M8 anomalies

[Nat Burgess]

Curious whether anyone else is experience the following anomolies:
1) Occasional corrupted DNG files. Shooting 10mp DNG and JPG, the JPG comes out fine but the DNG file is a quarter the size that it should be, and will preview but won't open. Or, the DNG file shows up as a file name but is 0kb. Using Sandisk UltraII 2gb cards. This problem seems to come up when the buffer is filled.
2) Occasional stripe across the DNG file (the JPG is fine); see sample below.

 

[James Roberts]

Nick--this looks strange. I've never seen--or heard of--defective DNG files, though I have had an SD card (ExtremeIII) go south on me.

Do you see the line often? This is another artifact I've never seen; I'd get that camera in for repair.

 

[Bart_van_der_Wolf]

Curious whether anyone else is experience the following anomolies: ...

I'm not a Leica user but it is not listed on the Leica user forum "Wichtig: Official Response from Leica on Laundry List" thread. I therefore suspect it is not a common thing to happen.

Assuming your firmware is up-to-date, does this also happen with other Sandisk memory cards (it's not likely to be a Sandisk thing but maybe it's a specific card issue?).

Bart

 

[John Sheehy]

2) Occasional stripe across the DNG file (the JPG is fine); see sample below.

How many different converters? Can you link to a DNG that renders this way? I can tell you if it's actually in the RAW data itself or is some other artifact.

 

[Asher Kelman]

Hi Nat,

I've not seen this but have seen errors when deltions are done of pictures taken amongs early pictures as opposed to just taken. This, Will Thompson suggest might be due to errors in filling the vacant space obtained when files are deleted. Whatever the case, I switched to Sandisk and no longer chimp and delete ... so much!

In thousands of M8 pictures I didn't see this although I'll check in iview to see that aqll the files generate a picture.

1.Do your pictures generate images to review and examine from within iview or just the jpgs?

2. What are the file sizes of the files that show small or no image from the DNG's.

If the files are small, then they are not recoverable of course. If the files are full size then they could be repaired. I have sometimes spent days working every angle to repair files and learned the hard way to throw away cards that show this.

Asher

 

[Nat Burgess]

Link to corrupted DNG files

Thanks for your ideas. I have zipped three DNG/JPG pairs (six files in total), one exhibiting the "0 KB file size" issue, one with a small and non-openable DNG, and one with a stripe across the DNG. The zip file is 16 megs.

I will use a different SD card today and see if that makes a difference. Unfortunately, it is the same make as the one that has been capturing corrupted files.

Here is the link:

Download leica error samples

PS I was able to re-create the corrupted DNG issue this morning with a different SD card. I filled the buffer, waited 30 seconds and took a single picture. The LCD displayed "Attention! Data Transfer". The DNG file was quarter size and corrupted.

 

[scott kirkpatrick]

where to report M8 problems

None of us is taking bug lists for leica, but on the Leica Camera Users' forum, www.leica-camera-user.com that is exactly what is done. Please repeat your description over there. Once you establish that the problem occurs on more than one SD card, and mention the connection with keeping the buffer filled, we'll see if you have some company in this problem. Are you running firmware 1.102, which was released today? Probably it will be good form to make sure that all new bugs get demonstrated on the current released firmware.

scott

 

[Nat Burgess]

Thanks, I will proceed accordingly

I didn't mean to try and turn this into a beta tester site - and I appreciate the time that people took to reply with suggestions, all of which were thoughtful and helpful.
-Nat

 

[Asher Kelman]

Nat,

I hope the issue gets solved!

Meanwhile any M8 pics to share without the line?

Asher

 

[John Sheehy]

Here is the link:

The one with the line in it has corruption on six different horizontal lines; a little bit of the right side of one, the next four, and then some of the sixth line. Looks like oscillations more than it does data. The fact that the corruption is not synchronized to the lines suggests file output corruption, as the 0 byte and 1/4 files also suggest. Same problem, different manifestations, most likely.

 

[Nat Burgess]

Thanks and request for clarification

Thanks for looking at the file. I am meeting a Leica rep on Saturday and would like to take advantage of your findings. Would you assume a problem with data flow, rather than with the sensor itself?

 

[scott kirkpatrick]

Thanks for looking at the file. I am meeting a Leica rep on Saturday and would like to take advantage of your findings. Would you assume a problem with data flow, rather than with the sensor itself?

The backlog of camera bodies seems to have worked itself down, so get yourself another body and give the rep a good description to go back to Solms with this one. The Leica field personnel vary a lot in their level of knowledge, so using sites like the LCU forum to identify the common elements in your problem and others' is also very important. I have seen over there more than one instance of partial horizontal lines occurring on just a few pixels' width, so I think this is a low level problem that does exist and will get sorted out through returns, as long as they are properly documented.

And I look forward to seeing your pictures here. When you post too many pictures on the LCU digital forum, people grumble and tell you to go over to the photo site side of things, where the comments are not always so useful.

good luck,

scott

 

[John Sheehy]

Thanks for looking at the file. I am meeting a Leica rep on Saturday and would like to take advantage of your findings. Would you assume a problem with data flow, rather than with the sensor itself?

Yes. The RAW data is clearly stored without corruption in the camera, because that is where the JPEG comes from; it is only when it is being written as a RAW file that the corruption occurs.