The Panasonic 20mm f/1.7 has horrific flare characteristics

texinwien wrote:

Tom Axford wrote:

texinwien wrote:

Tom, here's the same image with exposure boosted to 1.5, contrast dialed down to -33, luminance NR set to 0 and purple hue set to 0. The purpose of this image is to help you see where the banding occurs, after which I think it will be easier for you to pick it out in the image I shared earlier (which, as I stated elsewhere, I had applied several tricks to in order to try to minimize the appearance of the pattern noise).

Tex,

Thank you for that version.

You bet!

I can now see the bands, but only just (it takes a bit of effort)! I think perhaps my eyes are less sensitive than yours to colour variations in that part of the colour spectrum because I certainly wouldn't have noticed these bands unless I was specifically looking for them (although I am not colour blind in the usual sense of the term).

Maybe variations in eyesight account for at least some of the disagreement on the significance of banding?

I bet you're right. It's important for us all to keep in mind that there's not really an objectively correct visual interpretation of any given image. Each person will 'see' (in his mind's eye) a given image differently than everyone else, and I seem to be pretty sensitive to these blue/purple/magenta variations (and I find them rather distracting).

What I will say is that I am calibrating my expectations based on how my other 8 m43 lenses perform on my E-M5, and, in my experience, the banding tends to be obviously (to me) more prominent when using the E-M5 with the 20mm attached than with the same camera and any of my other lenses attached, all else remaining equal.

What I also find interesting is the variable nature of the banding, from shot-to-shot. To me it looks very much like an interference pattern that that varies in intensity and shape quickly, from shot-to-shot. AFAIK, no one's figured out the source of the interference (or, if they have figured it out, they don't seem to have expended much effort in publicizing their findings).

I think the best guess we have with regard to the source of the interference is still the one worked out by Detail Man and me in this lengthy thread about the issue

http://www.dpreview.com/forums/post/50092818

i.e., a noisy power supply/voltage regulator on the lens PCB. The forum poster who has gone to the greatest length in trying to test various hypotheses about the issue calls himself tsi. Here are some of the posts where he details his results:

http://www.dpreview.com/forums/post/51951047

http://www.dpreview.com/forums/post/51978743

http://www.dpreview.com/forums/post/51977740

What additionally seems pretty clear by now is that there is something about the (presumably Sony-designed) sensor used in the E-M5, E-PM2, E-PL5, E-P5, E-M10, and GH3 that makes it particularly sensitive to this interference. AFAIK, all other MFT sensors than that particular one are unaffected.