Sunday, 5 October 2008

On Olympus strategy and conspiracy - and the final question

In previous posts, we have written about the initial thoughts on the introduction of the Micro Four-Thirds format and the peculiarities of the strategy Olympus seems to follow. Back in mid September, also we could only speculate on ideas that the dpreview.com interview for example made clear.
  1. We did not know about the segmentation strategy, but now we know it: regardless of Panasonic, Olympus aims mFT products at the "lower" class, coming "up" from compacts or buying mFT instead of them.
  2. We did not know about chances of using FT lenses on mFT bodies. We still do not know for sure, but it seems that all of them will work, only perhaps slowly. (I'm personally wondering how slow the 12-60mm will be...)
  3. We did not know if the Olympus mFT offering will sport a viewfinder. Saying that the current mockup is a "mid-tier" mFT camera and even this might include a viewfinder, it is almost sure that the "top" tier cam will have one. So we shall have a Leica-killer after all? Hmm, it might be the case.

Besides other bits, we learned what we have already known to some extent: Olympus is very vague on timing and it reserves the right to choose a sensor manufacturer other then Panasonic (Remember the nice low-level performance of Fuji compact chips? Yumm!).

AND finally we heard that while FT is all about compact tele lenses (thanks, Jonas), mFT is all about compact wide-angle lenses. dpreview however could not manage to go into a very important detail of this matter. FT to many is equivalent to not compact, but high quality wide-angle lenses. The 7-14mm is still unmatched except by the Nikkor 14-24mm in many respects, including edge to edge sharpness. This, we were told, is achieved by telecentric lens design (I mean in the case of the Olympus. In the Nikkor, on 35mm full frame? Ehh...). Although telecentricity has nothing to do with flange back distance, ie. it is achievable with mFT, history suggestests otherwise.

Leica M lenses, to name the most prominent example, can achieve high image quality because they use small flange-back distance to get closer to the focal plane and project light at steep angles onto the film. This is the very reason Leica had to employ a shifted microlens design in the M8, which can never be optimal, onle for a single focal length (and limits focal lengths from above). The Leica experience suggests that using the smaller distance this way is not a terribly good idea (but even then, many photographers without MF gear claim that they use the M8 if they want to achieve the ultimate image quality - confusing, to say the least!). If mFT goes this way, it will be about not only compact, but also lesser quality wide angle lenses.

But what do we really know about telecentricity? Not much, and that is coming from Olympus marketing. Do not get me wrong - they bought me with this too. But have you ever compared the rear lens element of the 35-100mm and the 7-14mm, for example? While the former is so big that the sensor fits right "into" it, including IS movements, the latter is distinctly smaller than the sensor. Conclusion number 1: telecentricity is already sacrifised there, in that fabulous lens! Conclusion number 2: telecentricity is not important? Or is it important only to a certain extent? Well, I do not know - this is a secret yet to be explored. But Olympus going the Leica way or not introducing microlens shifting at all will be an indication of either telecentricity or top image quality not being as important as it was designing FT. The former will bug FT users who were believers of telecentricity being important. Such a move would indicate that the relatively bulky wide angle lenses they have to put up with are in fact unnecessary. The latter will not, but buyers of mFT have to be aware that the images will lack corner-border sharpness and overall definition compared to FT images. I am no saying this is always a terrible thing - Nikon users sometimes say (think about the 70-200mm Nikkor, for example): this is no problem as long as you are shooting people.

Given the marketing strategy of Olympus, this might be a good argument after all!

4 comments:

Anonymous said...

The article is very good, but I cannot understand any word from the Leica M part :-)

OPi,

csgzs said...

:) But hey, that is easy. The Leica M system has a small flange-back distance because of the lack of mirror. Leica engineers exploited this by designing lenses that go deep into the camera body - somtimes almost touching the film surface. Think abut it: this implies extremely steep light rays at the edges. But this is a problem with digital, as... you know the rest. The M8 uses shifted microlenses to circumvent the problem, but this works only to a certain extent.

Tobias Mennle said...

Dear Zsombor

I appreciate both yours and Krisztians thoughtprovoking articles in your blog very much. It is a rare find.

About this article. You cited the article mentioned wrongly, FT is about small TELElenses, while microFT is about small Wideangles. Otherwise, I personally do not care about microFT at the moment. I see Olympus is really showing commitment to FT. Of course nobody knows if they will be able to compete in the long run. I hope they will, because most of their arguments seem sound to me, and the lenses are probably second to none.

Some remarks connected with your worries about telecentricity, and some more general remarks:


I don´t know much about the importance of telecentricity and how important it is for quality digital photography, but there are some interesting empirical hints if you have a look at for example High Definition Cinematography (which I do being a cameraman and producer). Professional HD video cameras have 3 chips each being about 4 times smaller than a FT chip or about 16 times smaller than a 35mm Full Frame chip. The high quality lenses though like Fujinon cine style zooms or Canon Primes are incredibly HUGE. I mean, they frighten me just looking at them. One should guess they are made for 6x6 medium format, but no, just a tiny image circle. Same with analogue S-35mm cine film (about half the size of 35mm FF), an Angenieux 25-250mm/2.8 (55 000 Euro) weighs 12kg! Seems if quality really matters, and cost and size do not, lenses become huge.

