[hoanpham]

I have an insignificant trick to share regarding PF:

Most pentax and takumar lenses do have PF, especially in highlight areas and dark background, like metal/chrome reflections.
This bike is shoot with Smc Tak 85/1.8 at f1.8. Distance and angle are selected to cover the whole bike with hyperfocal.

I shoot multiple frames this bike, with different focus point: rear part, mid/engine part and front/front wheel part.
Result from these photos reveals the in-focus areas have least PF (or not noticable).
Based on hyperfocal, one might think that the photo with focua point in the middle/engine part would be the best/sharpest.
But due to PF from the shinny metal reflections, the photo with focus point in the front wheel gives least PF (note: f1.8 is not the best property of this lens)

Then in post processing/LR, i adjust the purple color saturation to remove the remaining of PF.
Since the photo does not contain any significant purple color, it's safe.
And here is the result:


Same trick on this photo - focus on the trumpet, not the face - I also have multiple frame with different focus point.
Smc Tak 85/1.8 at f1.8. No visible PF, and therefore no PP:


Now, grab your lens and have fun

[Ameiji]

Any ideas why PF is more obvious with some lenses while not so with the others? For example DA40 on K10D gives me more PF in comparison to DA18-55.

[Asha]

There may be several reasons, but the first that comes to mind for me is that more contrasty lenses will be more susceptible. That is to say, lenses that produce more contrast (a la MTF) for highly detailed objects could end up saturating a detector element more easily.

Any ideas why PF is more obvious with some lenses while not so with the others? For example DA40 on K10D gives me more PF in comparison to DA18-55.

[Ameiji]

There may be several reasons, but the first that comes to mind for me is that more contrasty lenses will be more susceptible. That is to say, lenses that produce more contrast (a la MTF) for highly detailed objects could end up saturating a detector element more easily.

You think it's a saturation issue? I was under impression that the PF occurs in sensor's upper layer, on it's micro-lenses.

[Lowell Goudge]

There may be several reasons, but the first that comes to mind for me is that more contrasty lenses will be more susceptible. That is to say, lenses that produce more contrast (a la MTF) for highly detailed objects could end up saturating a detector element more easily.

You think it's a saturation issue? I was under impression that the PF occurs in sensor's upper layer, on it's micro-lenses.

Somewhere, and I can't recall where now, or find where, I read that PF was also a saturation issue, not an issue with the micro lenses, but I can';t find that definition any more. My experiences seem to be much more related to points that are at the top of the histogram as opposed to just simply a light on dark image.

much of where i see it in the motor cycle is directly adjacent to areas where the greyscale value is 255, which somewhat supports the saturation theory

[stevebrot]

Any ideas why PF is more obvious with some lenses while not so with the others? For example DA40 on K10D gives me more PF in comparison to DA18-55.

That is the magic question. Theories:

• Incident angle of light to the sensor microlens array
• Diameter of rear element
• Constructive interference at blue/purple wavelengths

One thing that is clear is that PF is as old as the consumer market for digital cameras. When I purchased my Canon G2, way back when, resistance to PF was part of the test suite for all digicams. Excessive PF was considered to be a sensor design flaw.

Why good technical information regarding PF is hard to come by is anybody's guess. It may be that there is a lot of proprietary technology at work here and the sensor/lens makers are keeping mum for that reason.


Steve

[Asha]

The microlenses are photon collectors for each of the sensor element charge buckets. So, PF could indirectly be due to the microlenses, since they would collect the necessary signal to induce too much charge in the bucket. Having said that, no lens is lossless, so chances are the microlenses actually prevent some signal from getting in.