full frame digital backs for Contax 645

[MarkGamba]

Hi,
Old school guy here with no digital cameras at all and looking at going digital only for new client.  My biggest complaint with digital is that most of the cameras out there seem to be cropped versions of their previous formatt.  To my (very limited) knowlege only Canon offers a camera that is full frame.  I am hoping there is a digital back out there that I am unaware of for my Contax that is full frame.  I shoot wide angles alot and can't afford the cropping.
Thanks for any help.

All the best,

Mark Gamba

[BJL]

I am hoping there is a digital back out there that I am unaware of for my Contax that is full frame.  I shoot wide angles alot and can't afford the cropping.
Thanks for any help.
[a href=\"index.php?act=findpost&pid=52487\"][{POST_SNAPBACK}][/a]

The largest sensors available for medium format backs are about 37x49mm, so about 61mm diagonal, still a bit of a crop relative to the 42.5x56mm (70mm diagonal) of the Contax's 645 format. More to the point, 35mm is the shortest focal length for Contax 645 or any MF system, about equal to the short dimension of the frame, so only as wide as about 24mm in 35mm format, 16mm in Nikon DX format, 15mm in Canon EF-S, or 13mm in Olympus FourThirds format.

All of those smaller DSLR formats offer far wider angular coverage options: Canon 16-35mm or 14mm for 35mm format, Nikon 12-24mm for DX, Canon 10-22 for EF-S, Olympus 7-14mm for FourThirds. And third party zoom lenses for DX and EF-S reach 10mm (Sigma) or 11mm (Tamron).

To match the widest of those, 14mm with Canon 35mm or 7mm with Olympus FourThirds, would require about 20mm with current MF backs, and 22mm even with a "full frame" 645 sensor: way beyond the current 35mm lower limit. So unless new shorter focal length "digital specific" ultra-wide angle lenses are released for medium format digital, MF trails the smaller digital formats badly for extremes of wide angle coverage. For the discontinued Contax in particular, such new lenses are unlikely.

[Zuikoholic]

Yes, but "in time" (however long that turns out to be) full-frame backs for 645 and 6x6 will become available. If they weren't heading in this direction, they would have abandoned MF before even starting with digital backs. I won't be buying anything below full-frame for my 6x6 Hassy, but it's going to be a long wait before they are produced and then drop in price enough for me to afford...
That's OK. I have waited for full-frame in 35mm format and now have myself a Canon 5D (which is full-frame) - so patience is needed!

[Anon E. Mouse]

If it archtectural work you do, Horseman has just come out with a SW-D with shifts and a 25mm lens which gives quite a large angle of view for a medium-fomat digital back. I think Cambo has a similar model.

[Anon E. Mouse]

Horseman SW-D

[BJL]

Yes, but "in time" (however long that turns out to be) full-frame backs for 645 and 6x6 will become available. If they weren't heading in this direction, they would have abandoned MF before even starting with digital backs.
[a href=\"index.php?act=findpost&pid=52955\"][{POST_SNAPBACK}][/a]

Keeping MF in production is perfectly compatible with another approach of a slight downsizing of format accompanied by a modest additions to the wide angle lens line-up.


6x6 sensors are unlikely ever to be developed, as MF development has moved almost entirely to 645 rather than the larger 6x6, 6x7 or 6x8 formats. The only recent 6x6 efforts are very weak AF offerings from Rollei. Hasselblad more or less stopped developing its 6x6 format prodicts about a decade ago, and has now discontinued the majority of their 500 bodies and all of the 200 system. Most of what they do now is market rebranded Fuji 645 AF bodies and lenses and Phase One backs as the "H" series.

