K-1 white dot issue. ?

MJKoski · 11-13-2016, 07:03 AM

I do not see what is wrong with the gray card test. If I use exact same ISO, exposure time, lens (manual aperture leitaxed zeiss) and lights with both bodies it gets pretty simple. If there is difference it is very easy to measure.

Class A · 11-13-2016, 08:56 AM

Originally posted by MJKoski

If there is difference it is very easy to measure.

In this case, how do you know which of the two bodies is not "normal"?

How do you want to rule out that the other body (not having been serviced) may not exhibit reference behaviour with respect to noise (even when it does not show "white dots")?

How do you want rule out that the other body may be OK, but yours is performing extra well, due to the adjustment and calibration it has received? If you simply observe that yours delivers slighter darker images, that could just be due a lower effective ISO performance (which would be a good thing, provided it comes with a respective increase in dynamic range).

MJKoski · 11-13-2016, 09:52 AM

If there is any meaningful difference between these two bodies, it tells that production quality is all over the place or calibration did something extra. No way to know without trying it out. Gimme a switch back to uncalibrated mode please.

yurko_yr · 11-14-2016, 03:59 AM

I think it is matter of statistic. If 10 cameras are behaving one way and one is different, then I would assume that the faulty is the later one, rather than 10 cameras.
It is not questionable that the issue exist, at least for me.
From other hand, only Pentax/Ricoh can solve this situation by repairing the affected cameras. The only way that forum members can assis is to check thier K-1 and report here and/or to pentax if their camera has white dots/dark stripe problem or not. Only publicity of the issue might make Pentax to take an actions on it.
Regards,
Yuriy

rcolman · 11-17-2016, 09:47 AM

Ricoh just announced the availablility of the next K-1 firmware upgrade, aimed at "astrophotography" features.

I am very concerned that the 1.4 firmware "hot pixel removal" will bring a nasty problem from Nikon to Sony and now to Pentax. One of the reasons I left the Sony A7S for night and astrophotography was the "STAR EATER" problem - namely, Sony installed a firmware noise reduction feature, activated with any Bulb exceeding 30 seconds, meant to remove hot pixels or "white dots" from an image.

**& UNFORTUNATELY, it would also remove faint stars and the centers of large bright stars as well. The "feature" could not be turned off or modified by the user. **

Nikon also had this "feature" but had enough smarts to remove it after complaints. As far as I know, Sony still has not fixed this problem, and I left Sony for Pentax.

The k-1 Astrotracer runs only in BULB mode and a Pentax version of the "Star Eater" problem would be even worse !!

So, I am actually fearful of installing the 1.4 firmware until I learn more.

Does anyone have any additional details?

Read more at: https://www.pentaxforums.com/forums/190-pentax-k-1/333557-firmware-update-1-4...#ixzz4QHYdHpUc

JBox · 11-17-2016, 10:28 AM

Originally posted by rcolman

I am very concerned that the 1.4 firmware "hot pixel removal" will bring a nasty problem from Nikon to Sony and now to Pentax. One of the reasons I left the Sony A7S for night and astrophotography was the "STAR EATER" problem - namely, Sony installed a firmware noise reduction feature, activated with any Bulb exceeding 30 seconds, meant to remove hot pixels or "white dots" from an image.

I just looked at the release notes and it seems to be saying that this new hot pixel removal is applied only to JPEG. Am I missing something?

awscreo · 11-17-2016, 12:36 PM

Same, I think it says it only affects jpegs. Plus it can be turned off if I'm not mistaken.

MJKoski · 11-17-2016, 01:54 PM

Star eater would be fine for most long exposure work. I would use it for anything else than star trails / milky way stuff. And then, if it is an option, it could be switched off should one prefer pure RAW.

--

I received ISO25600, 30 seconds darkframe (roomtemp ~22C) today from another K-1 owner (never serviced) and his results were much less noisy just by looking at preview image of the file. Then, I subtracted his unit's output from my calibrated unit's output and got this:

These devices are all over the place it seems considering consistent output between different copies. And like my loaner unit before, the noise profile of this other K-1 is much smoother and random like.

