green banding on long exposures with 10x nd

[brill64]

Advertisement (gone after registration)

i've recently found strong green banding (possible sensor issue) and a strange ghosting artifact (possible internal reflection) appearing with my m9 on exposures taken at between 1.5-180s., at apertures between f4 & f16 using a 28 summicron & 75 summicron. The banding became apparent on exposures made during early evening & night time and the ghosting only during early evening with a 10x nd filter attached. the nd is a b+w 49mm which i used with a 46-49 step-up ring. It didn't appear in bulb long exposures with the nd filter off or when a polariser was fitted to increase exposure time. in all cases there was no strong sidelight or sunlight just outside or at the edges of the frame, which sometimes caused green banding on the m8 sensor. i wondered if the sole cause was the use of a 10x filter on this type of sensor causing this at b setting or whether i need to get a leica tech to look into it? thanks for looking

 

four examples posted:

9287 180s f4 iso 160 28 summicron+nd 10x

9774 90s f4 iso 160 28 summicron+nd 10x

9754 24s f4 iso 160 75 summicron+nd 10x

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!

[gravastar]

The green banding looks like a sensor problem. The sensor imaging area is read out in two halves, left and right hand side. If you look at the bands some are only on one half of the image. I don't know what the Kodak/Leica specification for acceptable performance is for very long exposure. The only way to find out is to send the images to Leica customer service and ask.

 

The second problem looks like flare caused by the sun outside of the left hand side of the picture. What causes the flare to have a straight vertical edge on the right is a mystery unless you were using a rectangular lens hood. Maybe it's a multiple reflection between the rectangular sensor and filter?

 

Bob.

[Alnitak]

Yeah, that looks like a sensor problem to me, but Bob is right, that may also be "out of spec" on that sensor. I had a similar problem a few years back with my Canon 1Ds Mark II, and Canon *eventually* replaced the sensor, but they also indicated that I was pushing the acceptable limits of their design (of course, I didn't have the same issues with a 5D or a 20D at that same time, so I always felt they were a little full of BS). The replacement was much, much better.

 

Jeff

[250swb]

I don't know because I wasn't there, but the banding looks like its in areas of the image that could be severely under exposed, and then perhaps over processed? For instance the brighter horizon doesn't have banding. Maybe the exposure time wasn't long enough.

 

Steve

[adan]

As to the ghost or flare - I'm not sure of the main cause, but I can venture a guess as to the reason for the sharp edge. The sensor essentially lives in a shallow cave formed by the shutter opening (noun, not verb). Which can cast a shadow if the source point of a reflection is well off to the side, giving a hard edge to an otherwise blurry patch of light. The rainbow coloring is diffraction or some other prismatic effect of the microlens pattern (I've seen it in the context of other flare-type events).

 

As to the bands: It looks like a sensor or circuitry problem to me.

 

I don't normally do long exposures (my M9 has 12,000 exposures - which between them probably don't add up to 120 seconds of total exposure time ) - but tried a 120-sec. exposure just now (indoors, night, f/16) and I'm not seeing any banding at all, let alone as severe as what you got.

 

I'm assuming the pictures were correctly exposed and didn't need any heroic brightening in post-processing.

 

Edit: actually 250swb's post makes me remember one other factor - the 28 Summicron vignettes rather a lot - the "correction" applied in response to reading the coding in effect "pushes" the corners to brighten them, and that is probably what is making the bands so emphatic in the corners. But I still suspect there should not be that much banding pattern to "push" in the first place.

[ho_co]

1) Breaking the line in half indicates a sensor or sensor readout problem.

 

2) Aside from that, the overall pattern looks similar to the M8 "Jesus stripes" elsewhere reported. Possibly a carryover from M8 to M9?

 

 

Send a file to Leica and ask their advice.

[brill64]

Advertisement (gone after registration)

thanks all. you have convinced me that i should send my camera in for checking. although reluctant to do that of course because otherwise it serves me perfectly well. i did send images to leica several days ago and i'm waiting to hear back from them before going further. i tend to slightly underexpose as a rule but the first image was brightened up and perhaps overprocessed a little in ps4 to show how prominent the banding was in this relatively long exposure. a lot of detail gets lost at such small size on the forum. the others, although shorter exposures were okie exposure-wise i guess. clever to spot that both sides of the image banding are not equal and i think this shows there could be something wrong so thanks for pointing that out. the prismatic ghosting was more prominent with the 75 than the 28 and yes, the 28 does vignette and tend to push the corners a bit. i didn't use a lenshood on the 28 but i did on the 75 which was fully out. i also wondered if this kind of exposure was a bit out of spec with the sensor especially when the 10x nd seemed to exaggerate the issue more. it does remind me of m8/m8.2 jesus stripes which i think is fixed now with a recent m8 update. will let you know any progress here. thanks again for your thoughtful input and all the best!

[ho_co]

You're in error in regard to the "Jesus stripes" phenomenon. As I explained in that thread, the problem:

 

a) was not fixed by a firmware update; and

was not fixed by an initial replacement of sensor and circuit board.

 

The camera is currently at Leica for a second replacement of sensor and circuit board.

 

The "Jesus stripes" were immediately recognized by Leica as a hardware issue, not firmware.

 

Your problem is similar but not the same, and seems to be hardware-related.

 

Keep us posted.

[brill64]

thank you, howard for correcting me on that. also a special thank you to adan for trying a 120s exposure for comparison. i will certainly keep you posted.

[mchular]

What was the response from Leica about the green banding?

[brill64]

i've not heard back from them yet.