Pixel shift calibration

biz-engineer · 06-05-2019, 10:42 AM

My original K1 did pixel shift just fine, but somehow it develops some line patterns over the full area of the sensor, especially for exposures longer than a couple of seconds. I'm not sure what does that. When I got my K1 II pixel shift doesn't yield line patterns on long exposures like the K1 does. I once ran a sensor check and I could listen to noise sounding like small displacement of the sensor. So I was wondering if the sensor checks perform some pixel shift calibration or not, and haven't found anything relevant in the K1 manual. Any idea how to figure out if sensor check perform pixel shift calibration or not?

**ProfessorBuzz** · 06-05-2019, 10:55 AM

Can you post a sample image with the problem?

**jbinpg** · 06-05-2019, 12:50 PM - **1 Like**

Digital sensors are prone to interference from EMF. Does the line pattern occur all the time or only when near potential EMF sources such as overhead powerlines/ operating microwaves, etc.?

photoptimist · 06-05-2019, 01:16 PM

If the line patterns are in the brighter areas, then perhaps the lighting is changing during the multiple long exposures -- sunsets are terrible in this regard and many artificial light sources change brightness as the warm up. (You might examine the histograms of the four DNGs inside the big PS RAW file to see if anything is shifted.)

If the line patterns are in the darkest areas, perhaps the sensor is warming up (or cooling off) during the exposures. (You might try turning on long-shutter-time NR to see if that corrects the problem.)

stevebrot · 06-05-2019, 01:54 PM

Originally posted by biz-engineer

Any idea how to figure out if sensor check perform pixel shift calibration or not?

What is "sensor check"? If you mean bad pixel mapping, I can add that my K-3 (no pixel shift) does a short burst of SR system noise when doing the pixel mapping.

Steve

biz-engineer · 06-05-2019, 10:02 PM

Originally posted by stevebrot

What is "sensor check"? If you mean bad pixel mapping, I can add that my K-3 (no pixel shift) does a short burst of SR system noise when doing the pixel mapping.

Yes, when I wrote "sensor check", I meant pixel mapping. So if your K3 does some SR motion while pixel mapping, I'd invalidate the hypothesis of pixel shift calibration on my K1 during pixel mapping operation.

---------- Post added 06-06-19 at 07:05 ----------

Originally posted by ProfessorBuzz

Can you post a sample image with the problem?

I couldn't find the JPEG from pixel shift, I guess I've deleted the pixel shift version and exported to JPEG of a single frame / non-pixel shift exposure.

---------- Post added 06-06-19 at 07:07 ----------

Originally posted by jbinpg

Digital sensors are prone to interference from EMF. Does the line pattern occur all the time or only when near potential EMF sources such as overhead powerlines/ operating microwaves, etc.?

I would exclude EMI at the location, but we never know... there are so much electromagnetic waves going on these days... 5G will come it will be worse. I'd trust the EMI shielding of the Pentax camera for now.

---------- Post added 06-06-19 at 07:17 ----------

Originally posted by photoptimist

If the line patterns are in the brighter areas, then perhaps the lighting is changing during the multiple long exposures

I think you nailed it. According to the JPEG above, it was a 25 sec exposure at f11 after sunset (blue hour) (Date 2016:08:01, Time 21:16:29). And I remember that the line patterns were on the top part of the castle courtyard and the blue sky, the part of the frame that was mostly lit by natural light and not the light bulbs. I was surprise to get line patterns from the pixel shifted exposure because the tripod was sitting on the big stones of the castle floor and I used the IR remote to trigger the exposure. But, the exposure time was 25 sec. per frame, there was 75 sec. time difference between the 1st PS frame and the 4th PS frame, enough for the natural light to have slightly changed.

stevebrot · 06-05-2019, 10:39 PM

Originally posted by biz-engineer

So if your K3 does some SR motion while pixel mapping

It sounds much the same as SR during live view.

Steve

**Rondec** · 06-06-2019, 03:03 AM

Originally posted by biz-engineer

Yes, when I wrote "sensor check", I meant pixel mapping. So if your K3 does some SR motion while pixel mapping, I'd invalidate the hypothesis of pixel shift calibration on my K1 during pixel mapping operation.

---------- Post added 06-06-19 at 07:05 ----------

I couldn't find the JPEG from pixel shift, I guess I've deleted the pixel shift version and exported to JPEG of a single frame / non-pixel shift exposure.

---------- Post added 06-06-19 at 07:07 ----------

I would exclude EMI at the location, but we never know... there are so much electromagnetic waves going on these days... 5G will come it will be worse. I'd trust the EMI shielding of the Pentax camera for now.

