Olympus-OM
[Top] [All Lists]

Re: [OM] Dude - Who stole my 0.56ms?

Subject: Re: [OM] Dude - Who stole my 0.56ms?
From: "Carlos J. Santisteban" <zuiko21@xxxxxxxxx>
Date: Thu, 8 Apr 2010 20:33:26 +0200
Hi Ken and all,

So many posts, so little time! There's A LOT of interesting ideas flowing on
this thread, and I'd like to add my experience and opinion before it gets
buried under a ton of late posts... :-) Apologies in advance for not
sourcing all quotes.

>Using three different cameras, the E-1, the OM-3Ti and the OM-2S, the
>aperture wide-open isn't F1.4, but closer to F1.8.
<snip>
>F2.8 at 1/250
>F2.0 at 1/500
>F1.4 at 1/640
>
>So, my confusion comes in with the the F1.4 reading. Shouldn't it be F1.4
at
>1/1000?  Who took my .56ms?

We've got three things to consider here:
1) Metering indication on OM bodies (wide-open metering)
2) Metering indication on adapted digital bodies (stop-down metering)
3) Actual exposure on digital bodies -- this was never an issue with film,
was it?

>Check the mechanical travel on the aperture you will know the difference
>from F1.4 to F2 is much less than from F2 to F2.8,

Yes, but that happens on _any_ OM lens -- the first stop is "shorter" than
the rest. On any F2 Zuiko, 2 and 2.8 are closer than any other stops. The
same with 2.8-4 on F2.8 lenses, etc. And those with a
slightly-more-than-a-full-stop maximum aperture (e.g. 1.2, 1.8, 3.5, 5)
usually have regularly spaced click stops all the way -- checking now a
135/3.5 (between 3.5 and 5.6) and a 200/5 (between 5 and 8)

But Pentax-K and Minolta manual lenses do pretty much the same: the first
stop is much shorter on the aperture ring -- to an even higher extent, I'd
say!

On the other hand, Nikon, Yashica/Contax, Konica AR and Canon FD have evenly
spaced apertures -- and so seem Tamron Adaptall's. Can't say about other
brands because I have no experience with them.

But I have no idea of why is this made -- and why some brands can live
without it! I don't think it's for simplifying the iris construction --
rangefinder and preset lenses, freed from any diaphragm coupling to the
body, are sometimes the quite the opposite (tighter spacing at the
_narrowest_ apertures)

>So on the OM
>bodies, the change in indicated shutter speed is simply a factor of how
>far the little tab on the back of the lens moves and how the body
>interpretates that movement, as there is no change in the actual amount
>of light hitting the meter.

Certainly. In fact, OM bodies (and many other brands) have no idea of the
_actual_ absolute aperture -- nor have any need to know.  They just "ask"
how many steps will the iris close down when shooting, in order to adjust
exposure readings from what they "see" at full aperture. In other words, the
OM meter can't tell a F4 lens on bright sun from a F1.4 lens in the shadow.

But there's some mystery added here... in theory, any lens set to the widest
aperture _should_ have the meter coupling tab at the same home position,
with the meaning of "no correction, will shoot at metering aperture", much
like the 28/3.5 here <http://www.flickr.com/photos/zuiko21/4438421803/> --
check the _left_ side (as seen in the pic) as is what actually contacts the
metering ring on the body.

This is true of, say, 2.8 and slower lenses, but fast lenses have the tab
somewhat displaced to the left -- several mm in case of the F1.2. Again, Oly
is not alone here: the same happens (as is documented in repair manuals)
with Nikon AI (a bit less, though), Yashica Y/C and Minolta MD.

Canon FD may seem the sole exception... but it ends doing the same: their
lenses have a pin in the back indicating the _maximum_ aperture of the lens,
which presses a button on the body that effectively offsets metering -- but
only for faster than f/2.8 lenses.

This time, however, I _think_ I've got an explanation for this... read more
later!

>On any of the OTF bodies in auto mode, of course, inaccuracies in the
>visual indication in the viewfinder aren't reflected in the actual
>exposure.

