Purple flare: Causes and remedies

Started Mar 1, 2013 | Discussions thread
Anders W
Forum ProPosts: 17,500Gear list
Like?
Re: Excellent!
In reply to kenw, Mar 3, 2013

kenw wrote:

Anders W wrote:

I have now tried to make the first of the two types of flare, that from light sources just outside the frame, extend diagonally rather than vertically by tilting the camera. But ... this simply doesn't work. The purple streak always appears vertically within the frame even with the camera tilted right or left. So in this respect, the two kinds of flare behave in exactly the same column-row-like manner.

Thanks for checking that. In your first images it hints that this would happen, the streaks don't point at the optical center but are vertical. Your further test makes it clearer.

So now what is strange is how does it extend so far? I completely understand off-axis light smearing into adjacent pixels and that causing the color shift. A few pixels is all that is necessary for the "purple effect". But this is demonstrating a vertical and horizontal constraint of hundreds of pixels even in a diagonally presented source. Seems remarkable for that to be happening between the CFA and the sensor... Of course that isn't the only way it could happen. AA filter or micro lens pattern could cause a vertical/horizontal streak in white light before the CFA but that light then strikes at a shallow angle causing the purple as you described.

Not sure how you reason when you say that this demonstrates a vertical and horizontal constraints of hundreds of pixels. Could you please elaborate a bit.

Just like you, I think of the pollutive streams as local (adjacent pixels). But as far as I can see, this wouldn't prevent the streaks from extending vertically over a pretty long distance but being blocked diagonally if pollution in that direction is impossible. Note that the pollution doesn't have an impact on color only but also on luminance due to the different amplification of the three channels. Note also that my attempt to eliminate the color-channel imbalance in the third image of the OP is not completely successful. The blue channel still trails red and green a bit, which means the streak remains a bit brighter than its surroundings and has a blueish tint (which is more apparent if I look at what LR tells me about the displayed strength of the three channels than if I judge the color subjectively, in which case the blueish tint isn't very apparent).

Anyway, interesting and perplexing. Would really like to experiment myself but I barely have the time to read this thread!

Again - thanks for taking the time and effort to come up with this and your excellent explanations.

-- hide signature --

Ken W
See profile for equipment list

 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