|
You have to be warned, however, that there are two limitations with this adapter:
-
It offers only the internal
bayonet, not the external one (which is used on just a few large and
heavy lenses). This is by design, but I’m not happy with it, especially
that the outer flange would not interfere with mounting of lenses with
the more common inner bayonet. In a piece costing $175 I would expect
full Exakta lens compatibility.
- A minor design error makes it impossible to use
some of the later Zeiss lenses, including the 2.0/50 Pancolar, 2.8/50
Tessar, and some others. I believe that all late, automatic Zeiss
lenses in black (or striped) finish may be affected by this problem.
Mr. Gandy of Camera Quest is aware of the problem and I hope it will be
rectified with the maker.
April, 2006:
No news if the problem was addressed; if you want to use your black
Zeiss lenses, check with the distributor first. I think that using a
Dremel tool to take off a small piece of the bayonet lip should help,
but I haven’t tried that yet.
Focal length and field of view
The Four Thirds camera sensor, used in the E-1 and E-300 cameras, has a
linear size almost exactly half of the standard 35-mm frame. (This is
exactly true in terms of the frame diagonal, and slightly varies for
image sides, as Four Thirds has a different aspect ratio. Risking a
slight simplification let’s just use the value of two.)
This means that when a lens of a given
focal length, F, is mounted on a Four Thirds camera, the part of the
optical image used by the sensor will be one half of the size of such
part used by the same (or just the same F) lens mounted on a 35-mm
camera. The camera’s field of view is reduced by half (linear; the area
is reduced four times).
(Some writers refer here to the “image
angle”, but this, while approximately OK for longer lenses, is not true
for shorter ones. If you paid attention in your high school math
classes, you remember the difference between an angle and its tangent
function value; if you haven’t, just take my word for it.)
To get the same field of view on a film
camera, you would have to mount on it a lens of the focal length twice
that value, i.e., 2F. (I’m usually referring to the latter as EFL, or Equivalent Focal Length, in terms of 35-mm film.)
Note: All I’m saying here is also applicable to digital cameras with
APS-sized sensors, except that the factor of 2x I’m using all around
the place has to be replaced with 1.5x.
This means that, for example, that any 50 mm lens mounted on an E-system body will have a field of view similar to that of any 100 mm lens mounted on a film camera.
The lens aperture, defined as the ratio of the entrance pupil diameter to the focal length, obviously does not change.
Lens resolution
There is one important difference, though. The recorded image (from
film or digital sensor) has to be enlarged to be viewed (as a print, or
on a screen). Because of smaller original size, the enlargement factor
for an E-system is twice as large as that for a film camera. This means
that all flaws of the image will be also enlarged twice as much; among
others, the effective resolution of a lens used on a Four Thirds body is only half of that for the same lens used with on a film camera.
We may put it differently: to achieve the same final image quality, a
lens designed for the Four Thirds system has to be twice as sharp (in
terms of absolute resolution) as a lens intended for a film camera.
A lens for a film-based system may, for
example, have a central resolution of 60 lpm (lines per millimeter),
and for many purposes this may be considered satisfactory, if not
stellar. When this lens is mounted on a Four Thirds camera, however,
the image flaws are magnified twice as much, so the resolution (while
still being 60 lpm) becomes equivalent to only 30 lpm (with the same
print size).
For nitpickers only:
the same argument holds when expressed in terms of the MTF (Modulation
Transfer Function), where different image enlargements can be
translated into rescaling the horizontal axis of an MTF
response-vs-frequency plot by a factor of two.
Now, camera and lens makers are not here
to save the world, whales and all; their basic goal is to earn money by
making us buy their products. Making lenses of higher resolution costs
more, so most lenses provide just enough resolution for their intended
use. The latter varies with the target market: lens resolution
requirements of a professional photographer will be higher than those
of an entry-level amateur, as the latter usually needs smaller prints
and has generally lower expectations.
What does this mean for you?
The second half of the preceding section boils down to two simple points:
-
Whatever you do, your images made with an non-4/3 lens on an E-series
camera will never be as sharp as when the same lens is used on a film
camera. (This assumes that the lens is the bottleneck of the process,
not camera shake and the medium, i.e., film or sensor, resolution.)
- If you are choosing such a lens to be used on
your E-1 or E-300, limit the choice to high-quality lenses. These
usually will be prime lenses, i.e., ones with fixed focal length, not
zooms.
What may help, is the
fact that a digital camera uses only the central part of a
“non-digital” lens image, which is usually the sharpest. Not always:
advanced lens designs may trade off some center resolution to gain more
resolution in the outer regions.
It may happen that, even with the 2x
factor in effective resolution loss, a good film-camera lens will be
better than a middle-of-the-road ZD (Zuiko Digital) or third-party one,
designed for the Four Thirds system. The only real way to tell is to
try.