It also seems unlikely to me that MF will upsize sensors another 40% in area to fill the 645 film frame. A far easier and more cost effective solution seems to be simply add one or two more lenses like an f=30mm, to recover the wide angle coverage they currently have with the shortest lenses for 645 format, f=35mm. Downsizing the format a modest 15% linear from 56x42mm to about 48x36mm keeps existing lenses, bodies and such perfectly usable, and keeps the sensor price gap between MF and 35mm format from being far greater than it already is. This shift seems more attractive for a medium format maker than abandoning all medium format assets entirely. Pentax seems to be aiming a bit smaller still, planning an integral digital body based on a 33x44mm sensor. Again, one new prime lens around f=28mm would fix any loss of wide angle coverage, and maybe one new zoom, since the Pentax MF system has an extensive range of zooms.


My prediction: square MF is a dead man walking: get over it! (As Michael might say.)

[BJL]

Horseman SW-D
[a href=\"index.php?act=findpost&pid=52959\"][{POST_SNAPBACK}][/a]

Indeed! To quote from that page:
"The 24mm lens unit in combination with a 38 x 48mm image sensor gives you the same wide angle as a 17mm lens with the 35mm format."
That is, wider than medium format film SLR's ever offered as far as I know!

And the idea of bringing the advantages of a view camera down to match the largest single shot sensors could be a useful additional raisond'etre for these very large, very expensive sensors, and be the future of digital view cameras. (Or does someone want to predict affordable 4"x5" sensors?)

[Zuikoholic]

The 6x6 sensor is not so much bigger than the 645 sensor. It makes a big difference today because of the relative number of useless sensors made in each batch (perhaps only 1 in 10 645 sensors are viable, compared with say 1 in 15 or 1 in 20 6x6 sensors). As sensor technology improves, the viability ratio of 645 to 6x6 could well change, making 6x6 drop in price realtive to 645 (6x6 will still cost more than 645 of course, but instead of say 5x the price, it might drop to 2x the price).

I believe that 6x6 sensors will ultimately be produced (yes, it is a matter of faith on my part, but I also had faith in FF cameras being produced in 35mm format dSLRs when everyone was saying "Get used to APS sensors as they'll never make FF sensors cheap enough").

If they really cannot produce 6x6 sensors then I would settle for 5x5. I think the square format is much much better than rectangular, particularly because it eliminates the need to turn the camera on its side to shoot vertical images. It would be a very great pity if they abandoned the square (I'm sure they won't: if Hasselblad does abandon 6x6 then another comapny will step in and offer 6x6 digital backs and they will sell those units to people like me; Hasselblad will lose sales by going down this path - they are there now because 6x6 sensors are not yet economic, but like the APS sensors, 645 will be a temporary phase: why? because we can still get 645 by cropping from 6x6).

I am prepared to wait it out. I am confident that 6x6 sensors will be produced: not for my benefit, no way! They will produce and sell 6x6 because if they don't, someone else will (and unlike the situation with the 35mm format dSLRs, medium format can use separate digital backs from different manufacturers, and so medium format is not so much at the mercy of the manufacturers as is 35mm digital).

It might take 5 years to produce 6x6; it might take another 5 years for it to become affordable... but I can wait.

[BJL]

Lately I am trying hard to avoid debating the backward looking wishful thinking of some people that camera technology will reverse all recent clear trends in order to revive the formats and technologies of the past. That includes fantasies of significantly market share recovery and product development efforts for film, moderate wide to normal primes, square formats, formats larger than 645, and for that matter, 645 format, 48x36mm digital format, 36x24mm (35mm) digital format, and perhaps even all digital formats larger than 2/3" (the great majority of digital cameras sold are now in formats smaller than 2/3", so under 1cm diagonal: even Four Thirds is relatively huge.)

When some people refuse to see that photographic progress is dominated by improving the resolution, sensitivity and dynamic range of a given format and pixel size, combined with sometimes using this to allow the shift to smaller formats and pixel spacing, I suspect that mere facts and reasoning are not going to change that thinking.


But one fact: the 6x6 format of 56x56mm is a substantial 80% larger in area than the current 645 sensors of around 36x48mm, and making 56x56mm sensors would probably require going beyond the already difficult double step masking (used by Dalsa at least) to perhaps quadruple step masking for each sensor. Such dreams might be technologically possible at sufficient cost (both development and unit price), but it is far, far easier to instead refine lens and body offerings to work better with the new "materials". Like Rollei and Hasselblad did with the new smaller roll film formats, and Leica did with the newer, smaller 35mm film. Especially since lenses are more and more the main limit on image quality.

