, but all were in focus. I didn't crop any of the shots, and this is the series. Noise reduction and shadows and highlight correction applied. Handheld, with SR on, Tv mode 1/500. I have had few opportunities to shoot a bird coming towards me or away from me, so I can't comment on how that will turn out.

11-13-2013, 10:30 PM	#44
dosdan Pentaxian Join Date: Dec 2007 Location: Brisbane, Australia Photos: Gallery \| Albums Posts: 1,741	Originally posted by derekkite . I was shooting a mallard, reasonably close in, and as it moved the dark or light parts of it's body set the metering It sounds like AE is linked the the AF point. You should be able to decouple this: Menu C1 5. Link AE and AF Point Sets whether to link the exposure and AF point in the focusing area. Default Setting: Off Dan.

11-13-2013, 10:48 PM	#45
derekkite Pentaxian Join Date: Sep 2011 Location: Nelson B.C. Posts: 3,782 Original Poster	I do have that set. I'll try turning it off to see it if changes anything. Would it change the size of the spot that the metering uses? I set the exposure mode to center weighting, and it seems to be more consistent. I was experimenting with the flash, specifically triggering a slave with the built-in flash. It worked quite well and consistent, except when the exposure mode was set to spot. I would point it at a white part of the dog, and it would underexpose. Pointed at the black part, it would overexpose. Quite dramatically. The spot seems as small as the af point. I'm certain the K-5 wasn't as dramatic as this. I can see situations where that would be very useful, being able to control highlights in some circumstances, or get the exact exposure for a specific spot. It will also cause issues for some people if they are not aware of it. But for most of the time, the center weighting seems to work. Another question arose. Since the tracking function uses the metering array, does the metering mode have any effect on the af tracking performance?