Somewhat on-topic, Carl Weese wrote a pretty decent post on TOP about printing large from 4/3 cameras. I am guilty of pixel-peeping quite a bit and was fretting over whether 4/3 would be "good enough". I think I can, after reading that, safely say that for my needs, it is.
That's a huge leap. "Just because they don't make it means they can't make it."
It's also a load of rubbish.
Why, exactly, is that?
They can make small compact mirrorless bodies with APS-c sensors at very reasonable prices -- note the NEX-5N. It's not a huge leap to put a bigger sensor in the body.
How do you know that? Why aren't you designing the full-frame mirrorless killer app that will dominate the industry?
According to the only company that's actually produced a full-frame mirrorless, it is a huge leap. According to another company that's put a rear element 5mm from an APS-C sensor, that was also a big leap.
The difficulty comes down to one major obstacle: digital sensors need light to come from a fairly restricted angle of incidence to avoid softness and color shifts and vignetting. With SLRs, this is not such a obstacle, you've generally got a large enough image circle and enough distance between the sensor and rear element to overcome the issues. Mirrorless systems minimize that distance and thus make the image circle relatively smaller. The puzzle is finding an affordable, quality solution to redirect light at the edges.
No one has successfully done that (affordably, Leica has for the price of a used Honda Accord, of course). There is no reason to believe it is easily and affordably surmounted - if it were, someone would have done it, because they'd be Scrooge McDucking in their large piles of money as we speak.
Even if it costs $1000 more to put a FF sensor (more than I think it would cost, IMHO) in a 5N-like body, what makes you think that $1599 isn't a reasonable price for such a body that wouldn't sell at that price? Or even $1999 for NEX-7 like body with a FF sensor?
As long as we're making up numbers, why not $20? $50 with a Sony EVF!
Given that the NEX-7 is, itself $1300, it seems rather presumptuous to assume that you can engineer and produce a body with a sensor 4x larger for $700.
Okay, not so much presumptuous as 'absurd.'
Still, by a large margin, the lowest cost FF camera on the market (Sony a850 discontinued, IIRC -- and it should be cheaper to produce than the a850 since there is no mirror box and pentaprism).
I find it highly doubtful that a mirror box and pentaprism are the driving factors in 36x24 cost, given that high-end APS-C mirrorless bodies are actually more expensive than APS-C dSLRs.
miloz wrote:
The difficulty comes down to one major obstacle: digital sensors need light to come from a fairly restricted angle of incidence to avoid softness and color shifts and vignetting. With SLRs, this is not such a obstacle, you've generally got a large enough image circle and enough distance between the sensor and rear element to overcome the issues. Mirrorless systems minimize that distance and thus make the image circle relatively smaller. The puzzle is finding an affordable, quality solution to redirect light at the edges.
No one has successfully done that (affordably, Leica has for the price of a used Honda Accord, of course). There is no reason to believe it is easily and affordably surmounted - if it were, someone would have done it, because they'd be Scrooge McDucking in their large piles of money as we speak....Show more →
luckily leica told us what their solution is and it's pretty plain that it's not that expensive from what they've said (offset microlenses, no AA filter, thin filters in general, and software corrections). the price of the m9 is not largely due to the sensor just as the cost of the D3x is not largely due to the sensor (which is the same one in the a850).
miloz wrote:
Given that the NEX-7 is, itself $1300, it seems rather presumptuous to assume that you can engineer and produce a body with a sensor 4x larger for $700.
Okay, not so much presumptuous as 'absurd.'
why? all you have to do is put a $500 dollar sensor in and maybe a better processor. do you think the difference in price between the NEX-5N and NEX-7 is due to solely to the cost to produce the sensor?
edit: leica's sensor is almost certainly much more expensive than the one in the a850, but not because it is much more difficult to make – it is more expensive because it is a very low volume product.
pingflood wrote:
Somewhat on-topic, Carl Weese wrote a pretty decent post on TOP about printing large from 4/3 cameras. I am guilty of pixel-peeping quite a bit and was fretting over whether 4/3 would be "good enough". I think I can, after reading that, safely say that for my needs, it is.
miloz wrote:
I find it highly doubtful that a mirror box and pentaprism are the driving factors in 36x24 cost, given that high-end APS-C mirrorless bodies are actually more expensive than APS-C dSLRs.