Sure. Modern rangefinders do a sort of OTF metering (Off the Curtains,
really, not real-time) which seems pretty accurate on matter the aperture...
even when using the low-light trick of metering at f/0.7 -- just remove the
lens! ;-)

>However, you are making another possibly unwarranted assumption - that
>the metering system works properly with MF lenses.

Be careful about metering mode: MATRIX metering (or whatever the maker calls
it) will be confused with adapted lenses -- that _needs_ to know the actual
aperture of the lens. Centre-weighted should be fine... except for the
widest apertures -- see below! It seems that spot metering in stop-down mode
is usually affected by small apertures, much like a split image screen.

>I tried your experiment with OM 50/1.4 >1,100,000 on the 5D. I got some
>somewhat strange and inconsistent seeming results. For example, there
>was often no difference in indicated shutter speed between f1.4 and
>f2.0.

My experience with the 300D mimics Moose's -- little increase from F2.8 to
F2, and _no_ increase at F1.4.

But I _think_ that TTL (non OTF) meters can't "see" more light than a f/2.5
(or so) aperture gives... similar phenomenon to the 2-x screens -- and I
believe it could be the cause for the "metering tab offset" discussed above.

>I remembered this test I did - with the same lens -and
>the 300D. <http://www.moosemystic.net/Gallery/tech/300D/300Dexp.jpg>

Interesting test. The 'auto' results are much like I had, although manual
mode gave me similar results -- IIRC. Another interesting test would have
been for 'unmetered mode', taking a central exposure as reference, and just
reciprocating for every aperture, in order to see the light actually
captured by the sensor... they all should be the same, shouldn't they?

>Olympus has designed the lens so that the travel of the aperture lever
inthe >lens mount is less than one stop when going from wide open to f/2, to
>keep the metering accurate.

I don't understand this... just tried the OM-1 and it meters reciprocally at
any aperture _except_ the widest, which reads about half a stop under (after
selecting a one stop faster speed). Tested with 50/1.2, 40/2 and 100/2.8.
Also tried the 135/3.5 and seemed reciprocal... but if I select one stop
faster speed while going from 5.6 to 3.5, it _should_ read a bit over, no
equal exposure! So, no compensation is done for the small first stop, anyway
:-(

>Comparing histograms and general appearance

Ahem... you all know that I'm not a fan of histograms... but anyway, they
could be fooled by vignetting -- which shouldn't be much of an issue for
in-camera metering otherwise, because it's centre-weighted anyway.

I think the best method for objectively "metering" off the pictures is to
measure the RGB value on pixels around the middle of the image -- the exact
value is meaningless, but should read _around_ 128 for all channels
(assuming a neutral subject). There's no exact relationship between
different RGB values and f-stops, because they depend on sensor,
firmware, transfer curves etc... but could be tested at a fixed aperture
;-)

>However, the final one at F1.4 is underexposed by 2/3 of a
>stop, which is exactly what I've been seeing all along with the exposure
>meters.

Well, here's the third question: the actual exposure, so metering seemed
OK... I have another theory for this: maybe SOME sensors can't get _all_ the
light from very fast lenses -- each pixel is in a delatively narrow well,
with a microlens on top, so it won't surprise me...

I have to do some testing with the GF1 and the fastest lenses I own! ;-)

>Hey, on a positive note, this late-model F1.4 lens is EXTREMELY sharp.
>Eye-popping sharp!

Well... it's a nice lens, and the latest verison is definitely above the
previous iterations. It shows the classic high-speed lens performance curve,
with limited contrast/sharpness (but fine resolution) at 1.4 _and_ 2,
boosting clearly at 2.8 with much slower improvement beyond that. Compared
with the older MC, sharpness is similar but with much better/stronger
colour!

Anyway, having tried many other brands... even these late Zuiko 50/1.4's are
below average at the widest apertures :-( But that's a technical comment,
phtography is much more than that, and they do nicer (to my taste) pictures
than the "technically superior" 50/1.8 miJ.

>Next, maybe you need to try a Super-Multi-Coated Takumar 50/1.4. ;-)
>(Do I have that right; fully spelled out version better than later,
>"SMC" version?) That one is famous even outside of Pentax circles.