Depth of field
A general introduction to depth of field (DOF) and its peculiarities
for digital cameras as compared to film ones is given in my DOF article, so I will not be repeating myself here, giving you just the bottom line instead:
When any lens of a focal length F is mounted on a Four Thirds body and
stepped down to aperture (F-number) A, its effective depth of field
will be the same as that of a lens with focal length 2F stepped down to
aperture 2A, and working on a film camera.
This is an ugly, but accurate,
sentence, and I wasn’t able to make it any friendlier. Just in case,
read it once again. Slowly.
This is good news if you want to
maximize the depth of field (macro, landscape), but bad news if you
want to keep it shallow (portrait).
For example, if you mount a 50 mm,
F/1.4 lens on the E-1 and use it wide open, the DOF will be the same as
when you mount a 100 mm, F/2.8 lens on a film camera. While the depth
of field is still quite shallow, making the setup suitable for portrait
photography (not to mention the perfect focal length), it is not as
shallow as it would have been in a 100 mm, F/1.4 lens. The
light-gathering capability is still defined by the actual aperture of
F/1.4.
What are you giving up?
The Four Thirds system introduced an entirely new, fully electronic
standard of communication between the lens and the camera body.
Obviously, this system is 100% incompatible with any lens-camera
coupling system existing before, regardless if the lens adapter you are
using to mechanically secure the lens on body.
What this means is that there is no
communication in either direction between your E-system camera and the
lens. In particular, you will be facing three limitations:
-
The camera must be set to manual focus, and you will have to focus manually with the lens focus ring.
-
After the image has been composed and focused at full aperture (the
latter for best accuracy), you have to step the aperture down to the
desired value by hand.
- The camera’s autoexposure system will work, but
only in the aperture-priority mode; the release button has to be
half-pressed (or AEL switch used) only after the lens has been closed down to the working aperture.
For some kinds of photography these are not really painful limitations,
especially if you shoot wide open, but the necessity to step the lens
down manually may be a disadvantage in action photography.
Exposure accuracy
We grew to expect that through-the-lens metering is usually accurate
and dependable; after all, the camera measures the same light which is
used to form the image or a film or digital sensor, right? Wrong!
During his extensive trials of OM Zuiko lenses on the E-1,
John Foster observed that the aperture-priority autoexposure was often
wrong at the widest apertures, resulting in overexposed images. We were
both puzzled by that, but John stands by his results. As soon as I’ve
put my hands on a 1.8/50 OM Zuiko, I was able to confirm the same effect at F/1.8, disappearing at F/2.8. What gives?
Then I noticed one more thing: closing the aperture down from F/1.7 to
F/2.8 (about 1.5 EV, or almost a factor of 3x in terms of light amount)
was not causing any noticeable darkening of the image seen on the
viewfinder screen. Moving down to F/4 and beyond was, as expected,
making the viewed image darker.
Hallelujah! Suddenly all seems clear (or
almost so). The fine print below is intended for investigative minds;
is you want just the bottom line, skip that discussion.
The optical image on the screen is
almost identical to that on the sensor, with one exception: there is a
finite-size mirror in the optical path. For a mirror large enough, this
difference would have no effect. As it is, the mirror effectively plays
a role of yet another, rectangular (or, more accurately, trapezoidal)
aperture, cutting off the peripheral parts of the light “beam”. Mind
it, not just the light falling beyond the sensor, but also within it,
as that aperture is still quite far from the image plane (i.e., the
screen). The mirror swings off the light path during the actual
exposure, therefore this additional aperture effect does not affect the
recorded image.
The E-1 and E-300 measure the light
off the viewing screen. The screen receives less light than the image
sensor, so the camera assumes the scene to be darker than it actually
is — and overexposes. At smaller apertures the whole light beam fits
within the mirror, and the effect does not occur. Simple.
The Zuiko Digital lenses, designed for
small-sized sensors, do not cause this problem, as the light beam is
already narrow before it strikes the mirror. While Olympus was touting
this feature in promotional materials, stressing its importance for the
performance of sensor’s photosites, it also has this extra effect.
I have noticed that the autoexposure
error also depends of the diameter of the rear lens element; the larger
it is the more AE problems you may expect, ant those may be not limited
to just the widest apertures.
The good news is that, for a given lens
and aperture, the over- or underexposure seems to be constant, not
depending on any other factors (a slight dependency on focused distance
can be safely neglected). Therefore the workaround is simple: for every
lens you need to know the exposure correction, applicable for every
aperture. Sometimes this will kick in only at widest lens openings,
sometimes it will keep changing in the whole aperture range. This
correction (applied on top of any others you may want to use, of
course) can be established experimentally by taking a series of test
pictures at varying compensation values; actually this is faster done
than described.
For example, the aforementioned 1.8/50
OM Zuiko requires just -0.3EV compensation at F/1.8 and no compensation
at smaller apertures (i.e., F-numbers from 2.8 upwards), as tested on
the E-300 with matrix metering.
