[Pål Jensen]

If you think FF (which is only in second generation at nikon) will not move down in price and squeeze from above you are sadly mistaken (lack of competition in the market is all that allows it to hold at the moment)

I believe it is sensor cost that prevents FF in being cheaper. I don't think any manufacturer have any interest to sell FF at a loss...

[Pål Jensen]

You make it sound as though anyone willing to pay this price for a camera would automatically go for a "full frame". I believe that is a fallacy: I, for one, wouldn't. At least, not just for the sensor size's sake. And I'm fairly certain that at equal price, an APS-C camera would be better specified than a "full frame".

Thats absolutely spot on. APS is undoubtly the sweetspot for DSLR; both in cost and in quality.

[Pål Jensen]

It doesn't have to be everyone. If the K5 replacement loses 25% of it's volume to a FF it's a big problem.
The other thing is attracting new users. with a selection of FF starting at a low price canikon have a huge edge even with people deciding on the apsC camera to buy. they already hafve an edge this just amplifies it. Like it or not the enthusiast end of the market will head into FF territory and it will become less viable for an enthusiast apsc model. not this year specifically but over the next few most definitely

The idea that FF is eating into APS from above have no basis in reality. FF has consistently been well below 10% share of the DSLR market. With improved sensors there are no reason for this number to increase.
Secondly, there are no signs in the data to suggest that miroirless have stolen customers from entry level DSLR. Both Nikon and Canoin are reporting record production volumes driven by the entry level. Pentax increase to is also due to increased sales of entry level DSLR.

[Kunzite]

5DII is dropping $300 next week apparently . MSRP of the 5D3 is not what it will stay at for long. either the 5D2 will be totally discontinued and the drop is to clear it (and keeping the 5D3 higher is a way of holding the 5D2 price a bit) or the 5D2 will stay as the value proposition (and by fall then the 5D3 will drop to $3000 after launch (if like the D800 it sells at predicted rates without a drop they will hold the price)
Remember all these FF models are MIJ with higher cost of production as well, and as soon as production moves so will price.
As long as they can sell everything at the asking price nikon and canon will hold the price....until they are ready to change the market Or someone else comes in and provides competition
Depending on Sony's price on the rumoured 3 models that may be the game changer. Pentax cannot go in currently at cnikon prices they need to represent value in the market until they are more established in it. If canikon keep prices at 3-3500 then Pentax can come in at 2400 with the right model.
If you think FF (which is only in second generation at nikon) will not move down in price and squeeze from above you are sadly mistaken (lack of competition in the market is all that allows it to hold at the moment)

Some price reductions for old cameras? What a surprise!
Open your eyes; the new FF were launched at over 3000$. That's reality. There's no moving down in price, there's no "squeezing". And I doubt Pentax will undercut Canikon by 1000$.

[eddie1960]

I believe it is sensor cost that prevents FF in being cheaper. I don't think any manufacturer have any interest to sell FF at a loss...

Sensor costs are not written in stone.they drop in price regularly and improve as well. this is no different than the arc of memory cards (I remember selling 256mb memory cards for close to $200 not that long back in the greater scheme of things.) If Nikon is as Thom Hogan has posted reducing the D700 to 2199 and removing dealer MAP agreements from it while continuing to produce it then FF apparently can be made at lower prices because Nikon definitely is not selling FF to lose money (neither is Canon where the 5D2 is apparently dropping $300next week - bet the mirror Nikon with no MAP as well if the drop isn't just a clear it out thing)
the only hard and fast rules of electronics are that prices go down and Technology gets better at the same time. 25 years of retail in that sector and I've never seen it go the other way, and I doubt that is going to change

being a latecomer to the tech Pentax' challenge is to get a sensor source and then produce a competitive product (and for them competitive will have to mean better featured at the same price or the same/lesser featured at a lower price
better featured than a D800 at the same price is likely unattainable for a first model, but if the competition is a D700 @ 1999-2200 then better featured at the same price should be attainable I would think (even if it means recycling some very old tech like the interchangeable finders from the LX, and pairing it with the rumoured 24MP sensor from Sony (which will likely being a NEX 9 and an entry slt IE S88, the36mp coming in the A99 over on the Sony side)
Certainly already there are a large number of enthusiasts moving into FF, almost all the pro market is there already and in many cases the D800 or 5D3 will be the third FF model they have bought

[Kunzite]

Don't forget sensors have a fixed size - which limits the price dropping quite a bit. What you see in memory cards is because much higher densities are now possible.
Again, open your eyes: the new Canikon FFs are not cheaper.