---------- Post added 06-06-19 at 07:17 ----------

I think you nailed it. According to the JPEG above, it was a 25 sec exposure at f11 after sunset (blue hour) (Date 2016:08:01, Time 21:16:29). And I remember that the line patterns were on the top part of the castle courtyard and the blue sky, the part of the frame that was mostly lit by natural light and not the light bulbs. I was surprise to get line patterns from the pixel shifted exposure because the tripod was sitting on the big stones of the castle floor and I used the IR remote to trigger the exposure. But, the exposure time was 25 sec. per frame, there was 75 sec. time difference between the 1st PS frame and the 4th PS frame, enough for the natural light to have slightly changed.

Do you use Raw Therapee when developing your pixel shift images? Because there is a "equalize brightness of frames" button there in the pixel shift options that can help with some of those sorts of situations.

MJKoski · 06-06-2019, 03:55 AM

Only way I have managed to get line or grid-patterns with PS is a situation where light changes during the exposure. Motion effects are different beast and very clear to spot (RawTherapee handles them fine). So far, my own results have been similar with K-1, K-1mk2 and KP.

Long exposure cancels out some light changes. For example 30 second PS exposure in ~stable candlelight results in mostly good outcome. But too short exposure in the same situation, like 5-10 seconds, gives 4 subframes with varied light levels -> grid pattern. Flickering flame is always a mess.

biz-engineer · 06-06-2019, 05:32 AM

Originally posted by Rondec

Do you use Raw Therapee when developing your pixel shift images? Because there is a "equalize brightness of frames" button there in the pixel shift options that can help with some of those sorts of situations.

I've used it a bit, but wasn't happy with colors. Too bad, RT has the best pixel shift processing functionality.

06-05-2019, 10:42 AM	#1
biz-engineer Pentaxian Join Date: Feb 2015 Photos: Gallery Posts: 12,247	Pixel shift calibration My original K1 did pixel shift just fine, but somehow it develops some line patterns over the full area of the sensor, especially for exposures longer than a couple of seconds. I'm not sure what does that. When I got my K1 II pixel shift doesn't yield line patterns on long exposures like the K1 does. I once ran a sensor check and I could listen to noise sounding like small displacement of the sensor. So I was wondering if the sensor checks perform some pixel shift calibration or not, and haven't found anything relevant in the K1 manual. Any idea how to figure out if sensor check perform pixel shift calibration or not?

06-05-2019, 01:54 PM	#5
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by biz-engineer Any idea how to figure out if sensor check perform pixel shift calibration or not? What is "sensor check"? If you mean bad pixel mapping, I can add that my K-3 (no pixel shift) does a short burst of SR system noise when doing the pixel mapping. Steve Last edited by stevebrot; 06-05-2019 at 03:16 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K1/Samyang 24mm Tilt/Shift and Pixel Shift	DDoram	Pentax K-1 & K-1 II	6	12-14-2021 06:47 PM
Pixel shift with pixel artefacts	Dericali	Pentax K-1 & K-1 II	4	07-18-2018 12:57 PM
Lloyd Chambers Prefers Pentax Pixel Shift To Sony's Pixel Shift	Fenwoodian	Pentax K-1 & K-1 II	2	07-16-2018 02:15 PM
Pentax KP Pixel Shift	bobbotron	Pentax KP	3	12-15-2017 08:57 AM
Pixel shift appears in the new Sony A7riii	travelswsage	Canon, Nikon, Sony, and Other Camera Brands	14	10-28-2017 07:18 AM

06-05-2019, 10:55 AM	#2
ProfessorBuzz Loyal Site Supporter Join Date: Feb 2018 Posts: 1,903	Can you post a sample image with the problem?

06-05-2019, 12:50 PM - 1 Like	#3
jbinpg Site Supporter Join Date: Mar 2008 Location: Prince George, BC Photos: Gallery \| Albums Posts: 3,546	Digital sensors are prone to interference from EMF. Does the line pattern occur all the time or only when near potential EMF sources such as overhead powerlines/ operating microwaves, etc.?

06-05-2019, 01:16 PM	#4
photoptimist Pentaxian Join Date: Jul 2016 Photos: Albums Posts: 5,128	If the line patterns are in the brighter areas, then perhaps the lighting is changing during the multiple long exposures -- sunsets are terrible in this regard and many artificial light sources change brightness as the warm up. (You might examine the histograms of the four DNGs inside the big PS RAW file to see if anything is shifted.) If the line patterns are in the darkest areas, perhaps the sensor is warming up (or cooling off) during the exposures. (You might try turning on long-shutter-time NR to see if that corrects the problem.)

