K-1 white dot issue. ?

photoptimist · 08-25-2016, 10:34 AM

Originally posted by Gimbal

I just noticed the weirdest thing while looking at your 100% crop, there is a lot of hot pixels of red, green and blue type but not white. That looks kind of normal (except perhaps the huge amount of them), but "my" long exposures do not have any green hot pixels, they are blue, red and white.

So much depends on the raw converter used that I don't now what to think anymore. The red and blue ones are usually larger than one pixel, forming a little cross or line of 2-3 pixels, while the white dots are single pixels standing alone. (How is that even possible with a bayer pattern?).

Is there a chance you could upload the 120s exposure so that we (or I) could see how it looks in the raw converter I have? This to see if there is an actual difference in the files or if it is the raw converter that makes them look different.

It's probably caused by differences in different raw convertors and the interaction of two issues:

1) RGB channel gain: how much each of the 3 channels are boosted to compensate for Bayer color filter attenuation, sensor quantum efficiency, and the convertor's white balance setting. Generally, the red and blue channels get the biggest boost.

2) RGB interpolation: the red and blue channels are sparsely sampled so the convertor needs to synthesize the missing pixels which might convert a 1x1 hot pixel into a cross or blob depending on the surrounding RGB values and the image structure models the convertor uses for interpolation. The green channel is more densely sampled so the reconstructed green channel has smaller artifacts.

(it would be interesting the analyze the coordinates of the various types of hot pixels -- are they always located on even (or odd) rows or columns??)

MJKoski · 08-25-2016, 11:56 AM

I asked about white dots on a finnish forum, three k-1 owners replied that they have very very few dots in iso3200 30 sec exposure. I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. No, I certainly was not trolling when I said I switched from A7R2 to K-1 to get better LE performance. All for nothing. Maybe a working unit is fine with LE's. Mine certainly is not. Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long.

Gimbal · 08-25-2016, 12:58 PM

Originally posted by MJKoski

I asked about white dots on a finnish forum, three k-1 owners replied that they have very very few dots in iso3200 30 sec exposure. I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. No, I certainly was not trolling when I said I switched from A7R2 to K-1 to get better LE performance. All for nothing. Maybe a working unit is fine with LE's. Mine certainly is not. Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long.

How is the result if you remove the white dots using Rawtherapee hot pixel removal function? As a work around in the mean time.

It seems to work pretty well on the few tests I've done so far, but then again I have only done simple test shots. Maybe it's not that good on a really critical shot, I don't know.

awscreo · 08-25-2016, 01:05 PM

Originally posted by MJKoski

My K-1 dealer should have one new body which arrived last week with my 15-30 zoom as I went to pick it up. If time allows I try to visit the shop on my way home tomorrow.

I'd much appreciate it if you could post a sample LE shot, even with a cap on. Just to see if issue persists on a newly manufactured bodies.

stevebrot · 08-25-2016, 01:06 PM

Originally posted by MJKoski

I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up.

Originally posted by MJKoski

Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long.

How long have you owned your K-1? Most reputable dealers will do an exchange of a defective camera within the first month. If you do have to send it in, I hope they find the source of your problem and provide you with a timely repair or replacement. I was unaware that the K-1 was supposed to have lower LE noise than the Sony A7RII, but with any luck Ricoh/Pentax will be able to provide you with a reasonable solution.

Steve

banep · 08-25-2016, 01:11 PM

I tested tonight very long exposures (10-30min, VND filter) with my K-1 against north sky.
Trails everywhere and only one visible and bright white dot - Polaris

---------- Post added 08-25-16 at 10:29 PM ----------

Originally posted by MJKoski

I asked about white dots on a finnish forum, three k-1 owners replied that they have very very few dots in iso3200 30 sec exposure. I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. No, I certainly was not trolling when I said I switched from A7R2 to K-1 to get better LE performance. All for nothing. Maybe a working unit is fine with LE's. Mine certainly is not. Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long.

If you can't wait for a solution, try to replace camera at your dealer.
If they don't have any servicing instruction for that particular situation they will probably contact Ricoh, but if Ricoh don't have any instructons there is a high possibility that they will agree to replace your camera instead of servicing.

MJKoski · 08-25-2016, 01:32 PM

A7R2 is not good for long exposures. It gets to about 30 seconds without messing everything with confetti style color noise. A7R1 however gets easily to 4-5 minutes without dark frame subtraction. But that aside. I did a rather extreme test.

