K1 AF issues at a wedding...

biz-engineer · 10-17-2016, 12:26 PM

Originally posted by Qwntm

This might also account for the poster who said his birding shots had slow AF. If your waiting around for the birds to show up, the camera could power off. I'm still testing all this but it seems to be the reason for my issue so far...

Not sure about this. When I receive the K1 , during the first couple of weeks I tested it, I saw this hesitating AF randomly occurring, especially in low light. Now I'm not experiencing it anymore, and I don't know why.

**UncleVanya** · 10-17-2016, 12:29 PM

Originally posted by Qwntm

I think I figured out the AF delay.

I had the K1 set to "Auto Power Off: 1 Minute." Which usually is no big deal because when you are using the camera it usually stays on while you are using it. But at a wedding there are times when you stand around for 5 minutes or more without touching the camera between shots. If the camera "Auto's Off" it there is a delay while it is restarting and the AF has a lag time. If the camera is awake, the AF is pretty much instant even in dim light.

So next wedding, Auto Power Off will be set to "Off" and we'll see how we go.

This might also account for the poster who said his birding shots had slow AF. If your waiting around for the birds to show up, the camera could power off.

I'm still testing all this but it seems to be the reason for my issue so far...

Is there a setting between 1 and infinity? Maybe 30 mins?
Nice to hear you are finding the problem that related to your issues and it seems easy to correct.

Unregistered User · 10-17-2016, 12:35 PM

Originally posted by UncleVanya

Is there a setting between 1 and infinity? Maybe 30 mins?
Nice to hear you are finding the problem that related to your issues and it seems easy to correct.

Minutes 1, 3, 5, 10, 30 and OFF are available.

MJKoski · 10-17-2016, 12:49 PM

It is the bubblegum tamron lens which fails there. It is unknown whether the AF somehow tries to fix focus shift by stopping down when focusing or not but anyway my 24-70 was very sluggish compared to tamron 24-70 on D810 which I tried out when trying to find an alternate to Nikkor 24-70.

If focus shift is worked around by stopping down when focusing there is not that much light available to AF module anymore and focusing becomes very sluggish.

Unregistered User · 10-17-2016, 12:55 PM

Originally posted by MJKoski

It is the bubblegum tamron lens which fails there. It is unknown whether the AF somehow tries to fix focus shift by stopping down when focusing or not but anyway my 24-70 was very sluggish compared to tamron 24-70 on D810 which I tried out when trying to find an alternate to Nikkor 24-70.

If focus shift is worked around by stopping down when focusing there is not that much light available to AF module anymore and focusing becomes very sluggish.

It does not stop down. It calculates where to set focus according to choosen aperture.

Winder · 10-17-2016, 01:15 PM

Originally posted by rawr

All the metering data is there in the K-1/K-3/645Z EXIF. Abundantly so.

That EXIF data lets you reconstruct with a great deal of precision exactly what the K-1 metering sensor sees, like so:

And the above represents only about 20% of the data the Pentax RGB sensor records.

Interesting. I would like to know just how much detail the RGB sensor is cable of.

rawr · 10-17-2016, 02:03 PM

Originally posted by Winder

I would like to know just how much detail the RGB sensor is cable of.

Well, the metering sensor 'resolution' is 45x30 (3:2) segments - ie 1350 segments in a 45x30 matrix, like so:

.

But in total there are 6750 data segments recorded by the chip: 4050 segments for the RGB (ie 3 X 1350 data segment blocks, one for each primary colour group) + 1350 segments for flash metering + another 1350 segments for slave flash metering.

So it's not just 'resolution' but [colour] depth too.

It's all a lot more than the old 77 (7x11 matrix) segments of the K-5, 645D, K-S2 etc meter:

Last edited by rawr; 10-17-2016 at 02:21 PM.

beholder3 · 10-17-2016, 02:27 PM

But the K-1 area actually is not fully centered, as you suggest with the red grid.
It is shifted towards the upper end.

