Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Interesting artifact w/ Canon 6D

  • Please log in to reply
90 replies to this topic

#1 jimegger

jimegger

    Vendor

  • *****
  • Vendors
  • topic starter
  • Posts: 1621
  • Joined: 22 Dec 2005
  • Loc: Palmer,Alaska

Posted 11 December 2015 - 02:40 PM

I used a Baader Contrast Booster filter with My modded Canon 6D while imaging NGC 404 and got this interesting artifact showing up around the bright Beta Andromeda. I used my 9.25 inch ACHRO refractor for 4 shots of 7 minutes each at ISO 1600. The filter was placed on the end of an extension about 4 inches from the sensor and no other optical window was in between the sensor and the filter. I suspect this is the result of a bright point of light interacting with the Bayer matrix in some way. Has anyone else experienced such an artifact ?

Attached Thumbnails

  • NGC404 Combine1.jpg


#2 Joepie

Joepie

    Explorer 1

  • -----
  • Posts: 78
  • Joined: 15 Oct 2012
  • Loc: The Netherlands

Posted 11 December 2015 - 03:39 PM

I know about someone who had similar problems, however his camera had been mono-fied, not just modded.



#3 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23290
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 11 December 2015 - 03:40 PM

That almost looks like a reflection of the sensor itself, highly magnified of course, but showing the pixels. Kind of seems like the microlens diffraction issue people using KAF-8300 sensors have had, but different...more precise. 


  • Michael Covington likes this

#4 Jerry Lodriguss

Jerry Lodriguss

    Vendor

  • *****
  • Vendors
  • Posts: 7061
  • Joined: 19 Jul 2008
  • Loc: Voorhees, NJ

Posted 11 December 2015 - 05:54 PM

I've gotten that before also on extremely bright objects when they are overexposed.

 

Jerry


  • Michael Covington likes this

#5 sharkmelley

sharkmelley

    Vanguard

  • *****
  • Posts: 2185
  • Joined: 19 Feb 2013

Posted 11 December 2015 - 06:13 PM

Very interesting effect - I've not seen anything quite like that.

 

I wonder exactly how the effect is caused and why is it red and not any other colour?

 

Mark



#6 jimegger

jimegger

    Vendor

  • *****
  • Vendors
  • topic starter
  • Posts: 1621
  • Joined: 22 Dec 2005
  • Loc: Palmer,Alaska

Posted 11 December 2015 - 06:43 PM

Beta Andromeda is fairly red and it is quite bright. I do not get this effect on any of the more faint objects with this camera. It does seem to indicate some magnified pixels doesn't it ? It does form a square shape which would indicate the grid formed by pixels. I'll have to try this again without any filter on and see if there is a difference.



#7 Jerry Lodriguss

Jerry Lodriguss

    Vendor

  • *****
  • Vendors
  • Posts: 7061
  • Joined: 19 Jul 2008
  • Loc: Voorhees, NJ

Posted 11 December 2015 - 07:52 PM

Here's the artifacts on a shot of Mars where I was trying to record Phobos and Deimos:

 

Jerry

Attached Thumbnails

  • Mars_Artifacts.jpg


#8 glend

glend

    Surveyor 1

  • *****
  • Posts: 1875
  • Joined: 04 Feb 2014

Posted 11 December 2015 - 10:21 PM

A new Borg Cube design?



#9 sharkmelley

sharkmelley

    Vanguard

  • *****
  • Posts: 2185
  • Joined: 19 Feb 2013

Posted 12 December 2015 - 02:05 AM

Here's the artifacts on a shot of Mars where I was trying to record Phobos and Deimos:

 

Jerry

 

It looks like the identical phenomenon.What combination of scope, camera and filters were you using?

 

Mark



#10 jambroseus

jambroseus

    Viking 1

  • -----
  • Posts: 842
  • Joined: 05 Apr 2007
  • Loc: Mount Laurel, New Jersey

Posted 12 December 2015 - 07:37 AM

Interesting artifact. I would guess that it is the mirrored back-reflection of the matrix off the filter.

 

John



#11 jimegger

jimegger

    Vendor

  • *****
  • Vendors
  • topic starter
  • Posts: 1621
  • Joined: 22 Dec 2005
  • Loc: Palmer,Alaska

Posted 12 December 2015 - 03:24 PM

You essentially got the same effect on your Mars shot Jerry. It even shows the fading out at the edges the same. What is interesting to me is how it appears to be an imitation of the pixel matrix but magnified and spread out as then magnified image over pixels that are actually quite distant from the bright objects image as if a mirror reflected it back onto the sensor.