[Zuikoholic]

Lately I am trying hard to avoid debating the backward looking wishful thinking of some people that camera technology will reverse all recent clear trends in order to revive the formats and technologies of the past...

When some people refuse to see that photographic progress is dominated by improving the resolution, sensitivity and dynamic range of a given format and pixel size, combined with sometimes using this to allow the shift to smaller formats and pixel spacing, I suspect that mere facts and reasoning are not going to change that thinking...

... Especially since lenses are more and more the main limit on image quality.
[a href=\"index.php?act=findpost&pid=53156\"][{POST_SNAPBACK}][/a]

Thanks for trying :-)

I am quite sure costs of large sensors will be high. I'm also sure that costs for all sensors will come down, and I suspect that the costs for larger sensors will come down faster than the costs for smaller sensors come down (but I admit that is speculation on my part - I'm willing to wait and see, and if I'm wrong then I save my money for something else - no big deal then, is it?). Manufacturers will only produce a 6x6 sensor if they can sell it at a profit. As I have said years back regarding the FF 35mm sensor, If they make one for sale, people will buy it. Canon are making large sales now on the 5D - people do actually want FF sensors! Most folk who express their opinion say that a FF sensor is a waste of time... but it seems that many of those who do not express their opinion are buying up these cameras anyway.

If a 6x6 sensor can be made and sold for a reasonable price, people will buy it. The pixels don't need to be as small as those on the smaller sensors, and who would want that? Larger pixels would give lower noise and better dynamic range, and existing lenses would work fine on them. There is no problem with this idea except for price.

When the full-frame Canon EOS 1Ds came out, it was $14,000 USD I seem to recall - correct me if I'm wrong. Its replacement, the 1Ds II, was $8,000 USD. And now the 5D is around $3,000 USD. All with sensors people said would never be made because there was no point and nobody would want them anyway. Never have I been so glad that probably none of the opinions expressed on sites like this are used by manufacturers to direct their policy on development. No doubt this will remain true. No-one will take what you say, or what I say, and use it to decide where their company is going in the next 5 years. If they can make a 6x6 sensor affordable then they will do so as they will sell like hotcakes. If they cannot, then I never could have afforded one anyway, so no big deal.

Up till now, sensitivity and dynamic range seem to be improved by bigger pixels, not smaller. Bigger pixels are most acceptable on bigger formats as resolution per mm is not required to be so high. So, I think that cost to produce is the only sticking point in the idea of a 6x6 sensor. Once it has been produced, it will no doubt give much better images than any smaller sensor. That alone does not mean that it will be produced - economics will determine that.

As for lenses, the old manual focus lenses of the 1970s have shown themselves to be up to the task of imaging onto todays best sensors. Therefore, all that is required is for the lens manufacturers to get their new lenses back up to the quality of the 1970s and this problem is solved (it is already solved if you can put up with stop-down metering). I'm sure they've got their old designs hidden away somewhere, and can dust them off and produce new AF versions of those wonderful lenses, if they wanted to.

Feel free to nay-say this as much and as often as you like. As I'm 100% sure that the people who make the decisions will not care what you say (nor what I say) then either of our predicitions are irrelevant. What will be will be regardless of what we say here and now. I'll be saving up so that I might actually be able to afford the 6x6 sensor when it's put on the market. You don't appear to want one, so it won't worry you whether it is ever made or not. Peace be with you; let's take some photographs!

[jani]

I am quite sure costs of large sensors will be high. I'm also sure that costs for all sensors will come down, and I suspect that the costs for larger sensors will come down faster than the costs for smaller sensors come down (but I admit that is speculation on my part - I'm willing to wait and see, and if I'm wrong then I save my money for something else - no big deal then, is it?).

You're most likely wrong.