[mecrox]

You make it sound as though anyone willing to pay this price for a camera would automatically go for a "full frame". I believe that is a fallacy: I, for one, wouldn't. At least, not just for the sensor size's sake. And I'm fairly certain that at equal price, an APS-C camera would be better specified than a "full frame".

Good point. I was talking about FF a few weeks ago to the guy in the shop near London from which I buy all my Pentax gear. I said that I wouldn't want one, simply because I wouldn't feel comfortable walking around a host of urban areas in the big bad city with several thousand bucks' worth of camera in my hand. I don't need the hassle. He said he felt exactly the same. It isn't just whether one can afford FF, there is a "what's this whole pursuit really worth to me?" question here as well. There are many good reasons for using an FF but for a more general go-anywhere, walkabout but highly capable camera which doesn't cost a fortune, APS-C still has major advantages, imho.

FWIW, I live in one of my country's major tourist towns which is always packed with visitors and tour groups, many from the Far East. I see them hundreds and hundreds of them with compacts, m4/3s and APS-C SLR cameras, but about one a month with FF (and about one a month using Pentax come to that). Outside of pros and jolly wealthy enthusiasts I don't see FF having as much going for it as some here think. For myself, I would sooner trade down to a top-end m4/3 than up to an FF. Trading up to FF would only mean I'd soon have to buy a second, smaller camera for travel and street photography. Can't see the sense of it.

[kenyee]

Sensor costs are not written in stone.they drop in price regularly and improve as well. this is no different than the arc of memory cards (I remember selling 256mb memory cards for close to $200 not that long back in the greater scheme of things.)

Ditto what Kunzite said.

You're comparing apples and oranges. I've done chip design before and the cost is directly proportional to the size on the wafer of your chip (wafer cost is a fixed cost, and the bigger the area, the more bad chips there are so yield goes down as size goes up).
Memory has gotten a lot cheaper because it has shrunk in size while gaining functionality (density/speed). You can't shrink a FF sensor....it's a fixed size. Same reason MF sensors are $$$$ expensive.

Sony can sell the A77 cheaper than the Nikon body w/ the same sensor because they don't take a margin off themselves when using their own sensors. They're also willing to take a bit of a loss to try to grab market share (it's not working though, but they're burning a lot of cash doing it that Pentax doesn't have :-)

I'd echo what others have said though...if something like the D800 were around $2K and the next gen Pentax APS-C body were $1500, I'd be sorely tempted to try it just to see if the size vs. image quality tradeoff is worth it. The D700 isn't worth it to me ;-)

[RonHendriks1966]

I've done chip design before and the cost is directly proportional to the size on the wafer of your chip (wafer cost is a fixed cost, and the bigger the area, the more bad chips there are so yield goes down as size goes up). You can't shrink a FF sensor....it's a fixed size. Same reason MF sensors are $$$$ expensive.

I was thinking about that. About Dalsa (or what ever other manufactuar). When Pentax would order a special design sensor. Say that 6x6 micron pixel size CCD sensor (but this is the same with CMos). And you order that for APS-H, FF 36x24mm and for MF 48x36mm sensorsize. Cut from the same wafer then you can make more out of the surface by cutting the wafer the smartest way. Would this make it cheaper?

[falconeye]

the cost is directly proportional to the size on the wafer of your chip (wafer cost is a fixed cost, and the bigger the area, the more bad chips there are so yield goes down as size goes up).

Right.
But try to actually compute the cost of a wafer for a given process, the yield (fault density) and waste (300mm placement).
I did from best public sources (chipworks and others) and cost is nowhere near what the current market makes us believe. I posted it here a couple months back. Maybe, you have better sources and can redo the exercise?

BTW, if image sensors can be made by avoiding the most recent process, machines are much cheaper or deprecated and machines are the dominant part in wafer cost.

Edit... Without looking up, a quick search tells me a 300mm wafer should be significantly below 10000$. Maybe, I should estimate it by going by DRAM cost and known DRAM silicon real estate. A wafer minus waste gives 60 FF dies, about 40 good ones from defect density numbers I could find (yield). This is $250 per FF die. Or less. Certainly not a decisive factor in overall cost. We're not talking cost here, we're talking about what the market is ready to pay. And as long as we still pay 1000$+ for APSC cameras, FF prices will stay high too.

What is happening now is that sensor fabs charge 100% margin (500$) and camera makers charge another 150% margin, making FF cameras 1500$ more expensive for us. With more competition, margins would come down.