#12 andysea

andysea

    Gemini

  • *****
  • Posts: 3443
  • Joined: 03 Sep 2010
  • Loc: Seattle, WA

Posted 12 December 2015 - 03:37 PM

Could it be caused by the micro lenses? I get weird artifacts that are caused but he micro lenses on bright stars.



#13 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • Posts: 8576
  • Joined: 12 Oct 2007

Posted 12 December 2015 - 04:29 PM

These kinds of images show up from time to time and yes I think it is related to the microlenses. Someone did a study of the images vs. wavelength with a ccd and confirmed that the pattern is wider with red than with blue - which means it is a diffraction effect. So each dot doesn't correspond to a single microlens - but to one of the diffraction orders like with a diffraction grating.

I don't know what to do to avoid them when they happen - but try dithering.

Frank

#14 jhayes_tucson

jhayes_tucson

    Fly Me to the Moon

  • *****
  • Posts: 6939
  • Joined: 26 Aug 2012
  • Loc: Bend, OR

Posted 12 December 2015 - 05:45 PM

Here are two possibilities:

1)  It may be due to stray reflections from the microns array.  Reflections from the high slope regions at the edge of each lens may be striking the surrounding lenses.  I haven't analyzed it, but it may be possible that the geometry favors red light being refracted to the sensor.  That's a complete guess!

 

2)  It could be diffraction from the sensor array being reflected from the filter.  The pattern seems so perfectly centered on the star that that seems unlikely.  The reflection could be coming from the surface of one of the filters within the cameras.  It seems that you would see other colors as well so I'm just not sure how that idea might work.

 

I've never seen this effect myself but it's pretty interesting.  Dithering might change things a little bit but if it's coming from the camera, it does't seem like it will change very much when you dither.

 

John



#15 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • Posts: 8576
  • Joined: 12 Oct 2007

Posted 12 December 2015 - 06:38 PM

For ccd's that show this effect, it is strongest in blue. I expect sensors to have different reflectivity depending on the sensor and the bayer array - and whatever other elements are reflecting it back to the sensor.

As I said above - I have seen a study where the pattern scaled with wavelength - but I can't find it now. It certainly has the look, to me, of a Fourier transform of an imperfect array.

I assume the star (or mars) spot is large on the scale of the pixels and that is why the star appears centered on the pattern. It would need to sample many pixels in order to generate a clean diffraction pattern. If the star spot were smaller on the scale of pixels, it would probably show less.

I think the star spot is sampling a large enough region of pixels that it generates a diffraction pattern on reflection - and that bounces off a nearby cover slip and back down to the sensor. Most things like this involve something very nearby - except for a pupil ghost, which may be in the reducing optics.

I don't think dithering would help much - but it is at least something to try. The pattern is probably well slaved to the star location. It's like a pupil ghost in that there is no way to remove it with a flat. I think it would need photoshop.

Frank
  • Jon Rista likes this

#16 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 12 December 2015 - 07:54 PM

Picture in post #1 is a full-resolution image.

 

The star with grid is about 250 x 250 pixel region, while there are about 13x13 shining zones.

If you consider the dark-black zones and shining zones, there are 26 x 26 regions.

 

Each region is about 10 x 10 (or 9 x 9 or even 8 x 8) pixel width and height.

 

This would be quite odd, if it's reflection directly from individual microlens.

 

However, I would suspect it is the light bouncing in between the AR layer (which is underneath the microlens and color filter layer, but right above the active sensing area of a pixel) and the Baader (Neodynium) Contact Booster filter.  Then individual microlens shapes the spread pattern (fairly similar with GLP with "star pattern" party filter in front of it.)

 

I am guessing if taking the Baader CB filter out, the pattern probably will go away.

 

Clear Skies!

 

ccs_hello



#17 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 12 December 2015 - 08:34 PM

A possible way to validate my speculation is to slightly move the Baader filter a bit away from the sensor.

This probably should have caused the grid pattern to be further spread out (or converged in.)

 

Clear Skies!

 

ccs_hello



#18 Jerry Lodriguss

Jerry Lodriguss

    Vendor

  • *****
  • Vendors
  • Posts: 7061
  • Joined: 19 Jul 2008
  • Loc: Voorhees, NJ

Posted 12 December 2015 - 11:15 PM

 

 

 

Mine was shot with a modified T2i with a replacement UV-IR filter and a C11 edge which has the concave corrector in the baffle tube.