I didn't say that. I said that a mirrorless FF camera should be cheaper to produce than a FF DSLR becuase of the cost of the mirror box and pentaprism.
LotusM50 wrote:
Even if it costs $1000 more to put a FF sensor (more than I think it would cost, IMHO) in a 5N-like body, what makes you think that $1599 isn't a reasonable price for such a body that wouldn't sell at that price? Or even $1999 for NEX-7 like body with a FF sensor?
miloz wrote:
As long as we're making up numbers, why not $20? $50 with a Sony EVF!
Given that the NEX-7 is, itself $1300, it seems rather presumptuous to assume that you can engineer and produce a body with a sensor 4x larger for $700.
Okay, not so much presumptuous as 'absurd.'
Wow. Totally missed the point. Let me restate it to be clearer. How much more does a FF sensor cost compared to a APS-c sensor? I picked a round number that is probably higher than the actual cost. The NEX-7 cost $1199 in the US. IF the cost of a FF sensor is $700 or $1000 more than an APS-c sensor, then the cost of a FF NEX-7-like mirrorless is about $2199. That number is very reasonable for a FF camera, and it will sell. Do you have an actual figure? Or do you think that a FF sensor costs $2000 more than an APS-c sensor? (that can hardly be true since FF DSLR's can be had for around $2499) A FF mirrorless will have some advantages (notably size) over comparable FF DSLR's (which currently cost $2499+). What other costs would have to be significantly higher? sebboh suggests maybe the processor. OK that might add a few dollars. The cost to engineering and design it about the same (really people know how to deal with FF sensors now, and much of the bits are off the shelf). Sure it will be a little bigger (+10mm higher + 15mm longer, due to the sensor, larger lens mount due to larger lenses, perhaps a bit thicker to handle a potentially longer lens registration distance, but not much else is all that different -- a NEX-7, for example, already processes 24mp of data). There is nothing that makes a FF mirrorless expensive compared to a APS-c mirrorless other than the sensor, and they CAN be produced for a reasonable, competitive price.
So much speculation. In a few days we'll find out a lot more about what exactly is the OM-D and more info about the sensor. Yes, many would like to see the built in EVF but I think most are anxiously awaiting to hear more about the sensor. From the rumors I've read, it should satisfy the cleaner IQ requirement and better dynamic range at low ISO, but what I can't decipher is how well it will do at high ISO, but in a couple weeks we'll know more and hopefully in a couple months we'll know for sure
I picked a round number that is probably higher than the actual cost.
Yes, you picked - you made up a number. Which may be related to the basic difference in the price of Sensor A and Sensor B, but ignores everything else about the difference in body design.
I'll say it again: if it were easy to produce an affordable full-frame mirrorless with compact lenses, why isn't anyone doing it?
Let's say NIkon and Sony are terrified of sacrificing DSLR sales (which is a questionable assumption - they stand to make excellent money selling a new line of FF lenses designed for the mirrorless mount) - why, then, did Samsung and Fuji produce APS-C systems? Why have Olympus and Panasonic ardently stuck to m43? There's a chance for four players to steal Canikon's thunder and establish a beachhead on a new and (presumably) popular market.
But none of them did, even when they didn't have established lines to protect. Must be a conspiracy.
miloz - look how long it took Nikon to get a FF sensor, they held out for quite some time and even told their users they were committed to APS-C. Something is amiss, why the hesitancy to go FF, maybe the fear they wouldn't recover on the R&D costs?
miloz wrote:
I'll say it again: if it were easy to produce an affordable full-frame mirrorless with compact lenses, why isn't anyone doing it?