Look at the Excel metering image and the roof of the shack in the top right corner.

Did you check my speculation on the colors and do you know in which order they are captured (if) in the metadata?

rawr · 10-17-2016, 02:34 PM

Originally posted by beholder3

do you know in which order they are captured (if) in the metadata?

It seems to be green/yellow, blue, then red, as one can see here:

beholder3 · 10-17-2016, 02:40 PM

Thanks, that is quite interesting.

And there are still so many fields in metadata which nobody has decoded yet.

Qwntm · 10-17-2016, 09:03 PM

Originally posted by biz-engineer

Not sure about this. When I receive the K1 , during the first couple of weeks I tested it, I saw this hesitating AF randomly occurring, especially in low light. Now I'm not experiencing it anymore, and I don't know why.

That's interesting and weird. Who knows. It's going to be hard to pin down. Time will tell I guess.

**Digitalis** · 10-17-2016, 10:13 PM

Interesting data you have uncovered there Rawr, is there any indication of there being a UV/IR blocking filter covering the metering sensor? the Red channel data seems to be abnormally saturated. Though this could also be contributed to a mismatch between the primary sensor CFA and the metering sensor CFA.

rawr · 10-18-2016, 12:08 AM

Originally posted by Digitalis

is there any indication of there being a UV/IR blocking filter covering the metering sensor?

From the hardware images of the sensor that I've seen, it's difficult to say. I don't think there is any such filter though.

In the hardware cutaways of the K-1 where you can see the RGB sensor, all I can discern is a small condenser lens in front of the sensor.

But it is hard to to be conclusive about what's in front of the RGB sensor without a more detailed image, and some information about the optical properties of whatever may be in front of the RGB sensor, and how that light is subsequently processed.

However it does not seem customary with other brand's to put an IR filter in front of their RGB sensors. I haven't seen mention of it in Nikon or Canon, for example. Canon even makes a special point about the RGB sensor of the 5D IV being sensitive to IR. So it is probably useful for the metering system to be aware [to a degree] of IR wavelengths.

**Digitalis** · 10-18-2016, 01:11 AM

Originally posted by rawr

it is probably useful for the metering system to be aware [to a degree] of IR wavelengths.

The reason why I brought up the issue of IR sensitivity in the metering system is due to the fact that IR focuses differently than visible light- it also predisposes the metering system to inaccuracies as CMOS sensors are particularly sensitive to it. And according to Pentax: on the K1, the AE system has has input on the AF system there may be some fine tuning to cancel out any discrepancy in spectral sensitivity between the sensors being used.

rawr · 10-18-2016, 03:55 AM

Canon mentions the IR capabilities of the metering sensor in many of it's newer cameras (like the 5DIV, 7DII and even eg the 80D) in terms of the contribution IR can generally make to more accurate exposure, perhaps including light source flicker detection, and maybe skin tone metering too.

Skin tone is important obviously for general shooting and portraiture, but is extra important for face detection to work reliably via a limited resolution sensor like a RGB sensor. Since some of those Canon's can do face detection AF via optical viewfinder shooting (like some Nikon's can), perhaps the problems of working with IR vs visible light can be compensated for, and balanced out by the ability of IR to help the camera meter find face targets.??

10-18-2016, 12:08 AM	#43
rawr Veteran Member Join Date: Jun 2009 Posts: 11,913	Originally posted by Digitalis is there any indication of there being a UV/IR blocking filter covering the metering sensor? From the hardware images of the sensor that I've seen, it's difficult to say. I don't think there is any such filter though. In the hardware cutaways of the K-1 where you can see the RGB sensor, all I can discern is a small condenser lens in front of the sensor. But it is hard to to be conclusive about what's in front of the RGB sensor without a more detailed image, and some information about the optical properties of whatever may be in front of the RGB sensor, and how that light is subsequently processed. However it does not seem customary with other brand's to put an IR filter in front of their RGB sensors. I haven't seen mention of it in Nikon or Canon, for example. Canon even makes a special point about the RGB sensor of the 5D IV being sensitive to IR. So it is probably useful for the metering system to be aware [to a degree] of IR wavelengths. Last edited by rawr; 10-18-2016 at 12:15 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Quality Control issues with K1	tomO2013	Pentax K-1 & K-1 II	107	02-24-2017 11:41 PM
People Some wedding shots from the K1	tomO2013	Post Your Photos!	4	07-26-2016 12:28 PM
K1 Wedding Photography Impressions	walro	Pentax K-1 & K-1 II	21	06-03-2016 12:41 PM
Shooting at a wedding.	LeDave	Photographic Industry and Professionals	12	09-28-2015 01:44 AM

