[clackers]

I can sense a knighthood ....

[Digitalis]

No, this is NOT Newton rings. Newton rings are uneven and require contact between two surfaces.

If they look like newton rings, which I have seen many times before, then the chances are high that that is EXACTLY what they are.

The effect is related to socalled Fabry-Perot filtering.

This is another theory, it is plausible. But the first problem is what we are seeing the Fabry-Perot lnterference would propagate across the image as the interference pattern isn't limited by physical contact the way newton rings are, which isn't what we are seeing here. the interference is occurring in a very small area. Secondly: that the two surfaces facing each other have to be planar to set up the conditions for Fabry-Perot effect to appear this isn't the case:



the first optical cell is dominantly spherical.

I might be stubborn as a pig but is there any reason why this wouldn't happen at the sensor array filters?

Let's put your stubbornness to good use then: I would like to see you intentionally cause an interference pattern to appear in the sensor stack.

[stevebrot]

EDIT: At least one example was seen using a 35mm film camera.

I vote for the gravitational singularity suggested by a comment on the linked post.


Steve

[Digitalis]

vote for the gravitational singularity suggested by a comment on the linked post.

@falconeye, get in here!

[GUB]

Let's put your stubbornness to good use then: I would like to see you intentionally cause an interference pattern to appear specifically in the sensor stack.

Na na my lot is finished here -- you can slug it out with DagT

[Sagitta]

The solution to dealing with this is to remove the lens from the camera and simply allow it to reach ambient temperature.

This is a good idea in general for optimal image quality (at least of you're a pixel peeper).

Moving air currents inside the camera and/or lens will cause a diffraction effect of one degree or another, as well as possibly result in condensation if the temperature difference is extreme enough. (ie, air conditioned car to hot muggy outside = wet glass)

[DagT]

If they look like newton rings, which I have seen many times before, then the chances are high that that is EXACTLY what they are.

This is another theory, it is plausible. But the first problem is what we are seeing the Fabry-Perot lnterference would propagate across the image as the interference pattern isn't limited by physical contact the way newton rings are, which isn't what we are seeing here. the interference is occurring in a very small area. Secondly: that the two surfaces facing each other have to be planar to set up the conditions for Fabry-Perot effect to appear this isn't the case:



the first optical cell is dominantly spherical.

Well, first. The Newton ring explanation is obviously not right, at the effect would occure in a place where to parts are in contact and then gradually moves away from each other. Also, it you have this problem with a lens you would see it all the time, not just with Auroras. Also, remember the days of mounting slides in frames? Nice, colourful patterns moving when the slides got warm. If it walks like a duck and quacks like a duck, it could in fact be a bird you haven´t seen before

Second: if you look at the links I referred to the shape of the pattern is the same as in the Fabry-Perot descriptions. You will have the centre of the rings at the optical axis, in the middle of the picture depening a little of the quality of the optics. The size of the pattern depends on the distance between the surfaces, i.e. how thick the filter is, and the clarity depends on the reflectivity of the surfaces at the relevant wavelength.

Third: I have tried, with Auroras. Remove the filter and the pattern disappears.

To the other questions this night (it is morning now in Norway):

The internal reflectivity and the external is the same. About 4% for uncoated glass, a lot less for coated glass, but coating efficiencies may vary with the light wavelength.

Could the pattern occure inside the lens or at the sensor? I wouldn´t say never, but the interference requires essentially parallel light rays (or wavefronts). You get this when focussing on infinity, which you get when taking pictures of the aurora. Inside the lens the rays are refracted, and are no longer parallel. Especially at the sensor plane where all rays from one place are focused at one spot and obviously are not parallel.

You do, of course, get interference at the sensor, but we know that as diffraction, which is a very local phenomena. Moving air inside the lens will be in the same range, and not cover a large part of the sensor.

Should you throw away your filters? No, but in same cases you should take it off.

---------- Post added 02-11-16 at 10:04 AM ----------

This is a good idea in general for optimal image quality (at least of you're a pixel peeper).

Moving air currents inside the camera and/or lens will cause a diffraction effect of one degree or another, as well as possibly result in condensation if the temperature difference is extreme enough. (ie, air conditioned car to hot muggy outside = wet glass)

I would not remove the lens during a transition in temperature. The weather sealing protects the camera and lens from condensation on the outside, but I would not want condensation inside.

---

Last edited by DagT; 02-11-2016 at 06:27 AM.

[stevebrot]

The weather sealing protects the camera and lens from condensation on the outside, but I would not want condensation inside.

Then you should avoid shooting where there will be sudden temperature changes. The camera is not hermetically sealed and there will always be some water vapor inside the camera and lens. How much depends on where you live.


Steve

[DagT]

Then you should avoid shooting where there will be sudden temperature changes. The camera is not hermetically sealed and there will always be some water vapor inside the camera and lens. How much depends on where you live.


Steve

Not really. Condensation comes from circulating warm air with relatively high humidity over a cold surface. There´s not enough humidity in the air enclosed inside the camera and lens to give much condensation.

We often go from -10C into a warm cottage with high humidity around here.