Here, ISO3200, 121 seconds, room temp @ +23C
http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1151.jpg/_full.jpg

I put the camera in freezer for 1.5 hours. -18C reading there. ISO3200, 122 seconds (sorry for this huge error)
http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1152.jpg/_full.jpg

See for yourself. I have not seen such radical change in noise with any CMOS camera I have owned. You can also see how the viewfinder blocker (little plastic piece) leaks light from one side but that is not the point.

mandale · 08-25-2016, 05:28 PM

I recognize you might be under time constraints, but the best way to judge this is to have the same test, done at the same place, under the same conditions with the Canon to see whether this is a flaw specific to Sony sensors. Seeing is believing.

awscreo · 08-25-2016, 11:27 PM

I guess there's just no way to know

Sucks, not the best feeling to be unsure in something you're spending so much money on.

Gimbal · 08-26-2016, 09:28 AM

Originally posted by MJKoski

A7R2 is not good for long exposures. It gets to about 30 seconds without messing everything with confetti style color noise. A7R1 however gets easily to 4-5 minutes without dark frame subtraction. But that aside. I did a rather extreme test.

Here, ISO3200, 121 seconds, room temp @ +23C
http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1151.jpg/_full.jpg

I put the camera in freezer for 1.5 hours. -18C reading there. ISO3200, 122 seconds (sorry for this huge error)
http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1152.jpg/_full.jpg

See for yourself. I have not seen such radical change in noise with any CMOS camera I have owned. You can also see how the viewfinder blocker (little plastic piece) leaks light from one side but that is not the point.

I have an older version of lightroom so maybe it isn't relevant, but that version removes a lot of hot pixels whether you like it to or not. I guess that is why these two examples look as clean as they do, also by default the croma noise filter turns the single green hot pixels into white pixels and doesn't remove them all.

Anyhow, the cold picture looks really good.

MJKoski · 08-26-2016, 09:39 AM

Yea, cold picture looks good to me too. A7R2 works the same way, it does very well with multiminute exposures without dark frame subtraction in -30C temp. Does LR really reduce these white dots? Luminance noise reduction gets easily rid of all the white ones but there goes all the detail the image might have as well.

When I asked other finnish owners about this few claimed their room temp dark frames with similar settings are like my cold dark frame. Go figure...

Gimbal · 08-26-2016, 12:34 PM

While waiting for a better solution I just "discovered" a pretty powerful noise reduction function in rawtherapee that also works on a jpeg to clean up the white dots without blurring up the picture. Quite impressive in my opinion. It's called impulse noise reduction and has a slider to adjust the strength.

I hope you don't mind MJKoski that I used your skeleton in post #43 as a test target. While that is the worst example I've seen so far it cleaned up pretty well without making a mush of it.

MJKoski · 08-26-2016, 12:58 PM

Heh...Mr. Skellie has experienced far worse abuse during the recent year shooting

Indeed it seems to work rather nicely. But I will not switch from LR to any other SW as of now. Too much hassle to replace current workflow with alternative one.

---------- Post added 08-26-16 at 12:59 PM ----------

Actually it works better than in-camera LENR with those shooting parameters...

MJKoski · 08-29-2016, 09:09 AM

All right, my dealer informed me that my camera should be sent to service. That was the reply from Pentax technicians after seeing the skellie crop. They will replace the whole sensor unit if that is needed.

fromunderthebridge · 08-29-2016, 11:25 PM

Originally posted by MJKoski

All right, my dealer informed me that my camera should be sent to service. That was the reply from Pentax technicians after seeing the skellie crop. They will replace the whole sensor unit if that is needed.