I had a late version in Pentax-KA mount -- definitely spelled SMC. Pretty
good from f/2, but in the soft side wide open. Sold long ago.

>The point is that by having it at infinity focus I had no
>light-loss due to extension which might have created another variable to
>question.

According to my old calculations, a 50/1.4 focused at 0.45m should be around
f/1.6... but proportionally the same at any other aperture, so it won't be
an issue here. Vignetting characteristics could change, though.

>I carefully duplicated the test
>sans-film with the OM-4T on my laptop screen

Not saying it's bad, but I won't trust such light source... or anything
coming from any kind of fluorescent :-(

>using the spot-meter mode and
>observed that the F1.4 position was 1/3 stop off from the other apertures.

Certainly because of the smaller-first-stop and/or rounding errors --
matches my OM-1 experience and what I recall from the OM-4Ti (it has no
batteries right now)

Now that you're talking about the spot meter... I know I will be flamed by
this, but I'm afraid the spot metering of the OM-2S/4/4T (and the 3 series,
I suppose) doesn't work OK with all lenses!

I did some extensive research many years ago and could check this with
several bodies and even several instances of the same lens... I will repeat
this test soon and I'll publish my results here. In short, _some_ lenses
(doesn't seem related to focal length or aperture) read somewhat lower in
spot mode than in non-OTF centre-weighted, leading to up to more-than-one
stop overexposure...

My first OM was a 4, but I didn't notice this... because my lenses back then
(Tamron SP 90/2.5, SP 70-210/3.5 and the Zuiko 35/2...) were in the category
of "negligible error". IIRC, the 21/3.5 is one of the worst, and the 28/3.5
is much worse than the later 2.8. But some other fast lense have big errors
too. I'll chech when I have some time.

>What I'm thinking here is that the actual mirror
>chamber itself isn't large enough to support maximum aperture of the 50
>F1.4.

Could be...

>As to the spot-metering of the OM-4T being off by 1/3 a stop, just
>maybe the same thing is happening there too.

OTF metering should tell... EEEEEK! My OM-2n with outdated film makes almost
no difference between f/2 and f/1.4 (about 5 seconds in very dim light)
whereas at f/2.8 it goes for 10 seconds... what's going on here??? :-(

>If so, this is a problem for regular FourThirds cameras, but with the
>shallow chamber of the MicroFourThirds, it might be just fine.

I definitely have to try it -- i had some metering issues with adapted
lenses, but exposure seemed to be coherent at any combination.

>What if this veiling glare is not a
>quality issue with the lenses themselves but the interaction of the lenses
>with the camera's tiny mirror chamber and the internals of the camera are
>reflecting the light from the front of the chamber back into the lens?

The image circle of an OM lens is much larger than expected -- there's a lot
of places where unexpected light could be reflected, adapter included.

>2. Loss of aperture linearity at the wider apertures with other lenses.

???

>but other 50/1.4 lenses may not have the same issue due to
>the position and size of the rear element and the location of the
rear-nodal
>point.

Should try them, of course. With the GF1 at least.

>but I was wondering if you've ever tried
>the Canon EF 50mm f/1.2L?

I tried the Canon FD version (L too), not sure if it has the same optics. It
was, in my experience, the best performer in its class... but now, fast CV
lenses draw circles around it...

>Moose's comments got me to thinking and I did pull out a couple of slides
>taken with my old Silvernosed 50/1.4. I still believe that the old one is
>sharper in the center but the corners were another story.

I had all iterations of the Zuiko 50/1.4. Sold them all except the latest
and a very old silvernosed version -- much softer but with an interesting
quality.

>However, I did spot some lens flare issues which I didn't expect to see
from
>a mega-multi-coated wonderpiece.

It's a Planar (double gauss) design which is naturally flare-prone, even
with the greates coatings :-( That's why the later and inferior Tessar was
developed.

>According to the talking heads around here, the switchover to the final
>version happened at 1100000.

IIRC, some say 1085xxx.

Some more testing is due... Cheers,
-- 
Carlos J. Santisteban Salinas
IES Turaniana (Roquetas de Mar, Almeria)
<http://cjss.sytes.net/>
-- 
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz