[Leica] Bokeh

[freakscene at weirdness.com (Marty Deveney)]

I thought I?d responded to this last night but my message seems to have 
vanished into the void.
?
>No argument from me on the two Leica lenses mentioned for extremely >good 
>out of focus rendition! ?But others would say 'Oh the new >designs lack the 
>smoothness or glow (or whatever) of the older >lenses.
?
Those others are confused.? Bokeh is blur, specifically the blur in photos 
where they are out of focus. ?It is not blur caused by camera or subject 
movement. ?Smoothness and glow are either terms that refer to the in focus 
parts of the photo or large unquantifiable.? So this has nothing to do with 
them.
?
>So I was saying to the OP, what do you understand by bo-ke or bokeh or 
>whatever you like to call it. Is it the way that OoF 
>highlights are rendered? Or the transition from sharp to unsharp or some 
>subjective Leica quality?
?
Bokeh is blur.? How out-of-focus highlights are rendered and how the 
transition from in- to out-of-focus appears are part of it, but they are not 
bokeh. ?Bokeh is the blur.
?
There is an hilarious thread here 
http://photo.net/classic-cameras-forum/00S2yV that shows just how confused 
many people are about this. ?The angst and agression that this engenders is 
why I don?t like photo.net.
?
>For example, the last pre-asph Summicron 35 is often held to be a bo-ke 
>champion. It doesn't appeal to me at all compared to its 
>asph. replacement. I see a lot of that smoothness as inability to resolve 
>un-sharp detail, if you like. What I call under-correction.
?
We?ve discussed and looked at enough photos together that I think I have a 
pretty good handle on what you mean and like (with many other people I 
don?t).? What you describe above are aspects of the in-focus image.? The 
biggest difference between the in-focus imaging characteristics of the 
version IV 35 Summicron and the asph is that the asph shows much greater 
contrast at wider apertures, particularly in fine detail (small image sizes, 
in optical parlance).? I think this is what you like about it, and I can?t 
really disagree.? But that?s not bokeh, although it is tangentially relevant.
?
However, if you look only at the bokeh, the version IV is both different and 
in most people?s opinion, better than the asph. ?The version IV is less 
likely to create hard edges, double lines and other harsh artefacts than the 
asph.
?
Why?? Well, those aberrations you mention.? It?s often discussed that 
contrast and resolution are trade-offs. ?This is only true in systems 
chiefly limited by spherical aberration. ?Modern glass and lens design 
technology have controlled spherical aberration in many types of lenses to 
the extent where both resolution and contrast can be raised significantly.? 
Spherical aberration becomes harder to control at short focal rations ? 
essentially the faster the lens the harder it is to control - testing the 
new Leica M 21/1.4 will be interesting.? What Leica did with the early asph 
lenses, in my opinion (caveat ? I have only tested the 90/2 and the three 
35s, not the 21 or 24) was to over-correct (?positive? control) the 
spherical aberrations and increase the contrast transmission in the image 
too much.? I think Leica thought this too, because the 50/1.4, the 75/2 and 
the 50/0.95 (it remains to be seen for the others) have more moderately 
controlled spherical aberration and lower contrast than the earlier asph 
lenses. ?The degree of control of aberration is hard to measure without an 
optical bench, because -0.1%, for instance, will cause a very similar amount 
of unsharpness in the corners of the image as +0.1%? If you can read German 
this is all explained in the summarised lectures of Ernst Abbe: S. Czapski 
and O. Eppenstein, Grundzuge der Theorie der optischen Instrumente nach Abbe 
(2nd ed., Leipzig, 1903).
?
So lenses that are slightly under-corrected or extremely well corrected for 
spherical aberration generally show good bokeh, but lenses that are 
over-corrected for spherical aberration will fairly consistently show 'bad' 
bokeh, characterised by double lines, rapid harsh transitions and an edgy 
appearance to objects in the bokeh.? Remember also that one lens will show 
different characteristic bokeh in front of and behind the in-focus image and 
that bokeh will change with aperture and distance to the in-focus areas and 
the distance between the in-focus areas and the out-of-focus areas.  Phew.

