Corrupt Images: Any last suggestions before I send the K1 in for repairs?

h4yn0nnym0u5e · 10-20-2018, 12:52 AM

Originally posted by lazarustx

6. Image 306 is the first image to show up as bad.

NOTE: I did take a look at the prior image and did a compare of the corrupt image. I see that the time stamp is off by milliseconds. That's the first change. However, that's where my reading raw image skills fall apart. Anyone know how to take a look at these files to see's different?

Thanks!

Any chance of posting images 305 and 306 somewhere we could take a look at them? Can't promise to find time, but I am quite intrigued by this, especially as it affects more than one camera model.

**lazarustx** · 10-23-2018, 08:54 AM

@h4yn0nnym0u5e I cleaned up and removed the original set, but yesterday I was able to create a few images to upload here:

Update your browser to use Google Drive - Google Drive Help.
Slot 1 Images:
IMGP7696.DNG - Good image taken right before Bad image
IMGP7697.DNG- Corrupt Image
IMGP7698.DNG - Next good image

IMGP7697.JPG - Good JPG produced in Camera.

Slot 2 Images are Binary equal using BeyondCompare.

Things to note:
1. Camera was set to shoot RAW+.
2. Camera was set to write to both to Slot 1 and Slot 2.

3. A binary compare of images IMGP7697.DNG and IMGP7697.JPG from Slot 1 and Slot 2 show them as identical.

4. If I use DCU to open the corrupt DNG, if I save the image to a TIF, I can see the corruption. However, I don't see the corruption until I do a Save/Save as.

This all suggests a hardware issue from what I I can tell. If you guys think otherwise let me know.
I'll have to figure out a way to have them try to recreate this inhouse @ Precision Camera.

Regards,
Rodney

**frankoz** · 10-23-2018, 07:54 PM

Originally posted by lazarustx

@h4yn0nnym0u5e I cleaned up and removed the original set, but yesterday I was able to create a few images to upload here:

Update your browser to use Google Drive - Google Drive Help.
Slot 1 Images:
IMGP7696.DNG - Good image taken right before Bad image
IMGP7697.DNG- Corrupt Image
IMGP7698.DNG - Next good image

IMGP7697.JPG - Good JPG produced in Camera.

Slot 2 Images are Binary equal using BeyondCompare.

Things to note:
1. Camera was set to shoot RAW+.
2. Camera was set to write to both to Slot 1 and Slot 2.

3. A binary compare of images IMGP7697.DNG and IMGP7697.JPG from Slot 1 and Slot 2 show them as identical.

4. If I use DCU to open the corrupt DNG, if I save the image to a TIF, I can see the corruption. However, I don't see the corruption until I do a Save/Save as.

This all suggests a hardware issue from what I I can tell. If you guys think otherwise let me know.
I'll have to figure out a way to have them try to recreate this inhouse @ Precision Camera.

Regards,
Rodney

Rodney,

Good luck with Precision Camera. Please let me know if Pentax/Ricoh acknowledge a hardware fault; I will hit them up for my Ricoh GRii as well.

Rgds,
John

h4yn0nnym0u5e · 10-24-2018, 10:44 AM

Originally posted by lazarustx

Slot 1 Images:
IMGP7696.DNG - Good image taken right before Bad image
IMGP7697.DNG- Corrupt Image
IMGP7698.DNG - Next good image

IMGP7697.JPG - Good JPG produced in Camera.

3. A binary compare of images IMGP7697.DNG and IMGP7697.JPG from Slot 1 and Slot 2 show them as identical.

Hi Rodney

Can I just check you exported the correct images?! You say IMGP769x.x but what's in the Google Drive are IMGP799x.x. Also, I'd assume the two images that you compared are both DNGs, not a DNG and a JPG!

I'm not getting a problem with these in RawTherapee 5.4, and my DCU ~~won't open any of them, I think I may need to update to a K-1 II-capable version~~ seems fine with IMGP7997.DNG, now I've updated to 5.8.1.

