Sensor reflections from eos-m

[Cinto]

I've heard that some of the M4/3rd's and Sony Nex(Alpha) suffer from reflections from the sensor. I've had an Eos-M for a couple of months and have seen my first example of it from the M. I'm just wondering if anyone else has come across it, and if it will be a problem across all the short flange digital systems?
Cheers,
Cinto

 

[ajfotofilmagem]

What you see in this picture does not seem to reflect the sensor, but "lens flare". Each specific lens has varying degrees of resistance to flare when shooting into light against. Use a lens hood can prevent this phenomenon, but when the sun is hitting the lens at an angle, it will still happen.

 

[BL]

Cinto,

What lens are you using? Is there a filter on it?

 

[Cinto]

I was using the 40mm pancake with no filter. I don't believe it is lens flair, I've seen tons of lens flare using this on my 5D 2 and never had the grid like pattern. It seems similar to this
http://photographylife.com/fuji-x-trans-flare-ghosting-issue
and this
http://www.fujixseries.com/discussion/comment/50181
and this
http://www.eoshd.com/comments/topic/4306-olympus-phase-detect-strips-showing-up-in-e-m1-video/

 

[jrista]

I was using the 40mm pancake with no filter. I don't believe it is lens flair, I've seen tons of lens flare using this on my 5D 2 and never had the grid like pattern. It seems similar to this
http://photographylife.com/fuji-x-trans-flare-ghosting-issue
and this
http://www.fujixseries.com/discussion/comment/50181
and this
http://www.eoshd.com/comments/topic/4306-olympus-phase-detect-strips-showing-up-in-e-m1-video/

The first two links are definitely lens flare, starburst, and ghosting.

Lens flare is the loss of contrast, NOT RAYS, due to a high angle of incidence off a bright off-axis light source (like the sun).

Rays are just that, rays or starburst, usually caused by the aperture. The number of rays in a starburst are usually dependent on the number of aperture blades, they can very quite a bit.

Ghosting is reflection off of lens elements, again caused by intense off-axis light. You can differentiate ghosting from sensor reflection, because ghosts have an aperture shape.

Sensor reflection is rather specific. It will usually be rectangular in shape, roughly the same dimensions as the sensor. It may be harsh, but more frequently it presents with softened "feathered" edges, so it can sometimes be hard to pinpoint as sensor reflection. Sensor reflection is also usually around the center of the image, not the periphery.

Given your sample image, that most definitely looks like sun flare, not sensor reflection.

 

[Drizzt321]

I was using the 40mm pancake with no filter. I don't believe it is lens flair, I've seen tons of lens flare using this on my 5D 2 and never had the grid like pattern. It seems similar to this
http://photographylife.com/fuji-x-trans-flare-ghosting-issue
and this
http://www.fujixseries.com/discussion/comment/50181
and this
http://www.eoshd.com/comments/topic/4306-olympus-phase-detect-strips-showing-up-in-e-m1-video/

The first two links are definitely lens flare, starburst, and ghosting.

Lens flare is the loss of contrast, NOT RAYS, due to a high angle of incidence off a bright off-axis light source (like the sun).

Rays are just that, rays or starburst, usually caused by the aperture. The number of rays in a starburst are usually dependent on the number of aperture blades, they can very quite a bit.

Ghosting is reflection off of lens elements, again caused by intense off-axis light. You can differentiate ghosting from sensor reflection, because ghosts have an aperture shape.

Sensor reflection is rather specific. It will usually be rectangular in shape, roughly the same dimensions as the sensor. It may be harsh, but more frequently it presents with softened "feathered" edges, so it can sometimes be hard to pinpoint as sensor reflection. Sensor reflection is also usually around the center of the image, not the periphery.

Given your sample image, that most definitely looks like sun flare, not sensor reflection.

It looks like sun flare, but there's that very interesting pattern in the exact center of the image, but only on top of where the dog is (eg where it's dark, so little incoming light). Maybe an odd reflection off of the adapter? Maybe off of the sensor? I dunno.

 

[Cinto]

I think I have to disagree, heres a 100% crop of what I'm talking about. That's not lens flare. Also based on the color of the flare it really looks like sensor refections to me. It's quite apparent in a lot of back lit photos, I've owned a 10D, 40D, XTi, 5d Mark 2, T3i, and a few more, I've never seen anything like this except for photos from the M.

 

[jrista]

I think I have to disagree, heres a 100% crop of what I'm talking about. That's not lens flare. Also based on the color of the flare it really looks like sensor refections to me. It's quite apparent in a lot of back lit photos, I've owned a 10D, 40D, XTi, 5d Mark 2, T3i, and a few more, I've never seen anything like this except for photos from the M.