[Pål Jensen]

Right.
But try to actually compute the cost of a wafer for a given process, the yield (fault density) and waste (300mm placement).
I did from best public sources (chipworks and others) and cost is nowhere near what the current market makes us believe. I posted it here a couple months back. Maybe, you have better sources and can redo the exercise?.

Cost is one thing but it is difficult to calculate developing cost and other overheads not directly related to the pure manufacturing process without inside info. In addition, the sensor manufacturer will want to have a profit on the thing as well. The camera manufacturer have to buy this expensive component that probably cost hundreds(?) of dollars. There's also lots of expensive (relative) stuff in a FF body apart from the senors. The magnesium bodies aren't cheap either. Then theres developing, marketing and distribution cost of the camera. The manufacturer wants a profit as well. Then the various distributors in various countries wants a profit; not to mention the camera stores. It is amazing that they can sell them for a mere $3000+

[kenyee]

Cut from the same wafer then you can make more out of the surface by cutting the wafer the smartest way. Would this make it cheaper?

Do this: print out a bunch of photos (different ones so you can't glue one part of a photo to another to make "one photo") of different sizes. Glue them all to a wall in a circular area. Take a shotgun (not loaded w/ slugs of course, or it'd just be one hole) and aim it at the circle and shoot it (those are impurities or problems w/ the manufacturing process). If you hit any image, it's unusable (hot pixels, dead pixels, dead rows, etc.). That's roughly what happens w/ a wafer
The bigger the area, the higher the likelihood of getting hit by something that renders that photo or chip useless. And no, you can't magically change the sizes of stuff to avoid the holes you didn't expect at specific spots

For FalconEye, I dug for a bit w/ google and couldn't find anything about the actual physical cost of a chip, but Sony has to make enough off each to pay for: A) the manufacturing facility and B) the design team and C) physical cost per unit. Physical cost isn't a direct relationship to final cost you have to sell to make a profit, which is also dependent on volume. You have to pay for all those unrelated things, otherwise you'd operate at a loss and who does that unless it's the US government that has infinite tax dollars to throw away?

I did manage to dig up the wholesale cost of the ipod nano/ipod sensors (you could argue from this that Apple should be selling them for $10 I suppose):
Image Sensors World: iPod Nano VGA Sensor Price is $0.20
and a guess on the cost of MF chips:
Understanding Sensor Design

[bobmaxja]

Canin is designed by engineer., Nikon by peoples who love taking photo. THat the didfferences between the d800 and Mark III

[jeffshaddix]

Canin is designed by engineer., Nikon by peoples who love taking photo. THat the didfferences between the d800 and Mark III

I think both of those companies make cameras via the marketing department.

[Ayoh]

What about the additional costs associated with stitching? The area of full-frame sensor exceeds the field size of existing CMOS steppers by ASML and Nikon. Consequently the full die exposures need to be "stitched" from segment exposures each of which require critical alignment and multiple sets of masks. The stepper mask sets for a modern CMOS geometries can be extremely expensive (I read somewhere that it can be in the six figures). Have you factored this additional alignment and mask cost in your estimate? Furthermore I suspect the alignment steps may also considerably reduce the yield below the estimates published for non-stitched dies. As APS-C sensor area fits within the existing stepper field size it avoids these additional overheads, so perhaps the sensor cost difference is in fact proportional to the difference in APS-C and 35mm camera bodies. I am keen to hear your comments. Thanks.

Right.
But try to actually compute the cost of a wafer for a given process, the yield (fault density) and waste (300mm placement).
I did from best public sources (chipworks and others) and cost is nowhere near what the current market makes us believe. I posted it here a couple months back. Maybe, you have better sources and can redo the exercise?

BTW, if image sensors can be made by avoiding the most recent process, machines are much cheaper or deprecated and machines are the dominant part in wafer cost.

Edit... Without looking up, a quick search tells me a 300mm wafer should be significantly below 10000$. Maybe, I should estimate it by going by DRAM cost and known DRAM silicon real estate. A wafer minus waste gives 60 FF dies, about 40 good ones from defect density numbers I could find (yield). This is $250 per FF die. Or less. Certainly not a decisive factor in overall cost. We're not talking cost here, we're talking about what the market is ready to pay. And as long as we still pay 1000$+ for APSC cameras, FF prices will stay high too.

What is happening now is that sensor fabs charge 100% margin (500$) and camera makers charge another 150% margin, making FF cameras 1500$ more expensive for us. With more competition, margins would come down.