10-17-2016, 12:26 PM	#31
biz-engineer Pentaxian Join Date: Feb 2015 Photos: Gallery Posts: 12,239	Originally posted by Qwntm This might also account for the poster who said his birding shots had slow AF. If your waiting around for the birds to show up, the camera could power off. I'm still testing all this but it seems to be the reason for my issue so far... Not sure about this. When I receive the K1 , during the first couple of weeks I tested it, I saw this hesitating AF randomly occurring, especially in low light. Now I'm not experiencing it anymore, and I don't know why.

10-17-2016, 12:29 PM	#32
UncleVanya Loyal Site Supporter Join Date: Jul 2014 Photos: Gallery \| Albums Posts: 28,445	Originally posted by Qwntm I think I figured out the AF delay. I had the K1 set to "Auto Power Off: 1 Minute." Which usually is no big deal because when you are using the camera it usually stays on while you are using it. But at a wedding there are times when you stand around for 5 minutes or more without touching the camera between shots. If the camera "Auto's Off" it there is a delay while it is restarting and the AF has a lag time. If the camera is awake, the AF is pretty much instant even in dim light. So next wedding, Auto Power Off will be set to "Off" and we'll see how we go. This might also account for the poster who said his birding shots had slow AF. If your waiting around for the birds to show up, the camera could power off. I'm still testing all this but it seems to be the reason for my issue so far... Is there a setting between 1 and infinity? Maybe 30 mins? Nice to hear you are finding the problem that related to your issues and it seems easy to correct.

10-17-2016, 12:35 PM	#33
Unregistered User Guest	Originally posted by UncleVanya Is there a setting between 1 and infinity? Maybe 30 mins? Nice to hear you are finding the problem that related to your issues and it seems easy to correct. Minutes 1, 3, 5, 10, 30 and OFF are available.

10-17-2016, 12:49 PM	#34
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	It is the bubblegum tamron lens which fails there. It is unknown whether the AF somehow tries to fix focus shift by stopping down when focusing or not but anyway my 24-70 was very sluggish compared to tamron 24-70 on D810 which I tried out when trying to find an alternate to Nikkor 24-70. If focus shift is worked around by stopping down when focusing there is not that much light available to AF module anymore and focusing becomes very sluggish.

10-17-2016, 12:55 PM	#35
Unregistered User Guest	Originally posted by MJKoski It is the bubblegum tamron lens which fails there. It is unknown whether the AF somehow tries to fix focus shift by stopping down when focusing or not but anyway my 24-70 was very sluggish compared to tamron 24-70 on D810 which I tried out when trying to find an alternate to Nikkor 24-70. If focus shift is worked around by stopping down when focusing there is not that much light available to AF module anymore and focusing becomes very sluggish. It does not stop down. It calculates where to set focus according to choosen aperture.

10-17-2016, 01:15 PM	#36
Winder Veteran Member Join Date: Apr 2010 Location: Tennessee Posts: 6,617	Originally posted by rawr All the metering data is there in the K-1/K-3/645Z EXIF. Abundantly so. That EXIF data lets you reconstruct with a great deal of precision exactly what the K-1 metering sensor sees, like so: And the above represents only about 20% of the data the Pentax RGB sensor records. Interesting. I would like to know just how much detail the RGB sensor is cable of.

