Strange M9 image corruption - part of latter image in prior image.

[smoody]

Advertisement (gone after registration)

I'm getting worried about my M9. It's about three months old and, within the last few days, a series of errors started occurring.

 

FYI, I'm using 16gb SanDisk Extreme II cards (as recommended by Leica). I always shoot uncompressed raw. I tend to shoot with three-shot exposure bracketing enabled while doing street photography. I handle the camera with kid gloves -- it is better cradled than most babies. It has always worked fine (minus the occasional 'card locked' and 'card full' errors), but today things started to get strange...

 

The image below is a 100% crop from part of a capture from just a couple of hours ago. It's odd to me that the image could be corrupted and still result in a valid DNG, but that's not the strange part...

 

The part of the image that appears offset from the rest of the crop is from the NEXT FAME in the three frame exposure bracketing sequence. So, somehow, the M9 put part of a later image in the present image. About ten or twenty minutes later, a shutter fault error occurred (had to turn camera off and then back on to reset it and things worked fine after that). Now I've started noticing a delay between the first two shots of a a three-bracket shot and the third. It takes the first two at normal speeds and then pauses to the point where I think it's not going to take a third and then, when I've just about given up hope, it takes the third. It doesn't happen all of the time but it happens frequently.

 

Theories concerning rips in the time-space continuum aside, has anyone else seen this behavior? Is it a sign that my M9 is about to take a dirt nap? If so, should I send the camera to Leica now as a preemptive measure?

 

BTW, the image following the cropped image below (the second of three frames) is fine, as is the third image in the sequence and the rest of the images from the day's shoot. Also, today wasn't particularly cold: 40F/5C

 

(resized crop from part of image)

 

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[wparsonsgisnet]

This problem has been reported before from two causes, as I recall.

 

One problem might be that the SC card needs reformatting. Do you reformat cards or erase the files, and do you do it in camera? There have been several recommendations to reformat and to do it in the camera. I do this, for example.

 

A second problem (my favorite bete noire) is a battery that with a low charge. On the M8, this was a serious problem for me and I always replaced at the last segment. I have been observing the same procedure with the M9, tho it seems sturdier in this area. However, if your battery is low, the camera may "misbehave" and this is a problem that I would expect.

 

Try a different card, formatted in the camera, and make sure your battery is charged up.

 

Keep us posted, too, please.

[250swb]

I would also say do you keep your card healthy by formatting it to erase files? The auto bracket problem may be a case of the camera having a full buffer, its only seven shots with .dng files so two or three bracket sequences one after the other is going to fill the buffer, the delay before the last shot is the wait for the buffer to clear enough space.

 

Steve

[Alnitak]

Funny, I heard a bit about this problem initially, but hadn't heard anything in a while. Today I had the problem occur. I have had my M9 for about 60 days and have run a few thousand shots through it without a problem. Today I got one shot with the problem you report.

 

It is a new Sandisk 16GB Ultra II card (the one recommended by Leica), and has been used one other time in the M9 without any problem. I format it in the camera. The battery on the camera was nearly totally full; I had changed it about 30 minutes earlier. I'm hoping it doesn't recur; luckily I took 2-3 shots of this subject and got one I wanted to keep.

 

 

Jeff

[smoody]

It is a new Sandisk 16GB Ultra II card (the one recommended by Leica), and has been used one other time in the M9 without any problem. I format it in the camera. The battery on the camera was nearly totally full; I had changed it about 30 minutes earlier. I'm hoping it doesn't recur; luckily I took 2-3 shots of this subject and got one I wanted to keep.

 

Jeff

 

Yes -- exactly! My SD card was also freshly formatted, same brand and model, fresh charge on the battery, three shots and one was a keeper. It's possible we've even taken close to the same number of shots with our M9s. I'll repost if the problem rears its ugly head again. Hopefully it's a one-in-two thousand shot problem.

 

Perhaps the good folks at Leica will come across this thread and find a fix? :-)

 

Scott

[Alnitak]

I'm skeptical its the card. My guess is that the code for writing to the buffer/flushing the buffer/writing to the card is a bit buggy. I hope so, as the speed issue is a huge problem as well.

[lejos]

Advertisement (gone after registration)

The same thing just happened to my M9 twice this week,

both times while shooting several shots in succession in single mode.

Full battery, two different cards (2gb Sandisk Ultra II, Extreme III) both

formatted in camera, uncompressed DNG no JPG, no bracketing.

I tried to replicate the problem using continuous mode, but this way the

camera works fine.

In one case it was the 6th shot, in the other case the 7th, so it seems to be

a problem of the buffer filling up.

