color cast DMR/M8/M9

[chrismuc]

Advertisement (gone after registration)

Using the Leica/Schneider 28f2.8 PC lens on a DMR shows similar color cast towards the edges like on M8/M9. So the cast seems to be related directly to the Kodak CCD and the micro lens structure and not (only) to the out-of-axis rays of light of the rangefinder cameras. The Leica PC lens does not show any color cast on a 5D2 (there even full frame).

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!

[lct]

...The Leica PC lens does not show any color cast on a 5D2 (there even full frame).

Interesting indeed. Would you mind to post a pic with this combo?

[masjah]

As a film only user, lurking on this Forum, can I ask three questions please:

 

(1) Were you shifting, and if so, how much shift were you using?

(2) What aperture were you using?

(3) (Apologies for asking, but I've done this myself!) Did you remember the manual stop-down?

[rsh]

With a deep blue sky, you may be getting a slight blue reflection on the white surface. I has happened to me on both film and digital in the past.

[pop]

In the copy of the image you placed here, I see not only a bluish cast in the top right hand corner. I also see fringes on edges. The shadow on the wall of the top balcony shows a distinctly blue edge as do some other edges. Other edges appear to be on the orange side, as far as I can make out. I can see it only in the top right hand edge of the image. I have no idea as to the possible causes.

[adan]

Using a PC lens with the lens shifted is equivalent to using a much wider lens (say ~19mm) and then cropping for just a part of the image near the edge. One is using the edge of the image circle instead of the middle.

 

Which can lead to 3 effects:

 

1) vignetting - which isn't part of the OP's question, but may be influencing the amount and visibility of color drift.

 

2) angle-of-incidence issues - cyan drift, red edges, etc. etc. In the posted sample, the light near the top of the image is passing through the DMR's IR cover filter at a shallower angle, and thus some visible red light is being filtered out (if it is behaving anything like an M8 and WA lens with IR filter attached) - leading to cyan drift in the white building.

 

This kind of color shift with shift/tilt lenses is common on digital - even MF backs.

 

3) Chromatic aberrations - red/cyan or blue/yellow fringes on sharp contrasty edges. Which is a completely separate issue from color drift, although tied to the general "wideness" of the lens. CA can appear in either digital or film images; AoI color drift is strictly a digital phenomenon, due to the use of the IR filtering needed for silicon sensors. Again, at the edges of the image circle, there will be more CA than in the center, so shifting the lens will bring it into the image area.

 

(Fortunately, with WA lenses, the CA is usually of the lateral kind, which is the easiest to correct for in digital processing. However, because the lens is "off-center" the CA will also be off-center - "I can see it only in the top right hand edge of the image" - so it requires Photoshop work rather than the quick-and-easy lens corrections available in raw developers.)

 

Now - as to why the DMR and 5D should show different results: First, I'd like to see the difference. But it could well be that the Canon and DMR use different IR filters (absorption vs. interference). It is also true that with the DMR, the lack of an AA filter can lead to moire effects along sharp slanted edges, leading to orange or cyan fringes that look similar to CA - or that the 5D's AA filter blurs those edges just enough to partially disguise any color fringes.

 

Finally - vignetting. If you tint a white wall slightly cyan and also darken it, it will appear more heavily colored than just a color tint alone. In the original sample, the top part of the semi-shaded white wall is not only bluer than the balcony just above the bushes, it is also 20% darker (brightness level 152 vs. 190)

[mjh]

Advertisement (gone after registration)

Using the Leica/Schneider 28f2.8 PC lens on a DMR shows similar color cast towards the edges like on M8/M9. So the cast seems to be related directly to the Kodak CCD and the micro lens structure and not (only) to the out-of-axis rays of light of the rangefinder cameras. The Leica PC lens does not show any color cast on a 5D2 (there even full frame).

Just one addition to Andy’s explanations. This is all about the incident angles of the light hitting the sensor. The distribution of incident angles solely depends on the position (and size) of the exit pupil and the sensor size, so for a given lens and a given sensor size they are the same – differences in the flange distances are immaterial. For this reason one should not expect any differences between rangefinder cameras and DSLRs.

 

On the other hand there are differences in the amount of microlens shifting. The arrangement of microlenses is optimized for a certain typical position of the exit pupil that differs between cameras. There are also differences in the IR-blocking filters shielding the sensor – Leica generally prefers absorption filters while Canon uses hybrid filters – a sandwich of absorption and dichroic filters.

[robsteve]

As Michael noted above, it is a problem when using shift lenses with sensors that have micro-lenses. The Flexcolor software has the solution for it, but sadly, Imakon/Hasselblad does not enable that function for the DMR.

 

It is common in most medium format digital back capture software. I forget what it is called, but may be called shading. After you set your shift, you shoot a white card or through a piece of white plexi to get a base file for the software to correct the rest of the images. A similar process can be done in photoshop.

 

Robert