Class A · 11-17-2016, 08:32 PM

Have you been sent a RAW file?

Are you still using LR (or PS) to render RAW files?
If so, you might as well save yourself the trouble of substracting images from each other (due to the adaptive interpretation by ACR).

Also, what do you expect from substracting noise from noise? The result should be noise again (with a mean value that is the difference of the original mean values).

You cannot expect two uncorrelated noise sources to somehow cancel themselves out. Dark frame substraction only works because thermal noise from one and the same sensor is not entirely random.

I don't see how you have made any progress.

stevebrot · 11-17-2016, 08:52 PM

Originally posted by Class A

Are you still using LR (or PS) to render RAW files?
If so, you might as well save yourself the trouble of substracting images from each other (due to the adaptive interpretation by ACR).

Ideally, it would be better to use dcraw to extract a linear conversion. Here is an interesting page describing the difference between ACR and a linear conversion:

mikeboers.com

When done with an image of an exposure wedge, gray card, or whatever, it is possible to do a quantitative evaluation of the actual sensor response. Here is another resource where they did just that:

Capturing Linear Images

From what I understand, DPReview and DXO Mark use similar numeric analysis.

Steve

MJKoski · 11-18-2016, 02:40 AM

Just by looking at RAW-Digger preview of four different outputs they all have difference in the amount of random noise @ room temp (~22C)

Order of amount of noise produced (visual examination):

1. My original unit, uncalibrated (black bar, lot of horizontal banding)
2. Loaner unit (no service done, smooth noise profile)
3. My calibrated unit (2nd calibration, horizontal banding remains)
4. Another K-1 (no service done, smooth noise profile)

I also tried subtracting ISO25600 output from another ISO25600 sample from my unit and this result in almost black image with very low noise. While subtracting those above samples from each other there is a large difference between the samples.

So, what is happening? Result from B-stock pile of sensors slapped into bodies for selling them ultra cheap?

Class A · 11-18-2016, 06:29 AM

Originally posted by MJKoski

So, what is happening?

You are unfamiliar with basics of sensor characteristics and ignore the advice of people who are trying to be helpful (stevebrot, to name just one). That's what is happening.

There might also be something happening with your sensor, but it seems we'll never know.

MJKoski · 11-18-2016, 07:47 AM

My previous D800E bodies (3 in total) and D810 (2) have given very very similar noise profile with high ISO and long exposures. My a7r1 bodies (2) were consistent as well. Not so with K-1. Now I have samples from three different K-1 bodies and output from each of them are somewhat different. In a blind test two (uncalibrated and calibrated results of my K-1) of the four samples would look like they are from different camera with different sensor. My unit's output is coarse and has slight banding all over versus the two other cases which produce smooth fine grained random noise. Something is not correct here.

ZeGaby · 11-18-2016, 08:49 AM

I've send some raws with my White Dots.

https://www.pentaxforums.com/forums/190-pentax-k-1/327326-k-1-white-dot-issue-6.html#post3754791

My retailer sends theses pcitures to Pentax , and they said that it's a normal behaviour of the K1.

aikaarska · 11-18-2016, 10:11 AM

Originally posted by ZeGaby

I've send some raws with my White Dots.

https://www.pentaxforums.com/forums/190-pentax-k-1/327326-k-1-white-dot-issue-6.html#post3754791

My retailer sends theses pcitures to Pentax , and they said that it's a normal behaviour of the K1.

Well, I think that's a normal behaviour of the Sony 36mp FF sensor (as Nikon D810 had similar white dot noise)...

But it is possible that after Pentax service (after sensor calibration), white dot noise decreases dramatically. So maybe Pentax know the solution to this issue after all.

11-13-2016, 08:56 AM	#362
Class A Pentaxian Join Date: Aug 2008 Location: Wellington, New Zealand Posts: 11,251	Originally posted by MJKoski If there is difference it is very easy to measure. In this case, how do you know which of the two bodies is not "normal"? How do you want to rule out that the other body (not having been serviced) may not exhibit reference behaviour with respect to noise (even when it does not show "white dots")? How do you want rule out that the other body may be OK, but yours is performing extra well, due to the adjustment and calibration it has received? If you simply observe that yours delivers slighter darker images, that could just be due a lower effective ISO performance (which would be a good thing, provided it comes with a respective increase in dynamic range). Last edited by Class A; 11-13-2016 at 11:05 AM.