I've explained this briefly in another post around here somewhere, but let me try to do this a bit more detailed:

Microchips (including CMOS imaging sensors) are made from silicon wafers. Canon uses, like everyone else, circular wafers. Canon's wafers now have a diameter of 300 mm.

If the chip you want to make is 56 mm x 56 mm, then you can make a total of 11 such chips on a wafer as long as the die spacers and margins are less than 2.3 mm. If these are greater, then you can only make 9.

If your chips are 48 mm x 36 mm, the numbers are:

0 mm spacer and margin: 24 (not likely)
0.3-1.0 mm: 23
1.1-1.8 mm: 22
1.9-2.3 mm: 21

It doesn't take much imagination to see that the bigger chips must be made in half the quantity of the smaller chips, and that this results in a significantly higher cost for the bigger chips.

Your chip manufacturing process may be good enough to avoid a significant difference in defects per wafer, but the number of chips you can make at a time is severely limited with bigger chips.

If you want to play around with the numbers and see a nice illustration, you can download free software for that here:

http://www.geek.com/procspec/software/wafer/wafer.htm

[Slough]

You're most likely wrong.

I've explained this briefly in another post around here somewhere, but let me try to do this a bit more detailed:

Microchips (including CMOS imaging sensors) are made from silicon wafers. Canon uses, like everyone else, circular wafers. Canon's wafers now have a diameter of 300 mm.

If the chip you want to make is 56 mm x 56 mm, then you can make a total of 11 such chips on a wafer as long as the die spacers and margins are less than 2.3 mm. If these are greater, then you can only make 9.

If your chips are 48 mm x 36 mm, the numbers are:

0 mm spacer and margin: 24 (not likely)
0.3-1.0 mm: 23
1.1-1.8 mm: 22
1.9-2.3 mm: 21

It doesn't take much imagination to see that the bigger chips must be made in half the quantity of the smaller chips, and that this results in a significantly higher cost for the bigger chips.

Your chip manufacturing process may be good enough to avoid a significant difference in defects per wafer, but the number of chips you can make at a time is severely limited with bigger chips.

If you want to play around with the numbers and see a nice illustration, you can download free software for that here:

http://www.geek.com/procspec/software/wafer/wafer.htm
[a href=\"index.php?act=findpost&pid=53331\"][{POST_SNAPBACK}][/a]

He might be right. The cost of the larger sensor is dependent on two factors, namely the increase in real estate, and the increase in the number of failures. I beleive the main factor for large sensors is the large number of failures, which is why they are very expensive. (The same number of dud photosites can in the worst case lead to 50% of APS sensors failing, but ALL FF sensors failing, for the same die size.) So if the failure rate comes down significantly, the result will be a disproportionate decrease in the cost of large sensors.

Of course, this assumes a large reduction in the failure rate which is no more than complete speculation. But the trend has always been for the price of an item to drop rapidly as manufacturers pay off research costs and improve manufacturing methods. Look at TFT screens where 20" is now commonplace.

But until then we will be in the current "my sensor's bigger than your sensor" situation.

Leif

[BJL]

OK, I will struggle again to stay with factual corrections, rather than further debating the futility of making sensors larger than any up-to-date SLR system, which top out at the Hasselblad, Fuji, Mamiya and Pentax 645 AF systems.

Canon are making large sales now on the 5D - people do actually want FF sensors!

"Large" meaning a production rate of about 10,000/month I believe, compared to about 2,000/mo for the 1Ds, 4,000/mo for the 1D, and 115,000/mo for all Nikon DSLR sales according to their most recent six months sales report. So I would guess about 250,000/mo or more total DSLR sales. The 5D is a relative success for a camera costing about twice as much as a Leica, but still a  niche product compared to DSLR's as a whole.

Most folk who express their opinion say that a FF sensor is a waste of time ...

That is a straw man: I and most "sensor upsizing skeptics" have not said that Film Format (FF) is a waste of time, just that it will always be a far smaller niche than the new, dominant digital specific SLR formats (and MF will be an even smaller niche, and ...)