Cheers

Jonathan

Last edited by h4yn0nnym0u5e; 10-24-2018 at 10:52 AM.

**lazarustx** · 10-24-2018, 04:10 PM

Originally posted by h4yn0nnym0u5e

Hi Rodney

Can I just check you exported the correct images?! You say IMGP769x.x but what's in the Google Drive are IMGP799x.x. Also, I'd assume the two images that you compared are both DNGs, not a DNG and a JPG!

I'm not getting a problem with these in RawTherapee 5.4, and my DCU ~~won't open any of them, I think I may need to update to a K-1 II-capable version~~ seems fine with IMGP7997.DNG, now I've updated to 5.8.1.

Cheers

Jonathan

Sorry about that. I've uploaded them correctly this time and removed the old ones. You might need to do a refresh...

- I did a binary compare of the DNG files from Slot 1 and Slot 2 - Meaning IMGP7697.DNG is the same on both Slots.

- The JPG for IMGP7697.DNG was produced in camera.
- The other 2 images were taken right before and right after the corrupt image.

Here's the link again: https://drive.google.com/drive/u/1/folders/1u70GX33WEGaZ1S76N6T1eOsiy2mgE4sl

Thanks!

leekil · 10-24-2018, 10:35 PM

Originally posted by lazarustx

- I did a binary compare of the DNG files from Slot 1 and Slot 2 - Meaning IMGP7697.DNG is the same on both Slots.

This would seem to rule out the cards being a problem. And likely the card writing, though not 100% sure.

prat1k · 10-27-2018, 09:31 AM

I am having the exact same problem with my K-500. It is intermittent, and independent of the lens I use. There is also an ISO issue (higher ISO gives me purple hue) - it started with infrequent corruption in images and now becoming more and more common. Using Tamron 18-250 mm lens and also tried Pentax DA 50 mm prime. No difference. Some images for example...

Attached Images

View Picture EXIF

View Picture EXIF

h4yn0nnym0u5e · 10-27-2018, 12:40 PM - **3 Likes**

Originally posted by lazarustx

Sorry about that. I've uploaded them correctly this time and removed the old ones. You might need to do a refresh...

- I did a binary compare of the DNG files from Slot 1 and Slot 2 - Meaning IMGP7697.DNG is the same on both Slots.

- The JPG for IMGP7697.DNG was produced in camera.
- The other 2 images were taken right before and right after the corrupt image.

Here's the link again: Update your browser to use Google Drive - Google Drive Help

Thanks!

Hi Rodney

Here's what RawTherapee 5.4 made of it without demosaicing, i.e. showing the bare RGGB pixels straight from the sensor. The bright red dots appear to be correct values, but with a couple of extra bits set; the magenta area to the right seems to be a different error, where the two green samples have swapped places with the red/blue ones (RGGB values have become GRBG). This looks like a memory fault in the camera; not the sensor (the JPG would be affected too), and not the SD card (only one DNG would be affected).

Interestingly, if I load the image into DCU, I seem to be able to save an uncorrupted JPG or TIFF so long as I don't "return image to full size" but leave it in "simple display mode". It only has 8 bits per pixel even if I save a 16-bit TIFF, but it might be a way of rescuing any corrupt images that you only have a DNG for. The dialogue has a check-box saying "in future do not show this message", so you might need to go to Tools/Options/Other/Message Dialog Display and reset your choice, if you ever selected that check-box.

This ought to be enough evidence for Precision to do a repair, without trying to reproduce the problem; it's possible that there's a diagnostic mode they can run to detect RAM faults, though they'd have to run for ... um ... 7s * 500 = 3500s ... OK, only about an hour, to be fairly sure of the fault showing up (7s image interval, your testing showed a problem within 500 shots). However, I'd guess it still might take them a while to fix, and/or need a return to Japan, so you might prefer to shoot RAW+ and accept the occasional failure, or at least time it for a period when you can manage without your K-1 II for a while...