[mjh]

The part of the image that appears offset from the rest of the crop is from the NEXT FAME in the three frame exposure bracketing sequence. So, somehow, the M9 put part of a later image in the present image.

This kind of issue has been reported here before. I am not totally convinced it is a mixup of two different frames rather than parts of the same frame, but in any case something appears to go wrong when shooting images in fast succession.

[Alnitak]

I've personally only ever seen it once. Lejos, what firmware version are you running? I wonder if this is something that was fixed in the new firmware?

 

Jeff

[jaapv]

We did not hear of this since the last firmware. I have a feeling that it had to do with the initial SD problems of the M9. What firmware are you running?

[lejos]

I'm running 1.116.

And it is definitely a part of the next frame showing up.

[ho_co]

I think it's part of the M9's IR-correction routine. You'll notice that on both these images, the main 'wrong' part is blacks and grays. The camera seems to be seeing that consecutive frames are similarly toned, and the comparison section is improperly getting dumped to the card. Just one of the safety fallbacks added after the M8's IR problems AFAIK.

 

Have you noticed the problem shooting anything magenta?

 

 

Okay, okay, it's a joke.

[Alnitak]

I'm running 1.116.

And it is definitely a part of the next frame showing up.

 

Can you make it happen by trying? To be honest, I never tried to make it happen after the one time it happened to me. If you can repeat it, that would be interesting.

 

I'll have to try my camera later tonight.

 

Jeff

[dalippe]

We did not hear of this since the last firmware. I have a feeling that it had to do with the initial SD problems of the M9. What firmware are you running?

 

This looks very related to a problem the M8 had (http://www.l-camera-forum.com/leica-forum/leica-m8-forum/53991-m8-glitch-merged-frames.html), although there it wasn't concentrated in subsequent pictures. So I suspect it is something that was never fixed and found its way from M8 to M9 since the M9 is so heavily based on the M8. I would be surprised if Leica ever fixes this on the M9 given how long it has survived so far. It will probably have to wait until a truly redesigned digital M comes along.

 

Although not discussed on that thread, I subsequently determined it was unlikely due to a card problem since it repeated itself with various cards at various times, and those cards were subsequently used with the M8 as well as other cameras without problem.

 

David

[scott kirkpatrick]

I had one of these last Saturday as well. First time in about 3000 shots with the M9. The moon is nearly full, so perhaps that is a factor. I was using a nearly filled 8 GB Extreme III card, formatted in the camera several days earlier. The inserted rectangle of slightly lighter image is from the following frame, which was produced without problem (although over-exposed). Between the two shots I reset the exposure by sampling in a darker region and holding the half-pressed shutter while I recomposed, so a second or two had elapsed.

 

Firmware 1.16 I only noticed this while editing the day's shots in Capture One, but it is visible in the camera as well, after the pause during which the full image comes up on the LCD.

 

scott

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[mjh]

It is a truly baffling phenomenon. Initially I had thought that maybe some process was writing to a buffer on the assumption that it had been released, but was actually still read from by another process. But even then one would have expected the superimposed image to appear at the edge of the image, not somewhere near the center. And when there is at least a second between the two exposures, it is unlikely that the process handling the second frame could catch up to the one still handling the first, so they would both access the same buffer. Also, why would the second image be complete then? It is like the second process realized its mistake: “Oops, wrong buffer! Sorry for making a mess of it; I’m starting from scratch using my own buffer.”

[k-hawinkler]

Maybe it's just an obscure programming bug in the FW?

 

K-H.

[mjh]

Maybe it's just an obscure programming bug in the FW?

I am almost certain that it is, but that doesn’t quite explain what exactly is happening here. And it’s the obscure bugs that are hard to squash. The fact that it happens only rarely doesn’t help either. Let’s hope Leica is more successful in understanding the issue than we are.

[Advertisement]

Hallo,

Look here: Leica M

[k-hawinkler]

I am almost certain that it is, but that doesn’t quite explain what exactly is happening here. And it’s the obscure bugs that are hard to squash. The fact that it happens only rarely doesn’t help either. Let’s hope Leica is more successful in understanding the issue than we are.

 

Michael,

 

Thanks. That would be nice. Maybe, based on the information they have, Leica could device a series of stress tests to force the error. Cameras are not the only area of technology that have to deal with intermittent exceedingly rare (and sometimes even soft) errors. Luckily, Leica seems to be using mostly standard parts - according to marknorton's disassembly and successful assembly of an M8. Otherwise, Leica would most likely have an even bigger problem to debug their cameras electronically, IMHO.

 

Best Regards, K-H.

[Shade]

Anything that has to do with image corruption means I will send my camera back to Leica. I would not want to risk taking pictures and ended up wanting to keep the image but then finding out its corrupted. Its a bummer..