When the full-frame Canon EOS 1Ds came out, it was $14,000 USD I seem to recall - correct me if I'm wrong. Its replacement, the 1Ds II, was $8,000 USD.

Way off on the original price. Canon first announced the 1Ds at US$9,000, but cut it to US$8,000 when actually released, and that has continued to be the list price for both versions of the 1Ds, though street price has eased down a bit, maybe 10%.
So your claimed rapid downward 1Ds price trend is non-existent.

As I'm 100% sure that the people who make the decisions will not care what you say (nor what I say) then either of our predicitions are irrelevant.

Exactly: predictions based on talk about the advantages of bigger sensors and bigger pixels falls flat in face of the clear fact that the decision makers are going in a different direction: the size, weight and cost advantages of improving performance and reducing pixel spacing at each of the current DSLR formats, from 13.5x18mm (FourThirds) up to 37x49mm.

[Zuikoholic]

OK, I will struggle again to stay with factual corrections, rather than further debating the futility of making sensors larger than any up-to-date SLR system, which top out at the Hasselblad, Fuji, Mamiya and Pentax 645 AF systems...

If a mythical 6x6 sensor does become available, I will buy it; if it does not, I won't. My concern is not being the one who was right all along, my concern is acquiring equipment which I think is worthwhile having (as long as it's acquirable). For the manufacturers to force us to put up with small sensor sizes is the best argument I've seen so far for going back to film. But let's wait and see shall we.

PS., I think you are right about the prices I gave being too high - that was my mistake.

[BJL]

If a mythical 6x6 sensor does become available, I will buy it; if it does not, I won't

That makes sense, with a qualification about price: I estimate that a 6x6 MF back would currently cost US$100,000 or more. Sensors in 6x6 format and larger are made for special markets by companies like Honeywell and Dalsa; Michael shows a sensor that looks about 6x6 size in his story about Dalsa at http://luminous-landscape.com/essays/dalsa.shtml

... For the manufacturers to force us to put up with small sensor sizes ...

There is no evil conspiracy of manufacturers forcing us to buy smaller sensors than we want, keeping the wished for larger formats off the market or artificially over-pricing them; it is a matter of customer choice based on price and performance differences between the different formats.


In medium format, sales are way down, with financial troubles almost throughout the industry. Clearly it would be in the interests of medium format makers to attract more customers with reduced prices, or to offer the much wished for larger sensors formats like full 645 or 6x6, if they could make an adequate profit on it.

And yet no one in the MF industry is doing so, despite the current ability of suppliers like Dalsa to make sensors 6x6 or larger. The only explanation I can see for the absence of digital backs in formats larger than about 36x48mm is the uniform judgment of the MF industry that demand would be insufficient, due to the combination of the extremely high price of those larger sensors and the excellent quality given by somewhat smaller, less expensive formats.


Even down at 35mm format, 24x36mm, digital sales are a tiny fraction of what they were with film, even with the efforts of SLR industry leader Canon. Even with the 5D, 24x36mm digital sales are only a few percent of what 35mm film SLR sales used to be, only a few percent of total current DSLR sales (estimated at 4 million per year in the latest Nikon financial statement), and barely more than one thousandth of total digital camera sales, whereas 35mm had a great majority of all film camera sales.


Photographers in general choose smaller formats for digital than for film, due to both smaller differences in quality and larger differences in price when you compare the same two formats in digital rather than film.  (The 2/3" digital sensor format is only 6.6x8.8mm. Imagine comparing 8"x10" prints from a tiny 6.6x8.8mm piece of film to ones from 35mm film: the quality difference would be far greater than comparing such prints from 2/3" to 24x36mm digital formats.)

