Matching lenses (was: Re: HTML help)

[Joel Alpers <joel.alpers@xxxxxxxxxxx>]

Greg Erker wrote:
> 
> >It looks to me like if you take out the first <blockquote>
> >that you'll get what you're wanting. Am I missing something?
> 
>   I don't think so. When I didn't
> have it in there the text just butted
> up against the photo and continued
> at the same width all the way down
> (didn't get wider below the photo).
> 

taking it offline, as it isn't really too "3D MF" related.


But here's something that might be more of interest... I got a request
from someone to buy two matched Pentax lenses (OK, this is miniature
format, but hold on).  I had about 10 50mm Pentax-K lenses, so yesterday
I did some matching using a technique Les Gehman thought up. I wouldn't
mind a "sanity check" on what I've done  ;)   and this may be helpful
for anyone trying something similar, with all the talk of twinning 
MF folders that's been going on.

With 10 lenses to check, I set up a tripod on my deck, and framed
a scene with a sharply defined vertical line at each end of the image.
(I used two roof lines). Then just took a photo using each lens, putting
the lens serial number in each photo. This was done on B/W film, which
I then developed in the darkroom and mounted in GePe glass slide mounts
(to eliminate slide "pop" when projecting. Slide film could of course
be used, but B/W is quite a bit cheaper (I buy in bulk rolls) and
quick, easy, and cheap to develop at home.

I projected all the slides onto a large sheet of paper, with the
image at about 6' (best I could do in the space available. 

Then I would line my chosen mark on the left side with a reference
mark on the paper, and pencil in a mark on the right point, with the
lens' serial number.

Since my two image reference marks were 6' apart on the projected
image, a difference of about 1/3 of an inch represents a 0.5% mismatch.
It was fairly easy to pick lens pairs that had MUCH better matching
than this. I picked three pairs with better than 1/8" matching, which
is (trying to allow for measurement error) 0.2% or better.

Thoughts? Anything I've overlooked?  I seem to recall that 0.5% is
a generally-agreed on number. Is that right?

Joel.