[Mrbossman]

Hi,
I am a beginner in photography. I currently own a Pentax K-5 with lens Tamron SP AF 17-50mm f/2.8 (IF).
Recently, I have noticed many of my night shots have these weird circular rings at the center of the photos. Please advice!
I was using manual focus. Anyone knows what is causing it and how i can fix those photos?

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[Unregistered User 8]

Hi Mrbossman. In the " search the forums " box, type in Newton's rings.

[Bruce Clark]

It certainly looks like Newtons Rings. These are an interference pattern caused by light passing through multiple layers of glass or other transparent materials in close contact. Glass and film in contact, prints on a scanner bed often produce similar effects.

I can only guess at the cause in your case. Other threads in the forum suggest filters or faulty lenses as possible causes. Other suggestions are internal reflections between the sensor and lens elements. How to fix it apart from identifying and removing the cause is a bit more problematic.

[Paul the Sunman]

If you have a filter attached, take it off. It's the most likely cause.

If that doesn't fix the problem, try shooting at other focal lengths as an experiment: both examples you posted used 28mm. Other focal lengths will alter the distance between elements, which should alter the Newton's rings pattern if internal reflections are to blame. I haven't used this lens, but I would be surprised if that was the cause.

You might also try stopping down a little. Again, both examples were shot wide open, which isn't a good idea anyway. See if that changes anything.

[beachboy2]

Are your photographs taken from inside behind glass or outside in the freezer? If outside are you getting dew or condensation on the lens.

[GUB]

Hi
Firstly great landscapes!!
Searching for "Newton rings" on the forums certainly shows some similar issues. The general consensus seems to put the blame on filters and elements but I am a little sceptical of that. Basic optics would dictate that interference patterns at the front of the lens would be so out of focus as to be invisible. Your rings are well defined. The other similar images as well as yours have another thing in common - a cold environment and a long exposure and I am wondering if sudden cooling has curved the aa filter and the patterns are interference patterns between it and the sensor. Someone with practical knowledge of how close the filter is to the sensor could well shoot this theory down.

[Digitalis]

Someone with practical knowledge of how close the filter is to the sensor could well shoot this theory down.

I intend to.

The AA filter is also combined with the optical high pass hot mirror for cancelling out IR and UV light, these plates of glass are very flat and close to the sensor* but still not directly attached to it. These materials are designed to have very similar expansion coefficients - therefore if subjected to high or low temperatures their expansion rates would be close to identical, and thus preventing any unwanted artifacts.

In the highly unlikely event of filters mounted so close to the sensor were to cause newton rings, the effect would be visible across the entire frame. Instead what we are seeing here is a small part of the image is affected, right at the expected apex of curvature of the entrance pupil of the lens. The front element of the lens is undoubtedly thicker than the glass of the filter, and therefore it needs to lose more thermal energy to reach temperature equilibrium with the environment - the filter can accomplish this much faster as the thickness of the glass in it is uniform. therefore the filter could be touching the lens, creating the newton rings, If you remove the filter the effect should disappear.



Looking at the optical design of the Tamron 17-50mm f/2.8 there are several points of contact between planar and curved surfaces**, and at least one of them in the rear group are between glass types of different chemistry which could theoretically have different expansion coefficients - this is another possible source of newton rings. The solution to dealing with this is to remove the lens from the camera and simply allow it to reach ambient temperature.

*It varies from manufacturer to manufacturer, but generally the AA /OPLF stack is about 0.5mm away from the sensor cover glass. Some degree of curvature of one or more elements in the optical path are a required for concentric newton rings of this type to appear.
** which are a known condition for concentric newton rings to appear.

---

Last edited by Digitalis; 02-07-2016 at 10:29 PM.

[GUB]

I intend to.

The AA filter is also combined with the optical high pass hot mirror for cancelling out IR and UV light, these plates of glass are very flat and close to the sensor* but still not directly attached to it. These materials are designed to have very similar expansion coefficients - therefore if subjected to high or low temperatures their expansion rates would be close to identical, and thus preventing any unwanted artifacts.

In the highly unlikely event of filters mounted so close to the sensor were to cause newton rings, the effect would be visible across the entire frame. Instead what we are seeing here is a small part of the image is affected, right at the expected apex of curvature of the entrance pupil of the lens. The front element of the lens is undoubtedly thicker than the glass of the filter, and therefore it needs to lose more thermal energy to reach temperature equilibrium with the environment - the filter can accomplish this much faster as the thickness of the glass in it is uniform. therefore the filter could be touching the lens, creating the newton rings, If you remove the filter the effect should disappear.