10-17-2016, 02:03 PM	#37
rawr Veteran Member Join Date: Jun 2009 Posts: 11,913	Originally posted by Winder I would like to know just how much detail the RGB sensor is cable of. Well, the metering sensor 'resolution' is 45x30 (3:2) segments - ie 1350 segments in a 45x30 matrix, like so: . But in total there are 6750 data segments recorded by the chip: 4050 segments for the RGB (ie 3 X 1350 data segment blocks, one for each primary colour group) + 1350 segments for flash metering + another 1350 segments for slave flash metering. So it's not just 'resolution' but [colour] depth too. It's all a lot more than the old 77 (7x11 matrix) segments of the K-5, 645D, K-S2 etc meter: Last edited by rawr; 10-17-2016 at 02:21 PM.

10-17-2016, 02:27 PM	#38
beholder3 Pentaxian Join Date: Dec 2011 Posts: 3,112	But the K-1 area actually is not fully centered, as you suggest with the red grid. It is shifted towards the upper end. Look at the Excel metering image and the roof of the shack in the top right corner. Did you check my speculation on the colors and do you know in which order they are captured (if) in the metadata?

10-17-2016, 02:34 PM	#39
rawr Veteran Member Join Date: Jun 2009 Posts: 11,913	Originally posted by beholder3 do you know in which order they are captured (if) in the metadata? It seems to be green/yellow, blue, then red, as one can see here:

10-17-2016, 02:40 PM	#40
beholder3 Pentaxian Join Date: Dec 2011 Posts: 3,112	Thanks, that is quite interesting. And there are still so many fields in metadata which nobody has decoded yet.

10-17-2016, 09:03 PM	#41
Qwntm Veteran Member Join Date: Jul 2011 Location: Eastern Oregon Posts: 856 Original Poster	Originally posted by biz-engineer Not sure about this. When I receive the K1 , during the first couple of weeks I tested it, I saw this hesitating AF randomly occurring, especially in low light. Now I'm not experiencing it anymore, and I don't know why. That's interesting and weird. Who knows. It's going to be hard to pin down. Time will tell I guess.

10-17-2016, 10:13 PM	#42
Digitalis Site Supporter Join Date: Mar 2009 Location: Melbourne, Victoria Photos: Gallery Posts: 11,694	Interesting data you have uncovered there Rawr, is there any indication of there being a UV/IR blocking filter covering the metering sensor? the Red channel data seems to be abnormally saturated. Though this could also be contributed to a mismatch between the primary sensor CFA and the metering sensor CFA.

10-18-2016, 01:11 AM	#44
Digitalis Site Supporter Join Date: Mar 2009 Location: Melbourne, Victoria Photos: Gallery Posts: 11,694	Originally posted by rawr it is probably useful for the metering system to be aware [to a degree] of IR wavelengths. The reason why I brought up the issue of IR sensitivity in the metering system is due to the fact that IR focuses differently than visible light- it also predisposes the metering system to inaccuracies as CMOS sensors are particularly sensitive to it. And according to Pentax: on the K1, the AE system has has input on the AF system there may be some fine tuning to cancel out any discrepancy in spectral sensitivity between the sensors being used.

10-18-2016, 03:55 AM	#45
rawr Veteran Member Join Date: Jun 2009 Posts: 11,913	Canon mentions the IR capabilities of the metering sensor in many of it's newer cameras (like the 5DIV, 7DII and even eg the 80D) in terms of the contribution IR can generally make to more accurate exposure, perhaps including light source flicker detection, and maybe skin tone metering too. Skin tone is important obviously for general shooting and portraiture, but is extra important for face detection to work reliably via a limited resolution sensor like a RGB sensor. Since some of those Canon's can do face detection AF via optical viewfinder shooting (like some Nikon's can), perhaps the problems of working with IR vs visible light can be compensated for, and balanced out by the ability of IR to help the camera meter find face targets.??