Hi, all. I've been very busy but it's time to get out from lurk mode...
>The story began when I observed that when I opened the 28/2.8 one stop from
>22 to 4, speed moved up one stop faster (or in manual clicking the speed
>ring towards faster speeds to keep the needle of my OM-2 centered) -
>*except* when opening up from f4 to f2.8.
>In that full stop opening, meter showed only a half stop rise. Yes, no
>matter how much troubleshooting I did.
Well, according to my experience, it's as Gerard said:
>There is nothing wrong with your lens. All Zuiko lenses are the same:
>the difference between full aperture and the the first stop is less
>that what you would expect.
*Many* SLR mounts have this 'narrower first f/stop': Olympus OM, P*ntax-K,
Min*lta MC/MD, K*nica AR... however, there *are* mounts with perfectly
linear aperture scales: Nik*n, Can*n FD, Y*shica-Cont*x, even the Tamron
Adaptall-2 system. OTOH, most M42 gear have no full-aperture metering
coupling, so I think it won't matter with them.
But I'm not sure about the rest of Gerard's explanation:
>The reason for that is that all markings on the lenses (except wide
>open) correspond to optical f/stops (light transmitted through the
>lens).
I would believe that all marked stops are either f/stops (geometrical) or
t/stops (optical), but never a mixture of them. Anyway, I think the
difference between both scales should be around 1/3 stop *with little
variation between lenses* -- I've checked this with a list of transmission
values of most Russian lenses, from various types and vintages.
>on 11/01/2005 13:23, John Hermanson wrote:
>
>> Normal! Try many lenses going to the maximum f stop and you'll see that
>> meter movement between the last 2 clicks is oftem just a half stop instead
>> of one. I heard about this back in the 80's. Just use the lens, it's okay.
Obviously, the position of the aperture ring won't matter to the OTF-Auto
exposure system, but if we assume that the marked stops are (at least)
evenly spaced, then there *is* an error in *manual* metering: +0.5 stops.
Maybe not too noticeable, and easily compensated by 'un-centering' the
needle a bit on the down side when shooting at full aperture. However,
shutter speeds can only be changed in 1-stop intervals, so... it's probably
not an issue.
>From OLYMPUS OPTICAL CO., LTD.:
>With reference to the aperture of 28mm F2.8 lens, the TTL Number System for
>Pen FT camera is applied to the OM-System lenses. According the the [sic]
>Pen FT system, the aperture numbers on the aperture ring are inscribed
>unevenly so that the light intensity can be reduced by half by stepping down
>one stop.
For those with no experience with the Pen FT, the TTL meter on these
cameras is uncoupled: the needle points to a number on the viewfinder,
which in practical terms means the number of f-stops (from full aperture,
where the metering is done) needed to stop down in order to get proper
exposure -- it's not *that* difficult, just a quick turn to wide-open and
then count click-click-click the other way round ;-)
Pen FT lenses have two separate (opposite), switchable aperture scales,
each with its own set of click-stops; one bears the typical f/numbers, and
the other is the 'metering scale' -- usually 0-1-2-3... And yes, the
distance between the '0' (full aperture) and '1' is, in most lenses... 1.33
stops!!! The 38/1.8 has '0' @ 1.8, '1' @ 2.8 (1+1/3 stops slower), '2' @ 4
etc. The 20/3.5 shows a similar pattern: if you double the f/numbers. And
the pancake 38/2.8 starts with '0' @ 2.8 and the '1' @ 4.5, '2' @ 6.3 etc.
>f-stop calibrations are
>mathematically formulated and apply only to the center area of the lens. If
>the f-stop system were correct, a lens with a maximum aperture of f2, when
>used wide open, would transmit the same amount of light as an f1.4 lens
>stopped down to f2. Tests prove that the latter transmits more light than
>the former.
Well, f/2 of a F2 lens *should* be the same as f/2 of a F0.00000001 lens...
at least in the centre. Possibly the smaller measured transmission of the
slower lens is due to vignetting or light fall-off at corners, while the
faster lens (with its larger lens elements) is sort of 'overkill' for the
same aperture. This agrees with the following paragraph:
>This accomplished by spacing the TTL numbers unevenly so
>that the light measurments are averaged at both the center and the edge of
>the film.
And then...
