Purple flare: Causes and remedies

Started Mar 1, 2013 | Discussions thread
Anders W
Forum ProPosts: 18,346Gear list
Like?
Re: Purple flare: Causes and remedies
In reply to micksh6, Mar 3, 2013

micksh6 wrote:

Anders W wrote:

micksh6 wrote:

Anders W wrote:

According to the broschure descring the new Leica M (p. 22), from which I borrow the drawing shown below of the sensor architecture of a "standard CMOS sensor" (not the one used by Leica M), "rays of incoming light at large angles of incidence can fail to reach the photodiode of the corresponding pixel and reach only the adjacent pixel". One possibility is the one shown by the drawing, where a photon passes directly through the filter of one pixel and then on to the photodiode of another. Another possibility, not shown in this drawing, is that a photon is reflected, after having passed the filter of one pixel, in such a way as to end up at the photodiode of another pixel.

Thanks a lot for the test and writing.

Glad you appreciated it.

I have just one comment regarding microlenses and CFA effect. Take a look at the picture at the bottom of my post. I took it from mu-43.com forum and it was shot with OM-D and 7-14mm wide open (aperture blades didn't impact the flare).

I'm sure the purple flare you see is a result of reflection between sensor and rear lens element. You can see the pattern of the blinds appearing on the wall. The pattern is smaller than real blinds pattern probably because the rear element is a bit concave? (I don't have 7-14mm)

Based on this I think the only thing we can tell for sure is that OM-D sensor is too reflective under certain conditions such as certain light angles and wavelength(s). The resulting reflected wavelength appears purple in sensor output. Reflectivity and coating of lens rear element most likely also plays some role.

How do you arrive at the conclusion that the flare results from a reflection off the sensor, hitting the rear element and then bouncing back to the sensor again. As far as I can see, this is not a necessary ingredient, and I don't it particularly likely (although I am not in a position to conclusively rule it out).

I arrived at this conclusion because it seemed the most logical explanation of the image I showed below.
If someone has better explanation of why blind pattern is shown as reflected on the left side I would be interested to hear it.

Also there are many other image samples showing multiple flare orbs, like in this post from this thread:
http://forums.dpreview.com/forums/post/50958176
If you think this is not a result of multiple reflections between sensor and rear lens element I would like to know a better theory explaining this.

Ordinary internal reflections within the lens. If you look at further examples from the same poster as the one you linked to here

http://forums.dpreview.com/forums/post/50958291

in this case using the GF1 along with the E-M5, you can see the same type of reflections from both cameras. The only difference is that with the E-M5, some of the reflections are intensely purple.

Essentially the same result has been reached in a large number of prior comparisons. The main difference between the E-M5 and other cameras is not that the E-M5 show reflections/flare that the others lack. It's that the reflections become intensely purple, and therefore far more noticeable (they might be barely visible otherwise), with the E-M5. Luminance can also be affected in addition to color. Since the red and blue channels are amplified far more than green (about twice as much in ordinary daylight), the average of the three colors also increases when green light escapes to the blue and red channels.

Where the reflection on sensor happens exactly is not clear. But it is possible that purple cast appears because reflected light passes through CFA and the CFA affects reflection in some irregular way. I would think purple color is a result of light passing through both red and blue filters, omitting green. In this case the light would have to be close to diagonal because red and blue only have common corners. But, since we also see horizontal and vertical flares the issue is probably more complicated.

If different wavelengths would behave slightly differently when it comes to reflection (as they are known to do; why is the sky blue?), then blue and red would likely behave most differently with green inbetween. Note also that the evidence we have so far suggests that channel pollution is more likely (perhaps only possible) vertically and horizontally, not diagonally.

OM-D, 7-14mm @F4

Original thread where I took the picture from: http://www.mu-43.com/f38/panasonic-7-14-gh3-39958/index2.html

PS. BTW, "color channel pollution" is called color cross-talk (sometimes color bleeding) in some imaging industries.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow