Monday, 2 June 2008

Is it really 5 frame per second?

The story began last Thursday. I have noticed a strange thing when I was trying my friend's new Zuiko Digital 14-35mm 1:2.0 lens. After taking the first pictures I could hear that the sound of the moving mirror is slower at 14mm focal length than at 35mm. To verify that I heard I have performed a brief acoustic analysis. It is a high precision method to determine the speed of sequential shooting (based on the visual analysis of sound waves). The results of the analysis is as follows:

Olympus E-3 + Zuiko Digital 14-35mm 1:2.0 @ 14mm f2.0 1/2000 sec.: 4.3 fps - mp3

Olympus E-3 + Zuiko Digital 14-35mm 1:2.0 @ 35mm f2.0 1/2000 sec.: 5.0 fps - mp3

It is sounds like madness, but the owner of the lens gave an idea what causes this phenomenon: the Zuiko Digital 14-35mm has very large rear lens-element, and it is closest to the image sensor when the lens is set to 14mm. The light-beam in the mirror box is the thickest in this case because the diameter of the rear lens-element is larger than diagonal size of the sensor.

I made an animation based on CAD drawing from "Passion for the best - episode 16." and my measurements. The maximum error in the size or position of the elements is 0.5mm.

Olympus E-3 + Zuiko Digital 14-35mm 1:2.0 @ 14mm

It is clearly visible that the mirror has to reach its highest position to pass the light onto the sensor when the lens is set to 14mm. At first we thought that the speed difference is caused by the fact that the mirror has to travel a longer distance than in any other case, but now I think that it is the timing of the shutter causes the difference in fps. Look at the following animation:

Olympus E-3 + Zuiko Digital 14-35mm 1:2.0 @ 35mm

In the first case (14mm) the camera has to wait for the mirror to reach its highest position, and it can open the shutter after that. The mirror can return after the exposure is taken. In the second case (35mm) the shutter can be opened significantly sooner, and although the mirror has to go right up to the highest position, it can do so during the exposure. This is of course only one of the possible reasons. Our first thought might have been correct as well. The traveling direction of the focal plane shutter makes the problem more complex.

I never use my camera in high speed sequential shooting mode so I don't want to draw a heavy-handed conclusion. The decrease of the speed is not published by Olympus, so it would be worth repeating the experiment with other lenses too. The Zuiko Digital 35-100mm 1:2.0 for example also has a very large rear lens-element but I can't try it right now. If the difference between real and official data is more than 10% (0.5fps), it can be critical because that lens is often used for action photography. If I remember well, its rear lens-element does not move behind the plane of the bayonet so there is probably no problem. I have measured the speed of the camera using the Zuiko Digital 35mm macro, the ED 50mm macro and ED 50-200mm SWD attached. All performed around 5.0 fps

Some thoughts about the lens itself:

Its building quality is excellent. The extending of the tube is negligible, there is no wobbling. The moving of focus ring is very smooth and the feeling is similar to "focus by wire" (which I find excellent). The zoom ring is little tighter than on other top-pro lenses but there is nothing to worry about.

Olympus E-3 + Zuiko Digital 14-35mm 1:2.0 @ 24mm f2.0

100% crop from previous picture

I have not taken enough pictures to judge its optical quality. I think this lens is optimized for shorter subject distances. Portraits and object-photos were very sharp even at F2.0 but when I focused to far subjects the pictures were not as perfect as I expected. The most conspicuous improvement on previous top-pro lenses is the extreme flare-resistance. You can take photos with the Sun in the background and the contrast will hardly decrease. Both types of chromatic aberrations are low. And yes, it is very large and heavy.