11-17-2016, 01:54 PM	#368
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	Star eater would be fine for most long exposure work. I would use it for anything else than star trails / milky way stuff. And then, if it is an option, it could be switched off should one prefer pure RAW. -- I received ISO25600, 30 seconds darkframe (roomtemp ~22C) today from another K-1 owner (never serviced) and his results were much less noisy just by looking at preview image of the file. Then, I subtracted his unit's output from my calibrated unit's output and got this: These devices are all over the place it seems considering consistent output between different copies. And like my loaner unit before, the noise profile of this other K-1 is much smoother and random like. Last edited by MarkJerling; 11-17-2016 at 08:53 PM. Reason: Masked profanity removed.

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

11-13-2016, 07:03 AM	#361
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	I do not see what is wrong with the gray card test. If I use exact same ISO, exposure time, lens (manual aperture leitaxed zeiss) and lights with both bodies it gets pretty simple. If there is difference it is very easy to measure.

11-13-2016, 09:52 AM	#363
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	If there is any meaningful difference between these two bodies, it tells that production quality is all over the place or calibration did something extra. No way to know without trying it out. Gimme a switch back to uncalibrated mode please.

11-14-2016, 03:59 AM	#364
yurko_yr Senior Member Join Date: Jun 2010 Location: Saarbrucken Photos: Gallery Posts: 332	I think it is matter of statistic. If 10 cameras are behaving one way and one is different, then I would assume that the faulty is the later one, rather than 10 cameras. It is not questionable that the issue exist, at least for me. From other hand, only Pentax/Ricoh can solve this situation by repairing the affected cameras. The only way that forum members can assis is to check thier K-1 and report here and/or to pentax if their camera has white dots/dark stripe problem or not. Only publicity of the issue might make Pentax to take an actions on it. Regards, Yuriy

11-17-2016, 09:47 AM	#365
rcolman Senior Member Join Date: Jul 2016 Photos: Albums Posts: 118	Ricoh just announced the availablility of the next K-1 firmware upgrade, aimed at "astrophotography" features. I am very concerned that the 1.4 firmware "hot pixel removal" will bring a nasty problem from Nikon to Sony and now to Pentax. One of the reasons I left the Sony A7S for night and astrophotography was the "STAR EATER" problem - namely, Sony installed a firmware noise reduction feature, activated with any Bulb exceeding 30 seconds, meant to remove hot pixels or "white dots" from an image. & UNFORTUNATELY, it would also remove faint stars and the centers of large bright stars as well. The "feature" could not be turned off or modified by the user. Nikon also had this "feature" but had enough smarts to remove it after complaints. As far as I know, Sony still has not fixed this problem, and I left Sony for Pentax. The k-1 Astrotracer runs only in BULB mode and a Pentax version of the "Star Eater" problem would be even worse !! So, I am actually fearful of installing the 1.4 firmware until I learn more. Does anyone have any additional details? Read more at: https://www.pentaxforums.com/forums/190-pentax-k-1/333557-firmware-update-1-4...#ixzz4QHYdHpUc

11-17-2016, 10:28 AM	#366
JBox Senior Member Join Date: Aug 2011 Location: Michigan Posts: 109	Originally posted by rcolman I am very concerned that the 1.4 firmware "hot pixel removal" will bring a nasty problem from Nikon to Sony and now to Pentax. One of the reasons I left the Sony A7S for night and astrophotography was the "STAR EATER" problem - namely, Sony installed a firmware noise reduction feature, activated with any Bulb exceeding 30 seconds, meant to remove hot pixels or "white dots" from an image. I just looked at the release notes and it seems to be saying that this new hot pixel removal is applied only to JPEG. Am I missing something?

11-17-2016, 12:36 PM	#367
awscreo Veteran Member Join Date: May 2016 Photos: Gallery \| Albums Posts: 2,722	Same, I think it says it only affects jpegs. Plus it can be turned off if I'm not mistaken.