Hope this helps

Cheers

Jonathan

**BigMackCam** · 10-27-2018, 01:14 PM - **4 Likes**

Originally posted by h4yn0nnym0u5e

Hope this helps

I've very much enjoyed reading your input in this thread, Jonathan. Whilst your knowledge and diagnostic ability are obvious, it's your desire to help that I applaud most. Well done, sir

**lazarustx** · 10-28-2018, 02:41 PM

Originally posted by h4yn0nnym0u5e

This ought to be enough evidence for Precision to do a repair, without trying to reproduce the problem; it's possible that there's a diagnostic mode they can run to detect RAM faults, though they'd have to run for .

This is awesome! Thanks for all your help. I can't guarantee that anyone at Precision will listen, but at least you've given me some ammo to use to get my foot in the door with some hope of finding a resolution.

Again, thank you so much for helping me out.

I'll keep you guys posted.

Regards,
Rodney

---------- Post added 10-28-18 at 04:52 PM ----------

Originally posted by prat1k

I am having the exact same problem with my K-500. It is intermittent, and independent of the lens I use. There is also an ISO issue (higher ISO gives me purple hue) - it started with infrequent corruption in images and now becoming more and more common. Using Tamron 18-250 mm lens and also tried Pentax DA 50 mm prime. No difference. Some images for example...

Good luck! There are so many areas where the corruption might be introduced. I'd double check every bit of hardware between your camera and your computer in order to nail the culprit. I'm going to open a ticket with Precision and see where it goes. I wish you luck.

Regards,
Rodney

**jbinpg** · 10-28-2018, 04:14 PM

The fact that you got a good 8-bit jpeg (the embedded jpeg?) but a crap raw file likely means it is an error in the on-camera image code somewhere. Time to send to the doctor.

**jbinpg** · 10-28-2018, 05:32 PM

Just a thought: have you reloaded the latest camera firmware yet? If not, do a camera reset both before and after reloading the appropriate latest firmware to see if that fixes things up. If not, send it in for service. - Jack

**lazarustx** · 10-29-2018, 06:38 PM

I recently reloaded the Firmware. However, I'm not sure I know what you mean when you say "Reset the firmward".

Regards,
Rodney

**jbinpg** · 10-29-2018, 07:44 PM

In one of your camera menus, you should find an option to reset the camera. You should do this to ensure that the camera will use the firmware you installed.

**lazarustx** · 10-31-2018, 02:36 PM

Originally posted by jbinpg

In one of your camera menus, you should find an option to reset the camera. You should do this to ensure that the camera will use the firmware you installed.

Thanks Jack, your reminder was very handy. I did do a reset. So here's where I'm at with debugging this on my final stretch:
1. New SD Cards installed and formetted with SD Card Formatter.

2. Camera set to RAW+JPG, Capturing PEF, and writing to Slot 1 and Slot 2.
3. Removed battery grip. There was a mention of a possible power loss being th culprit. I've removed that exposure (no pun intended).

I've made way too many changes to really attribute any one of these things as being the source of my problem. So I'll keep things static/stable for now and see if I can't recreate the problem.

I'll open a ticket with Precision if the problem reappears. Now I have some awesome tools/advice to move me forward slowly.
Regards,
Rodney

10-24-2018, 04:10 PM	#50
lazarustx Site Supporter Join Date: Mar 2010 Location: Austin Tx Posts: 209 Original Poster	Originally posted by h4yn0nnym0u5e Hi Rodney Can I just check you exported the correct images?! You say IMGP769x.x but what's in the Google Drive are IMGP799x.x. Also, I'd assume the two images that you compared are both DNGs, not a DNG and a JPG! I'm not getting a problem with these in RawTherapee 5.4, and my DCU ~~won't open any of them, I think I may need to update to a K-1 II-capable version~~ seems fine with IMGP7997.DNG, now I've updated to 5.8.1. Cheers Jonathan Sorry about that. I've uploaded them correctly this time and removed the old ones. You might need to do a refresh... - I did a binary compare of the DNG files from Slot 1 and Slot 2 - Meaning IMGP7697.DNG is the same on both Slots. - The JPG for IMGP7697.DNG was produced in camera. - The other 2 images were taken right before and right after the corrupt image. Here's the link again: https://drive.google.com/drive/u/1/folders/1u70GX33WEGaZ1S76N6T1eOsiy2mgE4sl Thanks! Last edited by lazarustx; 10-24-2018 at 06:52 PM. Reason: Added link to google drive.

