Purple Reflection

[Beyder28]

Advertisement (gone after registration)

I just got my used 50mm summicron and took my first few photos with it. One of them had this strange purple reflection off the silver necklace. Can anyone tell me what this is from? If you zoom in a bit, you can see it more clearly.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Lindolfi]

Longitudinal chromatic aberration of the lens enhanced by the lack of an anti-aliasing filter in the M9. Try to photograph reflection of the sky in a lake with tiny ripples from the wind: a field of purple dots!

 

Also see here: http://www.l-camera-forum.com/leica-forum/customer-forum/170947-convince-me-i-need-noctilux.html#post1637831

[mjh]

Two words: purple fringing.

[Beyder28]

so is this normal? Only way to prevent is with filter?

[jaapv]

No, filters make no difference. Best to use C1 as raw converter and use the defringing slider. ACR works decently as well. For the rest any technique in PP. There are several.

[Beyder28]

I actually tried going with C1 as my raw converter over what I currently use which is Aperture and looking at the converted files side by side, Aperture is miles better in handling the DNG files in terms of colors and WB so I backed away from C1.

 

Is there a way to fix this within Aperture?

[artspraken]

Advertisement (gone after registration)

Chromatic aberration.

 

 

Sent from my iPhone using Tapatalk

[ho_co]

Chromatic aberration would usually show one color on one side, the complementary color on the other side.

 

Looks like purple fringing, which happens when a greatly overexposed area borders on one with more normal exposure.

 

 

Only thing I can suggest after the fact is to desaturate those areas.

[Lindolfi]

Chromatic aberration comes in two flavors: lateral and longitudinal. The lateral type shows itself in achromatic lenses as purple on one side and cyan on the other in high contrast parts of the image and never in the exact center of the image. Longitudinal chromatic aberration in achromatic lenses appears as a purple ring around bright objects or a cyan ring, depending whether your are slightly in front or behind the focal point. This discolouration can become enhanced by blooming (the carry-over of over-exposed colour wells to neighbouring wells) and by a lack of an anti aliasing filter in the case of very fine details in the image. This effect is more strong in the purple type (red plus blue), thus the term "purple fringing".

[jaapv]

I actually tried going with C1 as my raw converter over what I currently use which is Aperture and looking at the converted files side by side, Aperture is miles better in handling the DNG files in terms of colors and WB so I backed away from C1.

 

Is there a way to fix this within Aperture?

Now that surprises me, as the best feature of C1 is the color handling, especially in Pro.Maybe you used the wrong profiles. Aperture is about the worst raw convertor for purple fringing ( and color aliasing for that matter:(), if you use it you will have to correct it in postprocessing. In photoshop it is very easy to use the color replacement brush.

[denoir]

And on a practical note, to get rid of the longitudinal CA you just stop down the lens a bit. Lateral CA can't be fixed by stopping down, but modern Leica lenses usually have very little lateral CA to speak of.

 

Here's an example from a set of comparisons I made between the 90 Cron V3 and the 90 Cron APO ASPH:

 

Full image:

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

f/2 crop:

 

f/5.6 crop:

[jaapv]

Chromatic aberration.

 

 

Sent from my iPhone using Tapatalk

No. this is an artifact in raw conversion for high-contrast edges.

[denoir]

No, Jaap, it has nothing to do with RAW conversion. It is longitudinal CA and it goes away if you stop down the lens. (see the example in my previous post)

 

There's a good article on the topic here:

Chromatic aberrations

[ho_co]

I think I've always seen chromatic aberration and purple fringing discussed as separate phenomena.

 

Are we now saying that purple fringing is always caused by longitudinal chromatic aberration? There is never purple fringing unless longitudinal chromatic aberration is present?

 

Then shouldn't there be some sign of longitudinal chromatic aberration in the rest of the image in question?

 

Just a question--what's the difference between saying that a) overexposure brings out the chromatic aberration and blooming takes over from there; and overexposure brings out the blooming which is the symptom of purple fringing?

