Poor results with filters?? (M246)

[Big_g]

Advertisement (gone after registration)

It is definitely a problem when you shoot woke open. I was using the Nocti with dark red filter yesterday and forgot to use the evf for the first couple of shots. Totally out of focus.

[adan]

While this could just be a "bad" filter that is not optically flat (which brand?) , I'd also suggest that backfocus with stronger red filters may be the influence of two things:

1. Infrared light. We know silicon is inherently sensitive to IR (see: Leica M8 IR contamination). Sensors are filtered to reduce IR light (it is partly why they have a greenish-cyan tinge when viewed, e.g. for dust removal). But those on-sensor filters are not perfectly opaque to IR, they just reduce it by 80-95%.

And infrared light does not focus in the same place as visible light, unless one has an Superachromat lens specifically designed to correct into the infrared range. (A regular "APO" lens doesn't, necessarily).

Back in the day, lenses came with "infrared index marks" - a red dot in the DoF scale. Taking IR pictures, one would focus visually with visible light at, say, 5 meters - and then move the focus ring so that the 5m-mark aligned with the IR Index (actual focus on the scale now at, say, 4 meters). In effect, "front-focusing" from what looked right with the eye. Then the IR image would be sharp. (image via flickr link, ex Canon FD lens instruction manual)

 

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!

Put a deep red filter on the lens (which removes blue and green light), and then add a filter factor of several stops (or allow the ttl meter to add exposure simply by reading through the dark red glass) and now you have an image that is formed by perhaps 70% visible red light and 30% infrared leakage, once adjusted by the sensor's IR filter. And that 30% IR image will be fuzzy, unless intentional front-focus is added.

2. Related to that, the visible colors in the spectrum also do not all focus in the same plane. This is called longitudinal chromatic aberration. An APO lens will reduce this - but may not eliminate it entirely, especially at larger apertures and longer focal lengths. It just depends on which three colors the lens designer chose to correct for.

Although it may vary from lens to lens, generally lenses with some residual LCA - especially with the proxy focus system of a rangefinder - will be adjusted for best focus in green light (the color the eye is most sensitive to, in the middle of the visible spectrum).

Throw away the blue and (correctly-focused) green light by using a deep red filter - and the RF focus may be "off" for the resulting single-color image.

Since it is a "focus" deviation, the increased DoF of stopping down will hide it to some extent, just like any other focusing error. And wide-open, it will be more obvious.

And - viewing through an EVF/LCD, you are not seeing (or failing to seeing) the original light wavelengths, simply what the sensor is capturing (including IR), translated electonically into visible-light pixels. If the sensor is sensitive to IR, it can "show it" to you (mixed with the visible red) on the TV set, just as IR goggles can show your eye a visible image of the IR world, even when there is no visible light ("night-vision").

 

[a.noctilux]

Thanks Andy for this explanation.

This is very clear that using dark colored filters with Monochrom can have some "more" things to understand.

I took time to learn what was going on and still in experimentation (to understand behaviors of Monochrom + colored filters).

With M246 + EVF2 or Liveview, I've learn a lot.

My conclusion is "each lens is different" in their behaviors, so no universal "one for all theory" 😵.

I've experienced some trouble ( = discrepency RF/LV ) with some lenses and those are :

 

1 - old LTM Elmar 3.5/5cm with dark red filter (Minolta R60 or Cokin R2 for example),

     at f/3.5 strange, no "red shift" at far distance but some at near about 1m - 2m

     without red filter, the RF calibration is ok for all distance no discrepency RF/LV

     at f/5.6 onward the dof can cover the discrepency

    this Elmar has engraved "R" marking same as f/5.6 for IR "film use" I think and the correction with red filter at 2m is about half way to "R"

2 - Latest optical formula Summicron-M 2/50mm with R60 filter

     no "R" marking to help, but at f/2 mostly no "red shift" nice to know when RF use wide open (for me very rare)

    but at f/4 , this has significant "red shift", strange so must have to do with "focus shift" while closing down of this particular lens

3 - Summilux-M 1.4/50mm pre-asph (E46 filter) with R60 filter

    no "R"marking and same behavior as Summicron-M (2) = no red shift at f/1.4 or f/2 =ok at all distance

    but at f/4 can have some red shift not compensated by dof

 

Experimentations not yet terminated...