 

No other filters were used.

 

Mars was quite large in my image compared to the pixels.  Note that the tiny white dots in the upper left and right of the image are actually stars.

 

Jerry



#19 jhayes_tucson

jhayes_tucson

    Fly Me to the Moon

  • *****
  • Posts: 6939
  • Joined: 26 Aug 2012
  • Loc: Bend, OR

Posted 13 December 2015 - 12:25 AM

For ccd's that show this effect, it is strongest in blue. I expect sensors to have different reflectivity depending on the sensor and the bayer array - and whatever other elements are reflecting it back to the sensor.

As I said above - I have seen a study where the pattern scaled with wavelength - but I can't find it now. It certainly has the look, to me, of a Fourier transform of an imperfect array.

I assume the star (or mars) spot is large on the scale of the pixels and that is why the star appears centered on the pattern. It would need to sample many pixels in order to generate a clean diffraction pattern. If the star spot were smaller on the scale of pixels, it would probably show less.

I think the star spot is sampling a large enough region of pixels that it generates a diffraction pattern on reflection - and that bounces off a nearby cover slip and back down to the sensor. Most things like this involve something very nearby - except for a pupil ghost, which may be in the reducing optics.

I don't think dithering would help much - but it is at least something to try. The pattern is probably well slaved to the star location. It's like a pupil ghost in that there is no way to remove it with a flat. I think it would need photoshop.

Frank

 

1) If the pattern scales with the wavelength, I'd expect to see lateral color in the pattern and I can't see that in either case.  The source is white and I just see pretty much red in the pattern.  How can the pattern scale with wavelength and still not show any lateral dispersion in the pattern?  Using some color filters in the beam path would easily confirm or toss out this idea.

 

2) I agree that it looks like an out of focus far field diffraction pattern; however, there are a couple of problems with that too.  First, I would expect to see dispersion increase with the order and that's not the case.  Second, Jim's example show pincushion distortion, which shouldn't happen with a diffraction pattern.  So, I'm having some trouble seeing this as a diffraction issue; though, I wouldn't totally rule it out.  Again, some color filters might be useful to try to look at whether diffraction is the cause.

 

3)  This looks to me more like a stray light issue.  Perhaps a stray is reflected from the top of each micro-lens from a nearby surface back onto the sensor.  Coatings could account for the color.  Angles within the illuminating beam might account for the distortion and small variations in reflectivity might account for the asymmetric pattern intensity.  If this is a stray reflection problem, the pattern should walk in position as the telescope is moved.  A number of images with the telescope position bumped a tiny amount between each exposure would tell us whether the pattern depends on the illumination position (or angle.)

 

It is certainly an interesting effect and the cause sure isn't immediately obvious.  I've shot some pretty bright stars with a modified 6D and I'm surprised that I've never seen this effect.  It may be that this is only a problem under certain "special" combinations of components and/or spacings.  The next time I'm out (like maybe by next spring when it clears again!!) I'll take a few shots of a really bright star to see if I can reproduce it.

 

John



#20 jimegger

jimegger

    Vendor

  • *****
  • Vendors
  • topic starter
  • Posts: 1621
  • Joined: 22 Dec 2005
  • Loc: Palmer,Alaska

Posted 14 December 2015 - 01:17 PM

Next time out I'll try a couple of things for an experiment with a control. For one, I'll shoot the same star without a filter on the camera. That will introduce more of the chromatic errors but will eliminate having anything between the objective and the sensor. For a second experiment I'll try using my 20 da with the same filter arrangement and see if the artifact appears and if it does if it is the same as the 6 D's appeared. I would imagine that the different sensors under the same conditions should give different artifact effects if it is the sensor arrangement. I thought about the filter reflections being responsible for the effect so removing the filter will resolve that issue possibly.



#21 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • Posts: 8576
  • Joined: 12 Oct 2007

Posted 14 December 2015 - 04:20 PM

It's unfortunate everything is being seen through the bayer array so you don't have a view of the raw pixel response, but I view this as diffraction and the pincushion shape is probably due to the square and abutting microlenses on the 6d.

If you can use a normal dslr lens and a long exposure on a point of light - that might not show the pattern if the spot is small on the sensor.

The baader contrast boost filter appears to be a visual filter - so it does not block the IR. I don't know what kind of mod your 6d has - but if it allows IR through, then this may largely be due to IR and you just need to block it.