[Zuikoholic]

Photographers in general choose smaller formats for digital than for film, due to both smaller differences in quality and larger differences in price when you compare the same two formats in digital rather than film.  (The 2/3" digital sensor format is only 6.6x8.8mm. Imagine comparing 8"x10" prints from a tiny 6.6x8.8mm piece of film to ones from 35mm film: the quality difference would be far greater than comparing such prints from 2/3" to 24x36mm digital formats.)
[a href=\"index.php?act=findpost&pid=53487\"][{POST_SNAPBACK}][/a]

I have no argument with this - I never said that smaller sensors were a bad thing in general. They have a benefit for long-lens photography, which is why I'm keeping my 10D. OTOH, they are not so good for wide-angle photography. OK, new lenses with smaller image circles can, and have, been made to address this problem.
My personal reason for prefering larger format sensors is DOF issues. I like the DOF associated with the FF 35mm sensor size and to make up for the larger DOF on APS sensors, lenses need to be 1-2 stops brighter - this means that we'd need lenses with a maximum aperture of f/1.0 or f/0.5 - this seems very unlikely to happen in the near future, and even if it did these lenses would be so large and heavy as to make arguments of "smaller & lighter" regarding APS sensor cameras a complete joke (not to mention the probable lousy edge/corner performance of these super-bright lenses).
This, then, was my own personal reason for wanting FF in my dSLR. Canon provided one, and I bought it. This doesn't mean that I'm campaigning for Canon to drop their 1.6x crop range - not at all! I'd be more than happy for them to continue both lines, which is what I believe that they will do. Sales are not so high with FF because these cameras still cost many times more than the top-of-the-range film cameras of the recent past. As costs come down - which they will - more & more people will buy into FF. APS likely won't die out though - why should it? For those who like telephoto lenses, APS are the way to go (even I can see that!).
Now, as for medium format, it has been shrinking (in terms of market share), that is true... but digital has been a major reason for that happening. Supplying digital backs - ones that real people can actually afford to buy - will beging to turn things around. In the same way as with dSLRs, I'm more than happy for smaller sensors to start the ball rolling. But, for me, cropped sensors on M.F. do not have the advantages they do in dSLRs - I'm not interested in shooting long telephoto shots on my Hassy, not one bit! And, as I mentioned about, I am fond of the characteristic DOF associated with 6x6 format, and shooting with a 50mm lens on a 4x4 sensor will give very different results indeed in terms of DOF - unless they are going to provide us with f/1.0 lenses for reasonable prices?
The advantage with medium format is that the digital backs need not be made by the camera manufacturer. I mean, if Canon got into the digital back game, pricing their backs down through profits earnt through dSLR slaes, then this would start the ball rolling. Of course these sensors will be expensive. But the economy of scale will bring prices down. The question remains though: Is digital medium format already dead in the water? If the companies think it is, then it will never happen. If it does get going, prices will come down in time - as the market share is so small at present, prices will come down in price much more slowly than we've seen in the dSLR market. So be it. As long as digital backs stay above my maximum affordable price, I will not buy one. Perhaps I will stick with film for medium format - hopefully, the films will still be available in this case. As far as 35mm format goes, I've given up on film as digital gives me all that I want. I'd like to go digital with M.F. too, but not at all costs, and a 4x4 sensor seems to me to be more of a curiosity than anything else (if 4x4 becomes 'cheap' enough then I probably would buy one because I like gadgets: for me, a 4x4 sensor cannot be much more than a gadget, no matter how well made it is).
So you see, I'm not holding my breath. I'll continue shooting with my 5D (I hardly ever pick up the 10D now - perhaps it's for emergency use only!) and forget about digital M.F. - if it comes on the market in my price range, then I'll look again. Personally, I think we will see it, eventually (yes, that is my personal opinion and mere speculation only) - you see, it doesn't matter to me if I'm right or if I'm wrong. But I'm stating in public that I want a 6x6 sensor. If nobody says anything regarding what they want, the companies might be silly enough to think that nobody wants these things at all, and then we certainly will not see them. It harms nobody when I say that I want 6x6 and that if they provide it at a cost I can afford then I will buy it. If I'm proved wrong, so what? If I'm proved right, so what? Let's wait...

[jani]

He might be right. The cost of the larger sensor is dependent on two factors, namely the increase in real estate, and the increase in the number of failures. I beleive the main factor for large sensors is the large number of failures, which is why they are very expensive. (The same number of dud photosites can in the worst case lead to 50% of APS sensors failing, but ALL FF sensors failing, for the same die size.) So if the failure rate comes down significantly, the result will be a disproportionate decrease in the cost of large sensors.