10-27-2018, 12:40 PM - 3 Likes	#53
h4yn0nnym0u5e Veteran Member Join Date: Mar 2016 Location: Cambridgeshire, UK Photos: Albums Posts: 333	Originally posted by lazarustx Sorry about that. I've uploaded them correctly this time and removed the old ones. You might need to do a refresh... - I did a binary compare of the DNG files from Slot 1 and Slot 2 - Meaning IMGP7697.DNG is the same on both Slots. - The JPG for IMGP7697.DNG was produced in camera. - The other 2 images were taken right before and right after the corrupt image. Here's the link again: Update your browser to use Google Drive - Google Drive Help Thanks! Hi Rodney Here's what RawTherapee 5.4 made of it without demosaicing, i.e. showing the bare RGGB pixels straight from the sensor. The bright red dots appear to be correct values, but with a couple of extra bits set; the magenta area to the right seems to be a different error, where the two green samples have swapped places with the red/blue ones (RGGB values have become GRBG). This looks like a memory fault in the camera; not the sensor (the JPG would be affected too), and not the SD card (only one DNG would be affected). Interestingly, if I load the image into DCU, I seem to be able to save an uncorrupted JPG or TIFF so long as I don't "return image to full size" but leave it in "simple display mode". It only has 8 bits per pixel even if I save a 16-bit TIFF, but it might be a way of rescuing any corrupt images that you only have a DNG for. The dialogue has a check-box saying "in future do not show this message", so you might need to go to Tools/Options/Other/Message Dialog Display and reset your choice, if you ever selected that check-box. This ought to be enough evidence for Precision to do a repair, without trying to reproduce the problem; it's possible that there's a diagnostic mode they can run to detect RAM faults, though they'd have to run for ... um ... 7s * 500 = 3500s ... OK, only about an hour, to be fairly sure of the fault showing up (7s image interval, your testing showed a problem within 500 shots). However, I'd guess it still might take them a while to fix, and/or need a return to Japan, so you might prefer to shoot RAW+ and accept the occasional failure, or at least time it for a period when you can manage without your K-1 II for a while... Hope this helps Cheers Jonathan Attached Images

10-28-2018, 02:41 PM	#55
lazarustx Site Supporter Join Date: Mar 2010 Location: Austin Tx Posts: 209 Original Poster	Originally posted by h4yn0nnym0u5e This ought to be enough evidence for Precision to do a repair, without trying to reproduce the problem; it's possible that there's a diagnostic mode they can run to detect RAM faults, though they'd have to run for . This is awesome! Thanks for all your help. I can't guarantee that anyone at Precision will listen, but at least you've given me some ammo to use to get my foot in the door with some hope of finding a resolution. Again, thank you so much for helping me out. I'll keep you guys posted. Regards, Rodney ---------- Post added 10-28-18 at 04:52 PM ---------- Originally posted by prat1k I am having the exact same problem with my K-500. It is intermittent, and independent of the lens I use. There is also an ISO issue (higher ISO gives me purple hue) - it started with infrequent corruption in images and now becoming more and more common. Using Tamron 18-250 mm lens and also tried Pentax DA 50 mm prime. No difference. Some images for example... Good luck! There are so many areas where the corruption might be introduced. I'd double check every bit of hardware between your camera and your computer in order to nail the culprit. I'm going to open a ticket with Precision and see where it goes. I wish you luck. Regards, Rodney Last edited by lazarustx; 10-28-2018 at 02:42 PM. Reason: Cleaned up my quotation