That pattern looks like something else, for sure. Can't say what that is. The purple bit still looks like flare from the sun, which I gather is directly behind the dog?

Only thing I can think is that the registration distance of mirrorless cameras is so small that maybe flare is exacerbated by slight reflection off the sensor. The registration distance of DSLR's is much larger, so any kind of reflection off the sensor will fall of to almost nothing by the time it reaches the lens again. With EOS-M, or pretty much any other mirrorless camera, the registration distance is only a couple/few millimeters.

I would expect true sensor reflection to be rectangular (and that funky pattern in your images seems to be), rather than an extension of flare (which is what the purple hue looks like in your images)...but perhaps there is something else going on between the back lens element and the sensor due to the extremely short distances.

Guess this is one of the downfalls of mirrorless designs, as the phenomena certainly doesn't seem to be limited to the EOS-M. Maybe the use of nanocoating on the back and inner lens elements would mitigate the problem.

 

[Drizzt321]

I think I have to disagree, heres a 100% crop of what I'm talking about. That's not lens flare. Also based on the color of the flare it really looks like sensor refections to me. It's quite apparent in a lot of back lit photos, I've owned a 10D, 40D, XTi, 5d Mark 2, T3i, and a few more, I've never seen anything like this except for photos from the M.

That pattern looks like something else, for sure. Can't say what that is. The purple bit still looks like flare from the sun, which I gather is directly behind the dog?

Only thing I can think is that the registration distance of mirrorless cameras is so small that maybe flare is exacerbated by slight reflection off the sensor. The registration distance of DSLR's is much larger, so any kind of reflection off the sensor will fall of to almost nothing by the time it reaches the lens again. With EOS-M, or pretty much any other mirrorless camera, the registration distance is only a couple/few millimeters.

I would expect true sensor reflection to be rectangular (and that funky pattern in your images seems to be), rather than an extension of flare (which is what the purple hue looks like in your images)...but perhaps there is something else going on between the back lens element and the sensor due to the extremely short distances.

Guess this is one of the downfalls of mirrorless designs, as the phenomena certainly doesn't seem to be limited to the EOS-M. Maybe the use of nanocoating on the back and inner lens elements would mitigate the problem.

Agreed, the purple seems to be flare. The pattern? No clue.

 

[Cinto]

Too bad, not extremely useful for backlighting. The flare, or reflections, happen quite often in back lit photos, although the grid like pattern occurs less often. Notice how the purple color is very wide spread.

 

[jrista]

Too bad, not extremely useful for backlighting. The flare, or reflections, happen quite often in back lit photos, although the grid like pattern occurs less often. Notice how the purple color is very wide spread.

I suspect the flare is due to the cheap quality of the 40mm pancake. RELATIVELY speaking, it is a decent lens, but in the grand scheme of things, it doesn't come close to say the kind of Zeiss-made pancakes used on Leica rangefinders (which also have a short registration distance, but few of these kinds of problems.)

The quality of the lens is apparently going to be even more important with mirrorless. To be more specific, the quality of the anti-reflection coatings will be of particular importance. The kind of basic and cheaper multicoating used on the likes of Canon's 40mm pancake is simply not going to be able to combat the kind of flare issues that seem prominent in cameras with very short registration distances. A higher quality multicoating, or better a nanocoating, would really be necessary to limit this problem. I also suspect that if sensor reflection is common on mirrorless cameras, some kind of anti-reflection coating will need to be added to the filter stack that sits on top of the sensor.

In other words...nothing you yourself can really do about the problem (outside of maybe buying better lenses.) You'll have to wait for manufacturers to actively seek to solve the problem.

 

[Orangutan]

Do you have the raw file or just JPEG? If you have raw, try processing it with different software. E.g., if you used LightRoom, try DPP and see if you get the same effect. Also, per JRista's description of sensor reflection, try looking at the full-res image and see if you can find the outlines of the speckling.

 

[Cinto]

It was a RAW, I tried it in Capture1, same thing.

 

[Mt Spokane Photography]

It might be worth trying it with a non pancake lens, the design of a pancake might be a big factor.

 

[raptor3x]

Only thing I can think is that the registration distance of mirrorless cameras is so small that maybe flare is exacerbated by slight reflection off the sensor. The registration distance of DSLR's is much larger, so any kind of reflection off the sensor will fall of to almost nothing by the time it reaches the lens again. With EOS-M, or pretty much any other mirrorless camera, the registration distance is only a couple/few millimeters.

That was my first thought, except he's using a 40mm pancake lens so the registration distance will be exactly the same as if it were on an EF mount DSLR.