Looking at the optical design of the Tamron 17-50mm f/2.8 there are several points of contact between planar and curved surfaces**, and at least one of them in the rear group are between glass types of different chemistry which could theoretically have different expansion coefficients - this is another possible source of newton rings. The solution to dealing with this is to remove the lens from the camera and simply allow it to reach ambient temperature.

*It varies from manufacturer to manufacturer, but generally the AA /OPLF stack is about 0.5mm away from the sensor cover glass. Some degree of curvature of one or more elements in the optical path are a required for concentric newton rings of this type to appear.
** which are a known condition for concentric newton rings to appear.

But why would such an issue within the lens be resolved on the sensor?. After all you can throw the iris pretty well anywhere inside the lens and while it lessens the light overall, it remains pretty well invisible. If the sensor filters cooled on their rims faster than their main body of glass then that could well temporarily create a dish effect bringing the centre closer to another surface. (I am thinking of heat effects on metal panels here)

[Digitalis]

After all you can throw the iris pretty well anywhere inside the lens and while it lessens the light overall, it remains pretty well invisible

No, you can't put an aperture just anywhere in a lens..just no.

But why would such an issue within the lens be resolved on the sensor?

Have you ever seen a dust spec on your camera sensor?

If the sensor filters cooled on their rims faster than their main body of glass then that could well temporarily create a dish effect bringing the centre closer to another surface.

If this did happen the temperatures would equalize in a few seconds the metal used to hold the filters in place are quite thin - the glass and crystals used in the UV/IR AA stack are small and are pretty good at reaching thermal equilibrium.

The "dish effect" can be debunked as glass and crystals used in the stack are cemented together, and glasses in general simply do not have that degree of malleability - the primary failure mode for glass under stress is fracture, not creep. The AA crystals would simply cleave along the weakest axis perpendicular to the stress, the UV/IR filter would simply shatter. To induce measurable creep in the UV/IR AA stack would only occur at temperatures above the glass transition temperature around 500 °C. The volumetric expansion coefficient of borosilicate glass is about 9.8 @20°C by comparison, the volumetric expansion coefficient of Invar, is about 3.6 @ 20°C

So If you want to really test your hypothesis, you're welcome to light up your oxy/acetaline blowtoch and stick it in your mirror box.

[GUB]

Have you ever seen a dust spec on your camera sensor?

Dust bunnies run amok on my sensor (another use for the " oxy/acetaline blowtoch")
But they are not within the lens as described and their shadow on the sensor is totally to be expected as well as the fact they sharpen when the lens is stopped down.
It is their presence that makes me look to the sensor area for this issue.

"No, you can't put an aperture just anywhere in a lens..just no"

Yes putting it in other areas within the lens will cause vignetting and alter the value of the f stop but would never create a sharp image of the iris on the sensor.
The image of the newton rings is well defined and about as sharp as actual photographs of them normally are. Maybe a filter with dramatic rings placed some distance in front of a well stopped down wider angle lens could create a visible image.

Mrbossman -- some interesting reading if you google "newton rings on sensor" especially in the astrophotography area.
and ; http://www.dpchallenge.com/forum.php?action=read&FORUM_THREAD_ID=168176
and a link from that link : http://www.ptialaska.net/~hutch/aurora.html
It is possible the monochrome green of the aurora is a factor.

[Digitalis]

It is their presence that makes me look to the sensor area for this issue.

The sensor cover glass and AA crystals and UV/IR are all planar surfaces: you're barking up the wrong tree.

The image of the newton rings is well defined and about as sharp as actual photographs of them normally are.

The geometric structure of the rings is a good indicator of their origins in a photographic optical system. They will appear regardless of F-stop because it is a very destructive interference pattern.

If the optical interference patterns were occurring between planar surfaces such as the AA UV/IR stack they would look like this:



Instead we are getting concentric rings that look like this:


(Image deliberately enhanced to show rings)

Which indicates a spherical surface impacting on a plane in the optical path. By definition, lenses have curved surfaces - filters are flat, this combination is all that is needed to cause newton rings.

It is possible the monochrome green of the aurora is a factor

The monochromatic nature of light from the aurora is certainly a factor in heightening the intensity of the newton rings. Under white light, they are prismatic - colours are split according to frequency, but under monochromatic light they take on alternating light/dark patterns.