If you buy into that, what Olympus should do to convince EVERYBODY of the advantages of FourThirds: Give us a model camera and lens in Full Frame and show us the size it would have if they would apply their FT quality standards to the Full Frame chip. Maybe it would be enough to take a Sigma 200-500/2.8, paint it black, put an Olympus sticker on it and have a FT 90-250 in the neighbourhood to be convincing enough... But probably, lenses would even be bigger. I feel the problem of FT little acceptance among aspiring or professional photographers really is that we can only compare today´s FT lenses to today´s FF lenses, the latter being made for yesterday´s image capture with film. If we could directly compare (touch and see) a 300/2.8 FT with a Canon 600/2.8 optimized for FF digital... and of course then compare the image quality... AND the price tag... things would be very different.

Looking at High Definition cinematography may also give us an idea about the reasonably possible future pixel densities on FT chips with today´s technology (although, admittedly, the videocameras have 3 chips, not one). The last two Star Wars episodes were shot on HD with less then 2 Million Pixel in the final images. I thought it looked pretty good on a huge screen. Sony EX1 has a chip size about 1/8th of FT and about 2 Million pixel. Image quality is very good. Translates into 16 Million pixels on a FT chip. But I think there are 4K cameras like the Dalsa which already do about 9million pixels on a chip 4 times smaller than FT, and these cameras are used for Hollywood productions, so this would make 36million pixels in FT. Numbers are not accurate to the point, but a rough and good enough estimate...

Some more thoughts on the future of FT vs. Full Frame and the future of FT (basically what I talk to myself thinking about buying into FT or better go for a 5DII...)

As pixel numbers increase, and pixel sizes shrink, depth of field shrinks, too. That means that todays already limited depth of field in FF will decrease considerably. On the other hand, for people who claim FT has too much depth of field, this situation will become better with new higher resolution chips.

Olympus cameras lack HD video capture, but once they will have it, they should have a much better practical value than a Canon full frame HD video capture because of more depth of field. I personally, having worked with HD video for 2 years now (pixel density and size and circle of confusion similar to FT), have not once had a situation with too much depth of field. The other way round, often.

Olympus cameras lack a time lapse photography feature. Stupid, but should easily be solved.

Ah yes, the lenses. I now mainly use Olympus Zuikos on my analogue Canon EOS cameras simply because they are the best lenses I ever had and used.. And, they are easily affordable. Especially the colour reproduction and separation is amazing. I can even see this on 15x10cm prints. Zuiko 100/2 beats anything else in terms of overall image quality, even the famous Planar 100/2 is a bit disappointing next to it, especially in close up and in colour reproduction (the brilliant Leica 90/2.8 also, let alone EF100/2). Zuiko Macro 50/2 same case - anywhere any sharper 50mm with better colours to be found? Zuiko 28/2 lets you wonder just looking at all the details that pop out until film grain gets too big. I had 3 Nikon wideangles being kicked in the dirt by it. Zuiko 35-80/2.8 is simply a marvel with high contrast and resolution at any aperture and focal length, it beats Canon EF50/1.4 from F2.8 on and at 80mm I doubt there is a sharper lens with skin colour reproduction as brilliant and accurate as this one. Olympus was building some of the best lenses in the world already decades ago. I really like their philosophy of building lenses, it seems to me to be similar to Leica´s, but they were probably always far ahead of the famous Germans.

And FT aspect ratio, to me it seems to be more harmonic and useful than the old Barnack 36x24 thing. Isn´t it ironic that movie making till the 1950s widescreen era was connected to a 4:3 frame?

Best regards,

Tobias

Volker Peter Schenk said...

Very interesting to come across such articles.
I believe mFT is the future. DSLRs just used that historically evolved concept, with all the limitations the neccessity of a mirror implies.
Until now I was successfully experimenting with SONY/Minolta + OLYMPUS FT, using the same lenses, and taking advantage of the different crop factor of the different sensor sizes: A900 1:1, A350 1:1.5, E510&E330 1:2. I use SONY on the wide side (starting at 12mm with SIGMA 12-24mm), FT on the tele side (300mmf/2.8 600mm f/2.8) .
When I started switching to DSLR, I was hopeful, to reduce my camerabag – what a foolish idea, 4 DSLRs, 13 lenses.. I am longing for my old OM 4….
At least with that system, I can focus more on primes and achieve the focal length with the crop factor.
Now, one of my expectations to make the change to mFT has happened with the appearance of the GH-1. My first digital was an RS-100, and the development of the eVF made huge progress…
..I am still waiting for OLYMPUS to offer mFT with IIS – one reason I stay abstinent from Canikon is the lack of IIS, and I do not want to (and do not have the funds to) invest in an IS in each lens I purchase.
I do not know, under the present economic situation, whether we will see that at all. OLYMPUS and PANASONIC seem to have split that business clearly between them; the quick disappearance of the E330 was an indication for me.. For now I hope that the7-14mm f/4 Panasonic will be good, adding the 25mm f/1.4 pancake as substitution for my C550Z streetfighter (f/1.7-Low Light) could leave me with only an E3 + GH-1( or preferable Olympus IIS mFT comparable); the 12-60mm f/2.8-4 Zuiko to me seems like the best Walka- &All- round lens. I use many” legacy” lenses, and the mFT system gives us an opportunity, to use legacy c-mount lenses with an adapter! I just need someone to convince me that I do not need a FF camera, and I am happy, to use the GH-1 as the “wide  mild ” and the E3 as the “mild  wild” system.
www.VolkerSchenk.com, www.alphalenses.com