Conservative marketing managers may be the answer to the first question. I mean, just look at Nikon's and Canon's answers to the growing MILC market. One chose to use a tiny sensor and the other chose to give up the IL capability altogether. Yikes!
OTOH, compact lenses for FF offer a true technical hurdle as the light circle is the same for a given sensor size. Of course, one may want to adopt RF designs but then, AF in such lenses is also another technical hurdle.
miloz wrote:
I'll say it again: if it were easy to produce an affordable full-frame mirrorless with compact lenses, why isn't anyone doing it?
the market for such a camera is perceived by camera manufacturers as quite small. the market for FF cameras themselves is quite small, or to put it another way, the market for $2000+ cameras is quite small. most of the people who are in the market for a $2000 camera don't care that it's small or they even want it to be big so they look 'professional'. sony has basically said they don't think FF is necessary or profitable given current technology (all cameras not just talking small cameras), but they realize they need to have a FF camera option in order to be taken seriously in the marketplace.
to summarize: the reason nobody is doing it is because they don't think there is a market for it, not because it's terribly hard. perhaps the success of the x100, NEX-7 (is it a success - to soon to tell?), and excitement over the new fuji will convince camera companies that there actually is a market for such a camera. (knock on wood)
sebboh wrote:
to summarize: the reason nobody is doing it is because they don't think there is a market for it, not because it's terribly hard. perhaps the success of the x100, NEX-7 (is it a success - to soon to tell?), and excitement over the new fuji will convince camera companies that there actually is a market for such a camera. (knock on wood)
Good point sebboh, I think this is exactly what happened with the original 5D, it broke the $7000+ barrier for FF and made it "affordable" and therefore, even though it had its shortcomings, it was a huge success.
I'm not saying you're wrong, but I've never seen evidence that it is "significantly" more, do you have any links to resources that show evidence of this?
wjmeyer wrote:
But do they really? I am just wondering if that is "assumed" more than known.
roughly an order of magnitude more. last i heard aps-c cost ~$50 to produce and FF ~$500.
edit: i'll have to look for the links, but i've seen it explained from the fabrication perspective a number of times. basically doubling the area that needs to be up to spec means a drastically higher percentage of subpar chips per wafer (and of course only half as many chips per wafer). in addition, if i recall correctly, FF chips are too big to be made whole using normal etching processes, so they have to kinda be stitched together (though on the same silicon).
Do you have a source you can point to regarding this information? One would assume that R&D would be the single most expensive part of making a camera and sensor production cost would be relatively the "same" regardless of size. I'm a computer scientist/software engineer and was thinking of the possibilities regarding production of medium format cameras few years back. I was in contact with Kodak for sensor prices at the time and larger, higher megapixel count sensors were not exponentially more expensive than the lesser ones and in no relation to price difference between the cheapest and the most expensive medium format cameras. When you buy an expensive camera it is expensive because the manufacturer expects to sell the item in low volumes. But theee are just my assumptions from my work and not facts.
Volume = low price -> R&D costs relatively less per unit sold
hauxon wrote:
Do you have a source you can point to regarding this information? One would assume that R&D would be the single most expensive part of making a camera and sensor production cost would be relatively the "same" regardless of size. I'm a computer scientist/software engineer and was thinking of the possibilities regarding production of medium format cameras few years back. I was in contact with Kodak for sensor prices at the time and larger, higher megapixel count sensors were not exponentially more expensive than the lesser ones and in no relation to price difference between the cheapest and the most expensive medium format cameras. When you buy an expensive camera it is expensive because the manufacturer expects to sell the item in low volumes. But theee are just my assumptions from my work and not facts.
Volume = low price -> R&D costs relatively less per unit sold
this is only partially true. higher pixel densities don't cost anymore to produce than lower ones assuming the same process. for any given etching process the cost to produce increases exponentially with the size of the chip though.
hauxon, that is what I was thinking as well - supply and demand can affect price more than anything else. They sell a lot more units with APS-C sensors than they do FF sensors.