If the failure rate comes down significantly, it will come down just as significantly for smaller sensors.

It will always be much easier to avoid failures on smaller chips than bigger chips.

Therein lies the rub.

[Jonathan Wienke]

If the failure rate comes down significantly, it will come down just as significantly for smaller sensors.

It will always be much easier to avoid failures on smaller chips than bigger chips.

Therein lies the rub.
[a href=\"index.php?act=findpost&pid=53513\"][{POST_SNAPBACK}][/a]

But if the failure rate is brought low enough, so that 24x36mm sensors had a failure rate <1%, the primary cost difference would become the wafer real estate. A 1% failure rate isn't that much different than .1% failure rate with regard to calculating the final per-unit cost.

Another possible large-format manufacturing strategy would be to make a mosaic of smaller sensors that have photosensors all the way to their edges. The trick with that approach is aligning the individual sensors and balancing their output to achieve consistent color across the entire frame, and eliminating seams.

[BJL]

If the failure rate comes down significantly, it will come down just as significantly for smaller sensors.
[a href=\"index.php?act=findpost&pid=53513\"][{POST_SNAPBACK}][/a]

Prices will come down all around, but by a larger factor for larger sensors.

For example, suppose that only 1 in 4 of 24x36mm (35mm) sensors are usable. (This figure has been floated on the internet for the 5D sensor, and fits loosely with 5D pricing, but do not trust it; this is just an example!)  Doubling sensor size roughly squares the yield, so it would be about 1 in a 16 for the 36x48mm "medium format" sensors. Half as many sensor dies per wafer and only a quarter as many of them usable, so one eighth as many usable sensors per wafer, and about eight times the unit cost.

If processes improve to 1 in 2 usable for 24x36mm, the squaring pattern gives 1 in 4 for the larger format. This would give about half the previous price for the smaller sensor, but one quarter the previous price for the larger sensor, and down to only four times the cost of the smaller sensor.


However, everything I have read suggests that yields for DSLR sensors are far below 100% and are expected to stay there for a long time. So prices will continue to go up far faster than in proportion to sensor area. For example, "APS-C" sized sensors now cost about US$100 to manufacturers (Cypress, new owner of FillFactory, announced one recently at US$90). If prices went up only with area, 24x36mm sensors like that in the 5D would cost under $250 to make, and Canon would be charging a lot less for the 5D.

[BJL]

smaller sensors ... are not so good for wide-angle photography. OK, new lenses with smaller image circles can, and have, been made to address this problem.
[a href=\"index.php?act=findpost&pid=53495\"][{POST_SNAPBACK}][/a]

We seem to agree then that with appropriate lenses, smaller formats are at no disadvantage for wide angle coverage. For a long time, 35mm format has offered wider angle coverage than medium format, with 14mm primes and 16mm zooms matching what would need about 28-32mm in medium format. Going smaller still, the Canon 16-35 is already matched for wide coverage in "APS-C" formats by lenses like the Canon 10-22 EF-S, Sigma 10-20, and Tokina 11-18, and outdone in the even smaller FourThirds format by the Olympus 7-14 ("14-28 equiv.") That last lens also matches the widest 35mm format SLR prime.

My personal reason for prefering larger format sensors is DOF issues.
[a href=\"index.php?act=findpost&pid=53495\"][{POST_SNAPBACK}][/a]

As to getting shallow DOF, 35mm generally matches or outdoes all other formats, larger and smaller. In comparison to medium format, this is due to the typically lower minimum aperture ratios of 35mm format prime lenses. For example, Hasselblad's current square MF lenses are limited to f/2.8, which matches about f/1.4 in 35mm format for DOF.

The extent to which the newer digital SLR formats will match shallow DOF options with the same strategy of shorter, faster primes remains to be seen, but I expect 35mm format to remain the "very blurry background champion".