 

[denoir]

LoCA can be seen in the bokeh as well (green rings around highlights). Sensor blooming is something different:

Concepts in Digital Imaging - CCD Saturation and Blooming

 

It can amplify the effect by increasing the contrast in the region, but the purple part is not related to the sensor. Purple fringing is detectable on film as well although it is generally more visible on digital. One of the reasons for that is that film has a certain thickness where the rays can converge (a digital sensor is much thinner) and the LoCA is less evident.

 

So to answer your question: No LoCA, no PF. Sensor blooming can make it look worse but it's not the cause of it.

 

This can be easily shown as in the example above by simply stopping down the lens, which removes the LoCA. The sensor blooming is still there, but no purple fringing.

[adan]

There are multiple things that can cause purple fringes. Sometimes more than one can happen at the same time. This can lead to confusion.

 

1. Overexposure of bright highlights (Violet wavelengths are the most energetic, thus spread further than green or yellow, thus the fringes tend to violet)

 

2. Lens CA, either lat. or long.

 

3. Moire artifacts (fine details being recorded by only a red or blue pixel)

 

4. (seems to me there is a fourth, but I can't recall what right now).

 

I'd say Beyder28's particular example involves 1 and 3, and maybe 2 as well (but not necessarily). The highlights are blown out and the blown area also contains fine detail (necklace engraving).

 

Denoir's examples involve 1 and 2.

 

To fix in Aperture (one method) - find the "color noise" control and increase it. That should blur colors (including the purple of the fringing) slightly into the surroundings, and the surrounding colors into the purple, without affecting luminance (the sharp underlying image). In Photoshop I use a "median" noise filter faded to color only for the same effect.

 

Beyder22 - the "AA filter" mentioned is normally built right into digital sensors as part of the cover glass (not a filter on the lens), to blur the image slightly precisely to prevent any one fine detail or edge falling on only one pixel.

 

Leica's digital system cameras (M, DMR back, S2) leave out this filter to preserve overall sharpness - but the downside is more potential for these effects.

[denoir]

There are multiple things that can cause purple fringes.

I'd say Beyder28's particular example involves 1 and 3, and maybe 2 as well (but not necessarily). The highlights are blown out and the blown area also contains fine detail (necklace engraving).

 

Very much necessarily. Had he stopped down, the highlights would have still been overexposed but the PF would have been gone. CA is always the cause of PF - blown highlights and sensor blooming can make it appear worse but it is never the cause of it.

 

---

 

 

Anyway, here's an example of the more unusual lateral CA purple fringing. This typically happens at the edge of the image (100% edge crop):

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

This is at f/8 and the left side image is the pre-ASPH 90 Cron and the right side of the image is the 90 Cron APO ASPH. Here we can see the benefits of the apochromatic corrections of the newer lens.

[adan]

"here's an example of the more unusual lateral CA purple fringing"

 

Actually, I run into lateral CA far more often than longitudinal. But I mostly shoot wideangles, where lat is more common and long less common. My only lenses that show a lot of long. CA are my 75 Summilux and a V.2 R 250 f/4 (which I got rid of). Both non-APO, of course.

[Advertisement]

Hallo,

Look here: Leica M

[denoir]

True, the "more unusual' is not correct. I was incorrectly generalizing from my own lenses. The wide angles that I have are modern and well corrected for lateral CA (with the odd exception), but that is not the norm. LoCA on the other hand is something that I need to deal with on occasion as they are typical of all fast lenses.

 

As for APO vs non-APO, I have yet to see any difference in LoCA. Both my 75 APO-Cron & 90 APO-Cron are as bad as my non-APO pre-ASPH 90 Cron. There's a big improvement in lateral CA though - the APOs are basically completely free of it.

[giordano]

As for APO vs non-APO, I have yet to see any difference in LoCA. Both my 75 APO-Cron & 90 APO-Cron are as bad as my non-APO pre-ASPH 90 Cron. There's a big improvement in lateral CA though - the APOs are basically completely free of it.

 

That's interesting, given that the usual definition of "apochromatic" is bringing three wavelengths of light into focus on the same plane - i.e. it's defined in terms of longitudinal rather than lateral CA.