Banding; advice needed

[Pointcolville]

Advertisement (gone after registration)

First shot with Lens detection UV/IR off, no UV filter, second one shot with filter and UV/IR turned on. Not an anomaly; this effect appears in every shot bifurcated by the chimney. Is this the legendary banding problem requring a firmware fix? I bought my M8 from Calumet in April, sn 3108xxx

[marknorton]

I would say the camera is faulty - the sensor is read out in two sections, left and right and when there's a mismatch between the two sides, it's called the "centre-fold problem" and normally it's very slight. What ISO were you using?

 

The banding occurs (on unmodified cameras) when there are bright light sources in the image and the bands spread to the centre. In addition, all cameras suffer to a greater or lesser extent when there's a bright light right at the edge. From the serial number, it will have already been modified for the first problem.

 

This is altogether more serious, send it back.

[Advertisement]

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

[ho_co]

Nat--

As Mark says, this is not a common problem. The split down the center is due to an imbalance between the two sides of the sensor as he said. The camera should have gone back to Solms for the upgrade, or been produced after those modifications, but it may not have. In fact, this 'center-fold' problem was fairly common with the pre-upgrade cameras.

 

Looking at the EXIF data may say something about the color differences.

 

In both cases, the firmware identifies the max aperture as 1.0, but doesn't identify the lens. That probably means you are using an uncoded lens. The picture looks to me as if you're using a wide-angle.

 

In the "no filter, no lens ID" image, the camera applies its default vignetting correction. Since you are using no IR-cut filter, you are getting the effect of IR. That effect is worse on one side because: 1) That's where your lamps are (the source of the IR); and 2) The sensor's output is out of balance.

 

In the "lens ID + filter" image, the camera still doesn't know what correction to apply since the lens is uncoded, but the camera applies a general cyan-drift correction in addition to the vignetting correction. Since you're using the UV/IR-cut filter, the color is better than in the first example. However, the split-readout nature of the sensor is still visible.

 

As I recall, the 'centerfold effect' was seldom if ever seen at an ISO as low as 160.

 

My guess is that the camera has not been to Solms for its initial upgrade. IIRC, only Solms can do the circuitry upgrade.

 

You say these are typical results in this situation. You might try shooting from the chimney back toward where the camera is standing in these images; the magenta cast should switch sides.

 

The center split will probably be visible in other cases as well, particularly at high ISO's.

 

Good luck!

 

--HC

[Pointcolville]

Thanks, this sheds a more even light on the problem (so to speak. . .). Yes, this was taken with an uncoded wide angle. The exposure was two seconds at f5.6, ISO 160. I will experiment as you suggested to see if the fault recurs on the same side of the sensor.

Your insights are much appreciated - thanks for sharing them.

[marknorton]

Are you seeing this on all shots?

[Jamie Roberts]

Forget experimenting. Unless you had a half-yellow filter over the lens, send the camera back; it's defective.

[marknorton]

Advertisement (gone after registration)

I agree!

[ho_co]

I agree with Mark and Jamie that the camera needs fixing.

 

But remember, this is the weekend. Checking for the split center at high ISO would be a simple and enjoyable way to pass a cold day indoors...

 

--HC