06-05-2019, 10:02 PM	#6
biz-engineer Pentaxian Join Date: Feb 2015 Photos: Gallery Posts: 12,247 Original Poster	Originally posted by stevebrot What is "sensor check"? If you mean bad pixel mapping, I can add that my K-3 (no pixel shift) does a short burst of SR system noise when doing the pixel mapping. Yes, when I wrote "sensor check", I meant pixel mapping. So if your K3 does some SR motion while pixel mapping, I'd invalidate the hypothesis of pixel shift calibration on my K1 during pixel mapping operation. ---------- Post added 06-06-19 at 07:05 ---------- Originally posted by ProfessorBuzz Can you post a sample image with the problem? I couldn't find the JPEG from pixel shift, I guess I've deleted the pixel shift version and exported to JPEG of a single frame / non-pixel shift exposure. ---------- Post added 06-06-19 at 07:07 ---------- Originally posted by jbinpg Digital sensors are prone to interference from EMF. Does the line pattern occur all the time or only when near potential EMF sources such as overhead powerlines/ operating microwaves, etc.? I would exclude EMI at the location, but we never know... there are so much electromagnetic waves going on these days... 5G will come it will be worse. I'd trust the EMI shielding of the Pentax camera for now. ---------- Post added 06-06-19 at 07:17 ---------- Originally posted by photoptimist If the line patterns are in the brighter areas, then perhaps the lighting is changing during the multiple long exposures I think you nailed it. According to the JPEG above, it was a 25 sec exposure at f11 after sunset (blue hour) (Date 2016:08:01, Time 21:16:29). And I remember that the line patterns were on the top part of the castle courtyard and the blue sky, the part of the frame that was mostly lit by natural light and not the light bulbs. I was surprise to get line patterns from the pixel shifted exposure because the tripod was sitting on the big stones of the castle floor and I used the IR remote to trigger the exposure. But, the exposure time was 25 sec. per frame, there was 75 sec. time difference between the 1st PS frame and the 4th PS frame, enough for the natural light to have slightly changed.

06-05-2019, 10:39 PM	#7
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by biz-engineer So if your K3 does some SR motion while pixel mapping It sounds much the same as SR during live view. Steve

06-06-2019, 03:03 AM	#8
Rondec Loyal Site Supporter Join Date: Mar 2009 Location: Gladys, Virginia Photos: Gallery Posts: 27,663	Originally posted by biz-engineer Yes, when I wrote "sensor check", I meant pixel mapping. So if your K3 does some SR motion while pixel mapping, I'd invalidate the hypothesis of pixel shift calibration on my K1 during pixel mapping operation. ---------- Post added 06-06-19 at 07:05 ---------- I couldn't find the JPEG from pixel shift, I guess I've deleted the pixel shift version and exported to JPEG of a single frame / non-pixel shift exposure. ---------- Post added 06-06-19 at 07:07 ---------- I would exclude EMI at the location, but we never know... there are so much electromagnetic waves going on these days... 5G will come it will be worse. I'd trust the EMI shielding of the Pentax camera for now. ---------- Post added 06-06-19 at 07:17 ---------- I think you nailed it. According to the JPEG above, it was a 25 sec exposure at f11 after sunset (blue hour) (Date 2016:08:01, Time 21:16:29). And I remember that the line patterns were on the top part of the castle courtyard and the blue sky, the part of the frame that was mostly lit by natural light and not the light bulbs. I was surprise to get line patterns from the pixel shifted exposure because the tripod was sitting on the big stones of the castle floor and I used the IR remote to trigger the exposure. But, the exposure time was 25 sec. per frame, there was 75 sec. time difference between the 1st PS frame and the 4th PS frame, enough for the natural light to have slightly changed. Do you use Raw Therapee when developing your pixel shift images? Because there is a "equalize brightness of frames" button there in the pixel shift options that can help with some of those sorts of situations.

06-06-2019, 03:55 AM	#9
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	Only way I have managed to get line or grid-patterns with PS is a situation where light changes during the exposure. Motion effects are different beast and very clear to spot (RawTherapee handles them fine). So far, my own results have been similar with K-1, K-1mk2 and KP. Long exposure cancels out some light changes. For example 30 second PS exposure in ~stable candlelight results in mostly good outcome. But too short exposure in the same situation, like 5-10 seconds, gives 4 subframes with varied light levels -> grid pattern. Flickering flame is always a mess.

06-06-2019, 05:32 AM	#10
biz-engineer Pentaxian Join Date: Feb 2015 Photos: Gallery Posts: 12,247 Original Poster	Originally posted by Rondec Do you use Raw Therapee when developing your pixel shift images? Because there is a "equalize brightness of frames" button there in the pixel shift options that can help with some of those sorts of situations. I've used it a bit, but wasn't happy with colors. Too bad, RT has the best pixel shift processing functionality.