What's your dealer btw? Another Finnish user here. I so far haven't done a lot of long exposure shooting and am wondering if I should perform some control tests on my unit as well to avoid nasty surprises later on.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
White Dot on 50mm f1.4 Super-Tak	roboticspro	Troubleshooting and Beginner Help	7	07-20-2018 01:47 PM
For Sale - Sold: White Pentax K-r + White Kit Lens + 2 Extra Batteries + 1 Extra Charger	MightyMike	Sold Items	2	04-14-2015 07:21 AM
Lost white dot	bassek	Pentax SLR Lens Discussion	15	09-06-2012 10:22 AM
white dot on live view screen	Shingkyo	Pentax K-r	7	03-14-2011 06:54 AM
old lenses- that little plastic white dot?	dandaniel	Pentax SLR Lens Discussion	6	02-25-2009 02:13 PM

08-25-2016, 10:34 AM	#61
photoptimist Pentaxian Join Date: Jul 2016 Photos: Albums Posts: 5,122	Originally posted by Gimbal I just noticed the weirdest thing while looking at your 100% crop, there is a lot of hot pixels of red, green and blue type but not white. That looks kind of normal (except perhaps the huge amount of them), but "my" long exposures do not have any green hot pixels, they are blue, red and white. So much depends on the raw converter used that I don't now what to think anymore. The red and blue ones are usually larger than one pixel, forming a little cross or line of 2-3 pixels, while the white dots are single pixels standing alone. (How is that even possible with a bayer pattern?). Is there a chance you could upload the 120s exposure so that we (or I) could see how it looks in the raw converter I have? This to see if there is an actual difference in the files or if it is the raw converter that makes them look different. It's probably caused by differences in different raw convertors and the interaction of two issues: 1) RGB channel gain: how much each of the 3 channels are boosted to compensate for Bayer color filter attenuation, sensor quantum efficiency, and the convertor's white balance setting. Generally, the red and blue channels get the biggest boost. 2) RGB interpolation: the red and blue channels are sparsely sampled so the convertor needs to synthesize the missing pixels which might convert a 1x1 hot pixel into a cross or blob depending on the surrounding RGB values and the image structure models the convertor uses for interpolation. The green channel is more densely sampled so the reconstructed green channel has smaller artifacts. (it would be interesting the analyze the coordinates of the various types of hot pixels -- are they always located on even (or odd) rows or columns??)

08-25-2016, 11:56 AM	#62
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	I asked about white dots on a finnish forum, three k-1 owners replied that they have very very few dots in iso3200 30 sec exposure. I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. No, I certainly was not trolling when I said I switched from A7R2 to K-1 to get better LE performance. All for nothing. Maybe a working unit is fine with LE's. Mine certainly is not. Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long.

08-25-2016, 12:58 PM	#63
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106 Original Poster	Originally posted by MJKoski I asked about white dots on a finnish forum, three k-1 owners replied that they have very very few dots in iso3200 30 sec exposure. I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. No, I certainly was not trolling when I said I switched from A7R2 to K-1 to get better LE performance. All for nothing. Maybe a working unit is fine with LE's. Mine certainly is not. Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long. How is the result if you remove the white dots using Rawtherapee hot pixel removal function? As a work around in the mean time. It seems to work pretty well on the few tests I've done so far, but then again I have only done simple test shots. Maybe it's not that good on a really critical shot, I don't know.

08-25-2016, 01:05 PM	#64
awscreo Veteran Member Join Date: May 2016 Photos: Gallery \| Albums Posts: 2,722	Originally posted by MJKoski My K-1 dealer should have one new body which arrived last week with my 15-30 zoom as I went to pick it up. If time allows I try to visit the shop on my way home tomorrow. I'd much appreciate it if you could post a sample LE shot, even with a cap on. Just to see if issue persists on a newly manufactured bodies.

08-25-2016, 01:06 PM	#65
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by MJKoski I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. Originally posted by MJKoski Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long. How long have you owned your K-1? Most reputable dealers will do an exchange of a defective camera within the first month. If you do have to send it in, I hope they find the source of your problem and provide you with a timely repair or replacement. I was unaware that the K-1 was supposed to have lower LE noise than the Sony A7RII, but with any luck Ricoh/Pentax will be able to provide you with a reasonable solution. Steve

08-25-2016, 01:11 PM	#66
banep Banned Join Date: Mar 2013 Posts: 360	I tested tonight very long exposures (10-30min, VND filter) with my K-1 against north sky. Trails everywhere and only one visible and bright white dot - Polaris ---------- Post added 08-25-16 at 10:29 PM ---------- Originally posted by MJKoski I asked about white dots on a finnish forum, three k-1 owners replied that they have very very few dots in iso3200 30 sec exposure. I would say my unit is not slightly out of tune but BROKEN. Like totally screwed up. No, I certainly was not trolling when I said I switched from A7R2 to K-1 to get better LE performance. All for nothing. Maybe a working unit is fine with LE's. Mine certainly is not. Not even slightly amused about this as only way to get it fixed is a service trip which takes god knows how long. If you can't wait for a solution, try to replace camera at your dealer. If they don't have any servicing instruction for that particular situation they will probably contact Ricoh, but if Ricoh don't have any instructons there is a high possibility that they will agree to replace your camera instead of servicing.