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Leaving for Europe a week from today... Any last minute suggestions?	UserAccessDenied	Travel, Events, and Groups	48	04-28-2019 09:56 PM
Suggestions For Repairs To A Tokina	hi_power	Repairs and Warranty Service	4	08-21-2017 01:16 PM
K5 giving occasional corrupt images	Darren8306	Troubleshooting and Beginner Help	28	03-03-2012 02:28 PM
K-7 + Rapid Fire = Corrupt images	RawheaD	Pentax DSLR Discussion	9	07-14-2009 07:27 AM

10-20-2018, 12:52 AM	#46
h4yn0nnym0u5e Veteran Member Join Date: Mar 2016 Location: Cambridgeshire, UK Photos: Albums Posts: 333	Originally posted by lazarustx 6. Image 306 is the first image to show up as bad. NOTE: I did take a look at the prior image and did a compare of the corrupt image. I see that the time stamp is off by milliseconds. That's the first change. However, that's where my reading raw image skills fall apart. Anyone know how to take a look at these files to see's different? Thanks! Any chance of posting images 305 and 306 somewhere we could take a look at them? Can't promise to find time, but I am quite intrigued by this, especially as it affects more than one camera model.

10-23-2018, 08:54 AM	#47
lazarustx Site Supporter Join Date: Mar 2010 Location: Austin Tx Posts: 209 Original Poster	@h4yn0nnym0u5e I cleaned up and removed the original set, but yesterday I was able to create a few images to upload here: Update your browser to use Google Drive - Google Drive Help. Slot 1 Images: IMGP7696.DNG - Good image taken right before Bad image IMGP7697.DNG- Corrupt Image IMGP7698.DNG - Next good image IMGP7697.JPG - Good JPG produced in Camera. Slot 2 Images are Binary equal using BeyondCompare. Things to note: 1. Camera was set to shoot RAW+. 2. Camera was set to write to both to Slot 1 and Slot 2. 3. A binary compare of images IMGP7697.DNG and IMGP7697.JPG from Slot 1 and Slot 2 show them as identical. 4. If I use DCU to open the corrupt DNG, if I save the image to a TIF, I can see the corruption. However, I don't see the corruption until I do a Save/Save as. This all suggests a hardware issue from what I I can tell. If you guys think otherwise let me know. I'll have to figure out a way to have them try to recreate this inhouse @ Precision Camera. Regards, Rodney

10-23-2018, 07:54 PM	#48
frankoz Site Supporter Join Date: May 2011 Location: Glenelg North, South Australia Posts: 127	Originally posted by lazarustx @h4yn0nnym0u5e I cleaned up and removed the original set, but yesterday I was able to create a few images to upload here: Update your browser to use Google Drive - Google Drive Help. Slot 1 Images: IMGP7696.DNG - Good image taken right before Bad image IMGP7697.DNG- Corrupt Image IMGP7698.DNG - Next good image IMGP7697.JPG - Good JPG produced in Camera. Slot 2 Images are Binary equal using BeyondCompare. Things to note: 1. Camera was set to shoot RAW+. 2. Camera was set to write to both to Slot 1 and Slot 2. 3. A binary compare of images IMGP7697.DNG and IMGP7697.JPG from Slot 1 and Slot 2 show them as identical. 4. If I use DCU to open the corrupt DNG, if I save the image to a TIF, I can see the corruption. However, I don't see the corruption until I do a Save/Save as. This all suggests a hardware issue from what I I can tell. If you guys think otherwise let me know. I'll have to figure out a way to have them try to recreate this inhouse @ Precision Camera. Regards, Rodney Rodney, Good luck with Precision Camera. Please let me know if Pentax/Ricoh acknowledge a hardware fault; I will hit them up for my Ricoh GRii as well. Rgds, John