If the pattern involves a reflection off something very local to the sensor like a cover slip - then changing filter distances or other things farther away wouldn't change the spacing of the dots. But I think it would be more noticeable with a wider spot on the sample and with slower f/number. So if you have a well focused 135 f/2 lens it may not show the pattern - but stopped down to f/16 and defocused to make a large spot - it might show the pattern. In a long enough exposure of a point source. And it may be much stronger in IR.

Frank

#22 Ron359

Ron359

    Apollo

  • *****
  • Posts: 1137
  • Joined: 21 Apr 2008
  • Loc: -105 +39

Posted 14 December 2015 - 05:06 PM

Next time out I'll try a couple of things for an experiment with a control. For one, I'll shoot the same star without a filter on the camera. That will introduce more of the chromatic errors but will eliminate having anything between the objective and the sensor. For a second experiment I'll try using my 20 da with the same filter arrangement and see if the artifact appears and if it does if it is the same as the 6 D's appeared. I would imagine that the different sensors under the same conditions should give different artifact effects if it is the sensor arrangement. I thought about the filter reflections being responsible for the effect so removing the filter will resolve that issue possibly.

When you do with filter off and on, you should also try rotating the camera and see if the rectangle rotates too.  That would nail it to inside the camera and perhaps a reflection off the filter.  Ron



#23 sharkmelley

sharkmelley

    Vanguard

  • *****
  • Posts: 2185
  • Joined: 19 Feb 2013

Posted 14 December 2015 - 05:28 PM

It seems this is a common problem and affects normal daylight photography with digital cameras and normal lenses. 

 

It's the red dot flare issue:

https://photographyl...dot-flare-issue

 

Other example images of the same phenomenon:

http://www.dpreview....s/post/53005750

http://www.dpreview....s/post/56018689

http://www.dpreview....s/post/51925256

 

No need for astronomical filters.  Maybe the surfaces of the lens (or in the case of telescopes, the focal reducer) are focusing an image of the brightly lit sensor back onto the sensor.

 

Mark


Edited by sharkmelley, 14 December 2015 - 05:31 PM.


#24 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • Posts: 8576
  • Joined: 12 Oct 2007

Posted 15 December 2015 - 02:56 AM

I think it's important to note that the overall pattern happens at a similar scale *on the sensor* whether using a dslr lens or a telescope with filters somewhere in it. That makes me think it is local to the sensor itself and involves a bounce off the sensor to something very nearby and back to the sensor.

I don't agree with some of the interpretations in those dslr threads because most did not consider the role of diffraction. The effect shows more clearly when you have a reasonably high f/ratio and the beam hitting the sensor is somewhat wide and parallel - like a laser spot. The microlenses would then act like a cd or star filter on a laser pointer - and create a spray of beams upwards that spread out going up and then come back down after hitting some particular interface in the sensor structure - which I assume preferentially reflects red.

The reports do seem to say it happens more with high f/numbers as I expected - and the particular pincushiony effect isn't normal pincushion because it isn't purely radial - it has a squarish shape with corners - exactly like the layout of the microlenses.

So - it all seems consistent with a diffraction pattern going up and back down after hitting something very near the sensor. But I could be wrong - if there is conflicting evidence and a better explanation.

Frank

#25 jhayes_tucson

jhayes_tucson

    Fly Me to the Moon

  • *****
  • Posts: 6939
  • Joined: 26 Aug 2012
  • Loc: Bend, OR

Posted 15 December 2015 - 10:14 AM

Hmm, is this why Cannon added the long wavelength filter in the first place?  It would be interesting to swap a modified and an unmodified camera to see if the pattern appears only in the mod'ed camera.

 

Jerry's image appears very symmetric and "diffraction like"; whereas, Jim's image shows distortion with a lot of asymmetry, which looks to me more like a stray issue.  Either way, there has to be a reflection to get the light back to the sensor.  I think that the surface causing the reflection is most likely in the camera and relatively close to the sensor.

 

Is there any way that this could be due to the anti-aliasing filter--say through multiple reflections?  The anti-aliasing filter operates on a pretty small scale so it's not immediately obvious how it could produce this kind of pattern.

 

If the weather ever clears, I'll try to reproduce it with my system.  With the moon coming back up, the clouds should part any day now...  ;-)))   I'm traveling right now but I have some ideas on how to simulate this effect in my shop so I'll try it when I get back home.

 

John




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics