[Leica] M8 Candles...

[pklein at 2alpha.net (Peter Klein)]

Whether part of an experiment or not, I think these candle shots show 
something.  Now, we must stipulate that where Steve actually focused may 
have been influenced by holiday spirit(s).  :-)  But I've seen other M8 
results with a similar effect. They point me to the following working 
hypothesis.

In Steve's picture of Tom,
http://gallery.leica-users.org/v/barbour/M8candles/tom.jpg.html
notice that the the pendant around his neck is sharper than his eyes.  This 
is even more apparent in the full-size version.  Assuming Steve focused on 
one of Tom's eyes, the plane of focus is farther away than where the 
rangefinder told Steve to focus.

Now, think about infrared photography.  I was just reading the old Theo 
Kisselbach "The Leica Book," where he describes how to focus for IR.  He 
mentions that you have to extend the lens further out for IR 
photographs.  In other words, focus closer than the rangefinder tells you 
to, because the IR image focuses as if it were farther away than the 
visible light image.  He even gives figures for lenses that don't have a 
special "R" mark for infrared.  For example, to get IR in focus at 
infinity, you should focus a 50mm lens on 15 meters (~50 feet).

So, let's take a candlelight picture. We are focus the rangefinder with the 
light we can see.  But there is a strong, invisible IR component in the 
light, which is in focus a bit behind the visible light focus point.  So, 
depending upon the ratio of IR to visible light, we may get a picture that 
is:

1. Sharp at the point on which we focused (mostly visible light)
2. Sharp at a point slightly behind where we focused (mostly IR).
3. A bit blurry at both the visible and IR focus points (about 50/50 ratio 
of IR to visible), because the OOF component of visible light blurs the IR, 
and vice versa.

The problem with all this is that our eyes can not tell us the ratio of 
visible light to IR, as we can't see the latter.  I would strongly suspect 
that candlelight has a higher percentage of IR in its spectrum than does 
household tungsten bulb light, which is why we see the more drastic focus 
shifts in candlelight pictures.  In the latter, we have to contend only 
with dark fabrics (low visible light) turning purple or shifting, but the 
IR/visible ratio is not so high that we notice much focus shift.  (BTW, in 
Steve's picture, Tom's shirt is distinctly purple if you boost the levels a 
bit).

I should mention that I have seen a similar effect with red light vs. green 
and blue.  I have lately been examining the individual R, G, and B channels 
when I create B&W conversions of some of my E-1 photographs (thanks to Mark 
Davison for this tip).  The red channel (longest visible wavelengths) 
always appears a little blurrier than the others.  Which means that my 
Zuiko lenses are not apochromatic, I guess.

All this should be easy to prove or disprove.  Some enterprising M8 owner 
needs to take a few pictures of a focus chart at 45 degrees to the lens 
axis, using both candle and tungsten light, both with and without an IR cut 
filter.

--Peter

At 08:35 AM 1/7/2007 -0800, Steve Barbour <kididdoc@cox.net> wrote:
>Leica M8,  firmware 1.09, Summilux 35mm/1.4 ASPH... all same lens,
>same room, same evening, focussing... comme-ci comme-ca (some after
>champagne)...
>
>5 images are in this folder...
>
>http://gallery.leica-users.org/v/barbour/M8candles/
>photos from New Years Eve...
>(definitely NOT taken as part of an experiment)