The theory that the spectrum selective lens coatings are destructively interacting, or amplifying the effect with the monochromatic light is an interesting hypothesis. But i'm skeptical as the coatings on the lenses we use are very broad, and have multiple thicknesses and the surfaces they are applied to have differing geometry* which makes the propagation of optical interference patterns, especially destructive ones at selective light frequencies: to be highly unlikely.


*such as Aspheric elements.

[GUB]

The sensor cover glass and AA crystals and UV/IR are all planar surfaces: you're barking up the wrong tree.



The geometric structure of the rings is a good indicator of their origins in a photographic optical system. They will appear regardless of F-stop because it is a very destructive interference pattern.

If the optical interference patterns were occurring between planar surfaces such as the AA UV/IR stack they would look like this:



Instead we are getting concentric rings that look like this:


(Image deliberately enhanced to show rings)

Which indicates a spherical surface impacting on a plane in the optical path. By definition, lenses have curved surfaces - filters are flat, this combination is all that is needed to cause newton rings.



The monochromatic nature of light from the aurora is certainly a factor in heightening the intensity of the newton rings. Under white light, they are prismatic - colours are split according to frequency, but under monochromatic light they take on alternating light/dark patterns.

The theory that the spectrum selective lens coatings are destructively interacting, or amplifying the effect with the monochromatic light is an interesting hypothesis. But i'm skeptical as the coatings on the lenses we use are very broad, and have multiple thicknesses and the surfaces they are applied to have differing geometry* which makes the propagation of optically resonant interference patterns, especially destructive ones at selective light frequencies: to be highly unlikely.


*such as Aspheric elements.

The theory that the spectrum selective lens coatings are destructively interacting, or amplifying the effect with the monochromatic light is an interesting hypothesis. But i'm skeptical as the coatings on the lenses we use are very broad

Couldn't this description also apply to the ir/uv cut filter? Their cut offs are very dramatic from what I remember.

[Digitalis]

Couldn't this description also apply to the ir/uv cut filter? Their cut offs are very dramatic from what I remember.

By which mechanism would you suggest this would be possible? the UV/IR filter absorbs frequencies higher than 400nm and reflects and dissipates frequencies lower than 700nm. The light from the aurora in the images is within 490-560nm frequency range.

[GUB]

By which mechanism would you suggest this would be possible? the UV/IR filter absorbs frequencies higher than 400nm and reflects and dissipates frequencies lower than 700nm. The light from the aurora in the images is within 490-560nm frequency range.

I don't pretend to know so that is why I am asking.
But an extract from that link we were both referring to ;
Thanks to the University of Alaska forecaster, Chuck Deehr, the explanation follows.
"These are interference fringes due to the parallel faces of the filter and to the narrow spectral emission at 5577 Angstroms in the aurora. That green, atomic oxygen emission line is the strongest emission in the aurora near our film and eye peak sensitivity, so it shows up first when there is any device in the optical path which sorts out the spectral emissions." So, don't use filters!.

I haven't found a reason why an interference pattern created by either the filter close to the front of the lens or from the lens itself would be visible as an image on the sensor. It breaks the basic rules of optics. The only exception to this I can think of is airy disks (which this isn't) which are created and visible due to understandable circumstances.

Given that the aurora light waves are at "infinity" then they are striking the front filter perpendicularly and that is why I feel that the parallel surfaces of it are not an issue. But not so on the sensor array where the focusing rays are on average radiating out from the optical centre of the lens.

[Digitalis]

These are interference fringes due to the parallel faces of the filter and to the narrow spectral emission at 5577 Angstroms in the aurora. That green, atomic oxygen emission line is the strongest emission in the aurora near our film and eye peak sensitivity, so it shows up first when there is any device in the optical path which sorts out the spectral emissions

There are two problems with this: the interference pattern would also need a second monochromatic light source out of phase with the primary light source to create the interference. Secondly the interference pattern wouldn't look like newton rings: the interference pattern would be irregular, and cover the entire image.

I haven't found a reason why an interference pattern created by either the filter close to the front of the lens or from the lens itself would be visible as an image on the sensor.

The interference pattern is quite large and the focal length is very short, even at wide apertures any feature larger than a pea on the front element will be visible. Judging from a focal length of 17mm this pattern would be approximately 7mm across

This is what a 5mm piece of blu-tak on the front element of my sigma 18-35mm f/1.8 looks like at 18mm@f/2.8