11-17-2016, 08:32 PM	#369
Class A Pentaxian Join Date: Aug 2008 Location: Wellington, New Zealand Posts: 11,251	Have you been sent a RAW file? Are you still using LR (or PS) to render RAW files? If so, you might as well save yourself the trouble of substracting images from each other (due to the adaptive interpretation by ACR). Also, what do you expect from substracting noise from noise? The result should be noise again (with a mean value that is the difference of the original mean values). You cannot expect two uncorrelated noise sources to somehow cancel themselves out. Dark frame substraction only works because thermal noise from one and the same sensor is not entirely random. I don't see how you have made any progress.

11-17-2016, 08:52 PM	#370
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by Class A Are you still using LR (or PS) to render RAW files? If so, you might as well save yourself the trouble of substracting images from each other (due to the adaptive interpretation by ACR). Ideally, it would be better to use dcraw to extract a linear conversion. Here is an interesting page describing the difference between ACR and a linear conversion: mikeboers.com When done with an image of an exposure wedge, gray card, or whatever, it is possible to do a quantitative evaluation of the actual sensor response. Here is another resource where they did just that: Capturing Linear Images From what I understand, DPReview and DXO Mark use similar numeric analysis. Steve

11-18-2016, 02:40 AM	#371
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	Just by looking at RAW-Digger preview of four different outputs they all have difference in the amount of random noise @ room temp (~22C) Order of amount of noise produced (visual examination): 1. My original unit, uncalibrated (black bar, lot of horizontal banding) 2. Loaner unit (no service done, smooth noise profile) 3. My calibrated unit (2nd calibration, horizontal banding remains) 4. Another K-1 (no service done, smooth noise profile) I also tried subtracting ISO25600 output from another ISO25600 sample from my unit and this result in almost black image with very low noise. While subtracting those above samples from each other there is a large difference between the samples. So, what is happening? Result from B-stock pile of sensors slapped into bodies for selling them ultra cheap?

11-18-2016, 06:29 AM	#372
Class A Pentaxian Join Date: Aug 2008 Location: Wellington, New Zealand Posts: 11,251	Originally posted by MJKoski So, what is happening? You are unfamiliar with basics of sensor characteristics and ignore the advice of people who are trying to be helpful (stevebrot, to name just one). That's what is happening. There might also be something happening with your sensor, but it seems we'll never know.

11-18-2016, 07:47 AM	#373
MJKoski Veteran Member Join Date: Aug 2016 Posts: 1,784	My previous D800E bodies (3 in total) and D810 (2) have given very very similar noise profile with high ISO and long exposures. My a7r1 bodies (2) were consistent as well. Not so with K-1. Now I have samples from three different K-1 bodies and output from each of them are somewhat different. In a blind test two (uncalibrated and calibrated results of my K-1) of the four samples would look like they are from different camera with different sensor. My unit's output is coarse and has slight banding all over versus the two other cases which produce smooth fine grained random noise. Something is not correct here.

11-18-2016, 08:49 AM	#374
ZeGaby Pentaxian Join Date: Jul 2014 Photos: Gallery Posts: 308	I've send some raws with my White Dots. https://www.pentaxforums.com/forums/190-pentax-k-1/327326-k-1-white-dot-issue-6.html#post3754791 My retailer sends theses pcitures to Pentax , and they said that it's a normal behaviour of the K1.

11-18-2016, 10:11 AM	#375
aikaarska Junior Member Join Date: Sep 2016 Location: Finland Posts: 44	Originally posted by ZeGaby I've send some raws with my White Dots. https://www.pentaxforums.com/forums/190-pentax-k-1/327326-k-1-white-dot-issue-6.html#post3754791 My retailer sends theses pcitures to Pentax , and they said that it's a normal behaviour of the K1. Well, I think that's a normal behaviour of the Sony 36mp FF sensor (as Nikon D810 had similar white dot noise)... But it is possible that after Pentax service (after sensor calibration), white dot noise decreases dramatically. So maybe Pentax know the solution to this issue after all.