10-24-2018, 10:44 AM	#49
h4yn0nnym0u5e Veteran Member Join Date: Mar 2016 Location: Cambridgeshire, UK Photos: Albums Posts: 333	Originally posted by lazarustx Slot 1 Images: IMGP7696.DNG - Good image taken right before Bad image IMGP7697.DNG- Corrupt Image IMGP7698.DNG - Next good image IMGP7697.JPG - Good JPG produced in Camera. 3. A binary compare of images IMGP7697.DNG and IMGP7697.JPG from Slot 1 and Slot 2 show them as identical. Hi Rodney Can I just check you exported the correct images?! You say IMGP769x.x but what's in the Google Drive are IMGP799x.x. Also, I'd assume the two images that you compared are both DNGs, not a DNG and a JPG! I'm not getting a problem with these in RawTherapee 5.4, and my DCU ~~won't open any of them, I think I may need to update to a K-1 II-capable version~~ seems fine with IMGP7997.DNG, now I've updated to 5.8.1. Cheers Jonathan Last edited by h4yn0nnym0u5e; 10-24-2018 at 10:52 AM.

10-24-2018, 10:35 PM	#51
leekil Pentaxian Join Date: May 2016 Photos: Albums Posts: 2,003	Originally posted by lazarustx - I did a binary compare of the DNG files from Slot 1 and Slot 2 - Meaning IMGP7697.DNG is the same on both Slots. This would seem to rule out the cards being a problem. And likely the card writing, though not 100% sure.

10-27-2018, 01:14 PM - 4 Likes	#54
BigMackCam Digitiser of Film Join Date: Mar 2010 Location: North East of England Posts: 20,704	Originally posted by h4yn0nnym0u5e Hope this helps I've very much enjoyed reading your input in this thread, Jonathan. Whilst your knowledge and diagnostic ability are obvious, it's your desire to help that I applaud most. Well done, sir

10-28-2018, 04:14 PM	#56
jbinpg Site Supporter Join Date: Mar 2008 Location: Prince George, BC Photos: Gallery \| Albums Posts: 3,546	The fact that you got a good 8-bit jpeg (the embedded jpeg?) but a crap raw file likely means it is an error in the on-camera image code somewhere. Time to send to the doctor.

10-28-2018, 05:32 PM	#57
jbinpg Site Supporter Join Date: Mar 2008 Location: Prince George, BC Photos: Gallery \| Albums Posts: 3,546	Just a thought: have you reloaded the latest camera firmware yet? If not, do a camera reset both before and after reloading the appropriate latest firmware to see if that fixes things up. If not, send it in for service. - Jack

10-29-2018, 06:38 PM	#58
lazarustx Site Supporter Join Date: Mar 2010 Location: Austin Tx Posts: 209 Original Poster	I recently reloaded the Firmware. However, I'm not sure I know what you mean when you say "Reset the firmward". Regards, Rodney

10-29-2018, 07:44 PM	#59
jbinpg Site Supporter Join Date: Mar 2008 Location: Prince George, BC Photos: Gallery \| Albums Posts: 3,546	In one of your camera menus, you should find an option to reset the camera. You should do this to ensure that the camera will use the firmware you installed.

10-31-2018, 02:36 PM	#60
lazarustx Site Supporter Join Date: Mar 2010 Location: Austin Tx Posts: 209 Original Poster	Originally posted by jbinpg In one of your camera menus, you should find an option to reset the camera. You should do this to ensure that the camera will use the firmware you installed. Thanks Jack, your reminder was very handy. I did do a reset. So here's where I'm at with debugging this on my final stretch: 1. New SD Cards installed and formetted with SD Card Formatter. 2. Camera set to RAW+JPG, Capturing PEF, and writing to Slot 1 and Slot 2. 3. Removed battery grip. There was a mention of a possible power loss being th culprit. I've removed that exposure (no pun intended). I've made way too many changes to really attribute any one of these things as being the source of my problem. So I'll keep things static/stable for now and see if I can't recreate the problem. I'll open a ticket with Precision if the problem reappears. Now I have some awesome tools/advice to move me forward slowly. Regards, Rodney