[Digitalis]

The Newton ring explanation is obviously not right, at the effect would occure(sic) in a place where to parts are in contact and then gradually moves away from each other.

This may be of some surprise to you but filters are commonly quite flat, the front element of the lens the OP has been using is curved. Materials shrink when it is cold, the frames for most decent quality filter frames is either brass or aluminium, the polymers used in most lenses have good dimensional stability at low temperatures but even so. Most new camera owners screw on filters rather tightly in an effort to protect their investment, all it takes is a lens with enough curvature and a filter that is close enough to contact said element to cause these rings.

Also, it you have this problem with a lens you would see it all the time, not just with Auroras.

not all the time, when the lens is warmer the materials will have expanded enough so that the interference pattern will no longer be visible.

You get this when focussing on infinity, which you get when taking pictures of the aurora.

If there is a destructive interference occurring anywhere within the optical path it will be visible no matter what focus distance your lens is set to. And in any case the OP's images show he was not focused accurately at infinity - the stars are all out of focus.




Furthermore I have attempted to generate the Fabry-Perot interference using flash with a green gel that had transmission resembling that of the aurora. I stacked a pair of thin framed Hoya multi-coated skylight filters* on my Sigma 18-35mm f/1.8 ART the filters were perfectly parallel combined with a narrow spectrum of light, I had Ideal conditions for Fabry-Perot interference to appear - However, I was unable to provoke the effect:



Which brings to mind green screen photography which uses monochromatic light to aid in the digital removal of backgrounds. Personally I use 550nm Dichroic filters for this as the purity of light they produce makes removal of the background very easy. Curiously, even under these conditions the Fabry-Perot interference patterns do not appear.

* I also stacked a pair of polarisers in an effort to provoke the effect as much as possible, again the rings didn't appear.

[GUB]

I think DagT has got this sussed in a way I am happy with the sharpness of the rings.

Looking at the OPs image (1200 pix hor) the the first couple of rings away from the centre go from light to dark in about 5 pixels (give or take). Field of view of lens (28mm) is 46degrees on the horizontal(apsc). Therefore in about 0.2 of a degree the image goes from in phase to out of phase.
(Really rough maths but suffices for the situation.)

Because the aurora is at infinity, light from any given point of it is striking the filter effectively parrallel so right across the filter it is either uniformly reinforced or cancelled. But a point of the aurora 0.2 of a degree away is oppositely treated.
So the filter does not have a visible artifact on it – I think if you were to move your eye around a little things would go light and dark. The pattern is optically at the aurora and therefore in focus.

---------- Post added 02-12-16 at 01:43 PM ----------

If there is a destructive interference occurring anywhere within the optical path it will be visible no matter what focus distance your lens is set to. And in any case the OP's images show he was not focused accurately at infinity - the stars are all out of focus.

Parallel waves from a given point (ie infinity) is a prerequisite for this effect. The fact the OP has not focussed exact doesn't alter the interplay on the filter.

[Digitalis]

If it walks like a duck and quacks like a duck, it could in fact be a bird you haven´t seen before

I grew up on a farm, I know a duck when I see one. In any case, this subject is rapidly becoming a rabbit hole of worms.

[GUB]

I grew up on a farm, I know a duck when I see one. In any case, this subject is rapidly becoming a rabbit hole of worms.

Even if your green screen was 100 metres away the light from it would be radiating out at .05 degrees which probably be still enough to upset the pattern. But I have a green laser and perhaps if I defocussed it a bit and fired it at a sheet and photographed it from 200 -300 metres away then maybe that would start to recreate the situation.

[Digitalis]

Newton rings are uneven and require contact between two surfaces

Pentax K5IIS - Sigma 180mm f/3.5 APO EX ~ ISO 80 F/32 + 82mm extension tubes, Godox AD180 flash used at 1/2 power. Image contrast has been enhanced.

yes they do require physical contact. I sandwiched a multi-coated Hoya +3 diopter macro lens with a a Hoya UV[0] filter to create the Newton rings in the above image Also note the evenness between the rings*. And I have to say I'm wrong about it being contact between the front element and the filter, the rings would be completely out of focus. But having said that: I don't think it is Fabry-Perot interference either. Because it requires so many environmental conditions to comply for it to occur.

The principal reason is that the interference would have to propagate throughout the entire optical path for it to be visible at the sensor plane - if this were the case it would be observable in live view or by looking through the viewfinder. The interference between the first two elements it wouldn't even be visible. It would have to occur near the rear element of the lens and I don't think it is possible for interference to spontaneously be created in an optical system like that. As for my newton ring theory: it transpires there are two elements that are made of different materials ( and different expansion coefficients) that I suspect that could also be a source of the interference pattern, and since the Lens in question is designed for APS -C the rear element will be closer to the sensor so any specks of dust or interference patterns even if the lens is being used wide open - will be visible to the sensor.


The point here is how to prevent the interference - and almost universally the advice is to remove the filter.

*perspective distortion has warped them somewhat.

[GUB]

So is this on a filter attached to your lens?
I take from your editing it is not attached to your lens.