>The uneven spacing of the TTL numbers also correct for the errors
>caused by the curvature at the edges of the lens
I don't understand this -- and I think I *don't* want to understand it ;-)
>Since the CdS cell of the TTL metering system is placed behind the
>viewfinder prism, the light which reaches the cell is influenced by the
>fresnel lens as well as by the camera lens itself. The
>
>-------------------------
>page 1-8
>
>fresnel lens centers the light on the CdS cell, and the longer the focal
>length of the camera lens, the greater the exaggeration of this centering
>effect. Therefore, readings taken with telephoto lenses differ from those
>taken with shorter lenses. A subject that reads "0" with a 60mm lens would
>indicate a reading of "1" when taken with a 100mm lens. On Olympus Pen FT
>lenses, corrections have been made so that the maximum aperture of each
>telephoto lens reflects a different TTL number for the maximum opening.
Hmmm, this explains why telephoto Pen lenses don't start their 'metering
scales' at 0; for instance, the 70/2 starts at '0.5', with '1' @ f/2.5, '2'
@ f/3.5, '3' @ f/5, etc.
>As John suggested, I tried my 21/3.5 which shows a full stop difference
>between 3.5 and 5.6,
But, but, but... between f/3.5 and f/5.6 there's a difference of 1+1/3
stops -- just like the FT scale!
>as does the 200/4 between 4 and 5.6.
I haven't got the 200/4 at hand, so I can't check, but I'm afraid this
first stop should be a bit smaller than the rest.
>Still studying the
>85~250 between 5 and 8 at different focal lenghts.
Again, 1 1/3 stops, like the 21mm.
>It was also very weird of an acquaintance of mine who has two plain 4 bodies
>and two 28/2.8 for sale. He would be happy if I bought him one, so he told
>me on the phone that his 28s on his 4s showed a full stop difference between
>2.8 and 4.
IIRC, the OM-4 with the 35/2 showed a 2/3 stop increment between f/2 and
f/2.8, but the little play on the aperture ring of my 35 allowed to get the
full stop difference. Maybe your friend is not lying, just not 100%
accurate.
>Basically, set up a camera on a tripod in a dimly lit room pointing at a
>blank wall
>about 2m away. Unexposed film in the film gate, time the actual exposure
>with a stop
>watch. I found using a 50/1.4 initially that within a certain lighting
>range the timed
>exposure always doubled as I stopped down (expected result), except for
>f1.4 to f2
>where the time was less than double (about 1.5 from memory
Quite odd... maybe vignetting/fall-off was striking.
Another issue with TTL metering, no matter the brand, is that it looks like
the meter isn't "able" to "see" the light from apertures larger than f/2.4
or so -- the error is small (about half a stop) down to f/1.8 or f/2, but
get over a full stop at f/1.2-1.4. In *all* the aforementioned mounts,
there is some kind of mechanical compensation on *fast* lenses (usually is
a small offset of the aperture signal pin) because exposure is metered at
full aperture, which would result in overexposure without compensation.
This effect is clearly detected with stopped-down metering (eg. EOS with OM
adapter)
I always thought OTF metering wasn't affected by this (the light won't get
thru the fresnel lens of the screen, which may be the culprit), but maybe
this is not 100% safe...
>I also tested another lens (can't remember which, possibly 100/2.8) with
>the same
>result.
Mmmm, but at f/2.8 and beyond, this metering issue dissapears. My vote goes
for the vignetting issue, then.
>- items I used to discuss offlist with Carlos Santisteban when the thread
>'Long exposure capability of OM2n' during May 2004.
Yo! I'm here again, wasting a bit of the list's bandwidth :-)
>From: "Ing. Radovan Faltus" <faltusr@xxxxxxxxx>
>
>So the conclusion is: The effective speed of
>Zuiko lenses is half stop lower than rated.
>
>Can anybody refute this with calibrated lightbox measuring?
I hope this is not true... but the best test would be with film, and a body
with a well-known calibrated shutter -- checking the reciprocity of several
equivalent aperture-speed combinations.
Maybe I'll do some testing tomorrow, with the EOS-300D -- now it's too
late, and tomorrow I expect to have a 'busy' morning'...
Enjoy,
...
Carlos J. Santisteban
<cjss@xxxxxxxxxxxxx>
<http://cjss.galeon.com>
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|