On the other hand, with the classic
Zeiss 1.5/75 Biotar, the E-500 very slightly overexposes at F/1.5,
stays even between F/2 and F/4, and then overexposes increasingly up to
almost 1 EV at the lens fully stepped down to F/16. (This was checked
with center-weighted metering.) In this particular case I have set the
WB by reference before the first exposure, and the colors were
consistent in the whole aperture range. By the way, the lens is
wonderfully soft at F/1.5, and this, together with the shallow depth of
field, makes it great for portrait shooting.
To be exact, the correction depends on
mirror size. If the E-1 has a larger mirror than the E-300 or E-500, it
will require less adjustment. I haven’t measured the mirrors in these
cameras, therefore I cannot drive any conclusions here.
The above discussion is based on an
educated guess, and consistent with my own results (E-300) and those
obtained by a number of people I know, most prominently John Foster
(E-1), who checked a wide range of OMZ lenses. At the same time, I
received a set of comprehensive image samples by Peter Klein (E-1), and his results seem to show the opposite trend.
According to Olympus, spot metering will also be unreliable with
non-4/3 lenses; use the ESP or center-weighted mode. Actually, having
used both these modes on the E-300 and E-500, I can say this still
depends on lens and aperture used; this is not just a problem of spot
metering.
White balance
Similarly to the exposure, the auto WB system in E-series cameras goes
bananas at the widest apertures, especially for short-to-medium focal
lengths. With many of my lenses, especially up to 75 mm, it provides
results which are clearly off-mark, mostly in the yellow direction. The
effect seems to go away aft F/2.8 or so.
I would recommend, therefore, using
manual WB with manual lenses fitted on Four Thirds cameras. Actually,
I’m in favor of using manual WB in almost all kinds of shooting,
regardless of lenses, so this will not be a problem for those who
follow this advice.
Focusing
As already mentioned, in addition to being obvious: you will be limited to manual focusing only, period.
The Olympus E-Series cameras have a manual focus confirmation feature:
the AF indicator blinks when the AF sensors determine the image is in
focus. Unfortunately (and ironically!), this feature works only with AF
lenses, so you will not get any help here.
Digital SLRs, as a rule, have finders
with much less magnification than film cameras. This is also true about
Olympus models. While in the E-1 and E-300 the image size is exactly
the same as in models from other makers using APS-sized sensors (except
for Pentax, offering slightly more magnification), in the E-500 and
E-330 the size is an extra 10% smaller. Either way, this makes manual
focusing problematic, at least for me.
While an eyepiece magnifier or a magnifying angle finder makes manual focusing easier, all this makes me rather skeptical regarding the usability of MF lenses on digital SLRs.
More and less useful lenses
Remember the 2× focal length magnifier. This means that your prized 14
mm lens will have an EFL of 28 mm on the Four Thirds body; nothing to
write home about, with all limitations and inconveniences discussed
above.
The widest-angle lenses are quite
difficult and expensive to design and make. They also may have to
sacrifice some of the central sharpness to achieve more acceptable
results on peripheries of the full frame. They may also show more
distortion, even in the central area used by the Four Thirds frame.
This may mean that such a lens, while
working OK on your E-System camera, may provide results not better than
your standard zoom set to 14 mm. If I am skeptical about using
wide-angle lenses from film cameras on Four Third bodies, that’s why.
On the other hand, a “normal” (50 mm)
lens will become a moderate telephoto (EFL of 100 mm) on your digital
camera; these lenses also offer widest apertures and usually, the
highest resolution. You may obtain surprisingly nice results.
Last but not least, long MF lenses may
simply have no equivalent (at least not in a non-exotic price range) in
the Four Thirds mount, while still providing the necessary extra
resolution, at least when stepped down. They can also often be bought
quite cheaply, therefore you will probably be using them regardless of
what I may say.
Advice
Adding all this up, here is some advice for those who would like to use
their “older” lenses on Olympus E-system cameras.
- Carefully select the
lenses you want to use; choose only the sharpest, as they will have to
provide twice the resolution for which they were designed. This usually
means limiting yourself to non-zoom lenses. (Sharpness is not all; if
you like the results, use the lens anyway!)
- Make sure that you can live with stepped-down
aperture priority metering. If not, forget it. Remember that most
lenses will perform best (here: satisfactorily) when stepped down one
or two F-stops from widest aperture.
- For every lens shoot a series of test images
under controlled conditions (preferably artificial light), at various
apertures. For widest apertures the images may be overexposed. At these
apertures apply negative exposure compensation until the image matches,
exposure-wise, those taken with aperture closed down; write down these
corrections and use them as needed.
-
Use manual white balance setting (which is generally a good idea), especially when shooting at wide apertures.
Have fun.
by J. Andrzej Wrotniak
|