08-25-2016, 01:32 PM	#67
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	A7R2 is not good for long exposures. It gets to about 30 seconds without messing everything with confetti style color noise. A7R1 however gets easily to 4-5 minutes without dark frame subtraction. But that aside. I did a rather extreme test. Here, ISO3200, 121 seconds, room temp @ +23C http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1151.jpg/_full.jpg I put the camera in freezer for 1.5 hours. -18C reading there. ISO3200, 122 seconds (sorry for this huge error) http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1152.jpg/_full.jpg See for yourself. I have not seen such radical change in noise with any CMOS camera I have owned. You can also see how the viewfinder blocker (little plastic piece) leaks light from one side but that is not the point.

08-25-2016, 05:28 PM	#68
mandale New Member Join Date: Jul 2016 Posts: 3	I recognize you might be under time constraints, but the best way to judge this is to have the same test, done at the same place, under the same conditions with the Canon to see whether this is a flaw specific to Sony sensors. Seeing is believing.

08-25-2016, 11:27 PM	#69
awscreo Veteran Member Join Date: May 2016 Photos: Gallery \| Albums Posts: 2,722	I guess there's just no way to know Sucks, not the best feeling to be unsure in something you're spending so much money on.

08-26-2016, 09:28 AM	#70
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106 Original Poster	Originally posted by MJKoski A7R2 is not good for long exposures. It gets to about 30 seconds without messing everything with confetti style color noise. A7R1 however gets easily to 4-5 minutes without dark frame subtraction. But that aside. I did a rather extreme test. Here, ISO3200, 121 seconds, room temp @ +23C http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1151.jpg/_full.jpg I put the camera in freezer for 1.5 hours. -18C reading there. ISO3200, 122 seconds (sorry for this huge error) http://mjkoski.1g.fi/kuvat/Miscellaneous/trash/_IMG1152.jpg/_full.jpg See for yourself. I have not seen such radical change in noise with any CMOS camera I have owned. You can also see how the viewfinder blocker (little plastic piece) leaks light from one side but that is not the point. I have an older version of lightroom so maybe it isn't relevant, but that version removes a lot of hot pixels whether you like it to or not. I guess that is why these two examples look as clean as they do, also by default the croma noise filter turns the single green hot pixels into white pixels and doesn't remove them all. Anyhow, the cold picture looks really good.

08-26-2016, 09:39 AM	#71
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	Yea, cold picture looks good to me too. A7R2 works the same way, it does very well with multiminute exposures without dark frame subtraction in -30C temp. Does LR really reduce these white dots? Luminance noise reduction gets easily rid of all the white ones but there goes all the detail the image might have as well. When I asked other finnish owners about this few claimed their room temp dark frames with similar settings are like my cold dark frame. Go figure...

08-26-2016, 12:58 PM	#73
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	Heh...Mr. Skellie has experienced far worse abuse during the recent year shooting Indeed it seems to work rather nicely. But I will not switch from LR to any other SW as of now. Too much hassle to replace current workflow with alternative one. ---------- Post added 08-26-16 at 12:59 PM ---------- Actually it works better than in-camera LENR with those shooting parameters...

08-29-2016, 09:09 AM	#74
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	All right, my dealer informed me that my camera should be sent to service. That was the reply from Pentax technicians after seeing the skellie crop. They will replace the whole sensor unit if that is needed.

08-29-2016, 11:25 PM	#75
fromunderthebridge Senior Member Join Date: May 2014 Photos: Gallery Posts: 181	Originally posted by MJKoski All right, my dealer informed me that my camera should be sent to service. That was the reply from Pentax technicians after seeing the skellie crop. They will replace the whole sensor unit if that is needed. What's your dealer btw? Another Finnish user here. I so far haven't done a lot of long exposure shooting and am wondering if I should perform some control tests on my unit as well to avoid nasty surprises later on.