So why over-correct in the first place?? It is rumoured that in the 1960s, 
Nikon?s marketing research showed that people perceived the in-focus part of 
pictures as ?sharper? if the lenses were slightly over-corrected for 
spherical aberration. ?This may have something to do with the way in-focus 
areas blend into out-of-focus areas, but needs further investigation.  
Investigating why perception works as it does is difficult.
?
Remember that all lenses that are available to us are LOADED with 
aberrations, even the absolutely best corrected.? It?s always a trade-off 
and there are things more important than sitting around wondering about 
optical correction of aberrations, unless you?re a lens designer.  If you 
are, tell me why can't I buy a still camera lens as good as the Zeiss Ultra 
Prime and Ultra 16 cine lenses?  The answer, of course, is that if someone 
made one, no-one (including me) could afford it.
?
Ultimately, this means that any lens can be designed to have excellent 
in-focus performance and excellent bokeh. ?The 50/1.4 asph and the 75/2 show 
this.? I hope that someday Leica will redesign the 35 asphs to better match 
the rendition of the newest 50 and 75, but I?m not holding my breath.  The 
main reason that this has not been happenning is the perception of sharpness 
issue described above and the fact that bokeh has only really been widely 
discussed over the last 10-15 years, although film makers have been aware of 
it for a lot longer than that.
?
What is good bokeh?? I once tried to undertake a survey of photographers and 
non-photographers to describe preferences in bokeh (in the hope of 
publishing it in a photography magazine) and found a few interesting 
things.? Most interesting is that most people don?t care, even if they are 
photographers ? my mother exclaimed ?it?s all just blur!? ?Of the people who 
did care, they generally disliked double lines (ni-sen bokeh) or sharp 
transitions where when viewed by normal vision the scene would appear to 
have smooth transitions.  People also generally liked similar kinds of bokeh 
(e.g. someone who liked double lines liked them from different lenses and to 
different extents)? Sharp transitions seemed okay to most people where they 
would naturally occur (such as in spotlights on a dark background). ?But 
there was sufficient variation for none of it to be anything other than 
indicative.  I tend to agree with Mark R. that good is, well, good and that 
in this case that means ?smooth?.  All this just affirms that in aesthetics 
anything goes.  Good is whatever you think is good.?

After all that, to reinterpret Doug?s question as ?what lenses do I think 
produce good out-of-focus blur and why?? I offer:
?
I think that the best bokeh in a lens for 135 film or comparably sized 
digital sensors is from the 35/2 version IV at f4 or 5.6 and focused between 
1.5 and 5 metres. ?Its bokeh shows no harsh transitions or double lines, 
objects almost seem to collapse onto themselves but smoothly, gradually.? 
Incredibly gradual, gentle transition from in focus to out of focus (this is 
important to me). ?Note that this lens flares readily, its bokeh is less 
pleasing and it is insufficiently contrasty for my liking wider open than 
f4. ?It also behaves somewhat strangely when focused closer than 1.5m.? It 
has other optical issues that displease me, mainly to do with distortion, 
but overall I think it?s a stellar lens, especially for its time.
?
As I said previously, the 50/1.4 asph and 75/2 have excellent bokeh and I 
would not fear using them from wide open to f11 (stopped down more than that 
diffraction starts to become apparent) or focal distance.? Smooth, though 
rapid, transition from in- to out-of-focus, no doubling.
?
The Zeiss 110/2 for the Hasselblad cameras has?smooth, even blur, even wide 
open.? When I borrowed this lens I used it wide open and close up just to 
look at the bokeh (sad but true).

The absolutely most consistent bokeh I?ve seen is in photos taken with the 
Rodenstock Apo-Sironar-S lenses. ?Wide open, stopped down, near or far 
focus.? Amazing.? Reason enough to get a large format camera.? I need to get 
one of these while I can.  I also need to take more pictures and think about 
bokeh less.
?
End bokeh rant.? Let?s all go take some photos.
?
Marty



Gallery: http://gallery.leica-users.org/v/freakscene


-- 
Be Yourself @ mail.com!
Choose From 200+ Email Addresses
Get a Free Account at www.mail.com