lines across image

[JohnSadlerAstro]

12 hours ago, Stub Mandrel said:

I looked up the referenced quote about the banding being caused by the AF and it categorically states that dithering can cure it.

I stand corrected. I understood that the banding would be strong enough to prevent elimination, but clearly this is not the case.

John

[Stub Mandrel]

I note that Canon dropped the Hybrid CMOS AF for the 850D.

It seems that there are seven rows of AF points, each of which leaves a dashed trail of four faint lines that appear more like a pair under normal observation.

The pattern seems to be largely 'density neutral' i.e. it appears as a pattern but doesn't have a big impact on luminosity at a scale that banding routines can easily remove it.

Presumably the AF pixels are connected differently which means they have a different sensitivity, and presumably this is non-linear meaning control frames can't remove it.

The older (and newer) diffuse AF patterns get lost amongst random noise, it seems.

32 minutes ago, ollypenrice said:

To be honest the cooled CMOS cameras are so much closer to DSLR prices than was the case with CCD that avoiding these hassles at source seems to me to be more desirable than ever.

I will be waiting and saving for used 1600MMC cameras to become available ?

[Stub Mandrel]

1 minute ago, JohnSadlerAstro said:

I stand corrected. I understood that the banding would be strong enough to prevent elimination, but clearly this is not the case.

John

I suspect that you need a good number of subs to do so.

[JohnSadlerAstro]

10 minutes ago, Stub Mandrel said:

I suspect that you need a good number of subs to do so. 

My (probably flawed) logic would suggest that in order for it to be entirely removed, the dithering would need to cover each patch of sky with those bands to prevent it from showing in the final image. That's a lot of dithering. ?

It's good that my 1000d never suffered with this. I know it has some very sparse AF points, which is probably as well to be honest!  

I hope the op will be able to find a solution. ?

John

[steppenwolf]

2 hours ago, ollypenrice said:

How wide are the bands?

Looks like 20 pixels deep with each pair 15mm apart and each pair repeated every 245 pixels (all figures approximate).

That would need a huge dither ?

[ollypenrice]

16 minutes ago, steppenwolf said:

Looks like 20 pixels deep with each pair 15mm apart and each pair repeated every 245 pixels (all figures approximate).

That would need a huge dither ?

Yup, I don't think dither has much hope of dealing with that.

Olly

[gonzostar]

Yikes, 20 pixels deep, That like the Mariana trench!  Well i much appreciated everyone's efforts in looking into this problem. If the experts are left scratching there heads then it leaves me to get a refund for this camera. Then back to the drawing boards.

Clear nights all

Dean