jonnyapple:
What makes getting AF sensors to function on the edge of the FX frame harder than getting them to function on the edge of the DX frame?

For if a DX lens mounted on a DX camera provides enough off-axis response to make the edgemost sensor on a D300 work than a FX lens mounted on a FX camera should be able to do the same, no? The image circle is scaled as well as the sensor size.

Testing123, I thought, "That's a cool illustration. Why didn't I think of doing that?" Then I saw it was me and remembered the discussion that brought it up in the first place—I think it was originally alphanikonrex's idea and I just clarified it with the color.

This seems like a good time to discuss something here that I've been meaning to but haven't had time [people groan knowing I'm going to post another long post]. I suspect there's something more than just convenience in the lack of extra coverage in FX AF sensors. I feel like if Nikon could have done it they would have. Heaven knows Nikon has consistently stayed ahead of the other players in AF and metering.

As I understand it (and I'm not sure I do!), phase-detected AF is similar to a rangefinder, using two prisms (four for cross-type sensors) to direct light that is coming from just one side of the lens each onto the AF sensor behind the set of prisms.

I wish I had an FX body to show this next part, but my DX body will have to do (I'm using a DX lens to try to make it more comparable, though). This is a scene with a lot of point light sources with the focus intentionally wrong to illustrate that the light at different points on the sensor comes from different areas of the front element of the lens. Each red circle (all are the same size) shows what I think is what the front element would let in so that you can see how some of the light gets blocked by the stop's blades (in this lens's case—other lenses might be limited by an internal element). In any case, the thing to notice is that light does not get to every point on the sensor from every side of the front element except at the center of the sensor. Why is this important? Because the same would be true for an AF sensor that is off-axis, as well, and if they rely on getting light from both sides of the lens then at best their performance is worse and at worst it's impossible for an AF sensor to operate there.

I might be completely off my rocker with all this. Feel free to correct anything that you see that's wrong. I would love to have a deep understanding of what happens with phase-detected autofocus. I would also love to hope that someday I'll have an FX body with the kind of AF coverage my D300 had, but as I see it the CAM-3500 might be pushing the physical limits at least as far as coverage is concerned.

To add a data point to the discussion, I was listening to some photographers and engineers ralking on a podcast (i forget which one, its been a while)

They suggested that as a rule of thumb, after yield issues are taken into account, a FX sensor costs around $500 to make, a DX costs $50 and cell phone sensors cost $5

This makes sense as you can fit 200 DX sensors on a wafer, but only 20 FX. so the loss of one affects FX much more than DX (19 vs 199).

And this has nothing to do with Moores law, so these costs will not change without some significant changes in fab methods.

So a sub $2000 FX camera with a quality body is currently a dream. But depending on how they do it, its not impossible.

PB PM said:
The D3 and D700 have the same Multi-CAM3500FX auto focus system, which is different from the Multi-CAM3500DX found in the D300/D300s, so nothing was shoehorned in.

I don't care what Nikon subjectively labels it.

Objectively, the coverage on the FX bodies is much less than on the DX bodies and the absolute spread (as measured across the sensor in mm) is MUCH MUCH closer to the DX spread than to a scaled spread.

See johnnyapple's illustration from way back in the day:

PB PM said:
The D3 and D700 have the same Multi-CAM3500FX auto focus system, which is different from the Multi-CAM3500DX found in the D300/D300s, so nothing was shoehorned in.

I don't care what Nikon subjectively labels it.

Objectively, the coverage on the FX bodies is much less than on the DX bodies and the absolute spread (as measured across the sensor in mm) is much closer to the DX spread than to a scaled spread.

Testing123 said:

Though the other cost differences between the D300 and D700 (mirror assembly, prism box) are likely fixed offsets and not percentages. And I believe we'll expect a new AF package on the next FX camera bodies, not the hack of shoehorning the D300's AF module into the D700 and having inferior coverage at the edges. That would be another cost. But maybe the D9000 (D7000 FX) won't do that upgrade?

The D3 and D700 have the same Multi-CAM3500FX auto focus system, which is different from the Multi-CAM3500DX found in the D300/D300s, so nothing was shoehorned in.

"R+D works into anything"
Correct, but that doesn't mean that (not) knowing R+D costs aids (hurts) his basic calculation. R+D is a variable in both the calculation of the D300 cost as well as the D700 cost. If one does a simple D700 price minus D300 price, it is valid because the term has been removed from both sides of the equation.

"I would bet my car Nikon is loosing money on FX cameras"
That speculation, if true, changes the validity of projecting the D700-D300 delta forward, but it would not lessen the delta. Therefore the D700-D300 delta is a minimum one would expect to see in a D9000-D7000 calculation.

"Fixed cost (R&D, creating the manufacturing machines, dies, assembly line etc.) + Variable costs (chips, memory, motors, parts, rubber, glass, shutters, etc.) To drive costs down you need to produce a ton so you can buy the variable cost items cheaper."
Correct except that you don't take into account yield rates, which need not improve as production ramps up. Obviously in the ideal world yield rates increase as production increases and issues get addressed, but when pushing the margins of current practice they often do not. In something which takes up as much of a wafer as a FX sensor yield rates are critical.

"What drives DX down, is that the assembly lines have been in place for years and paid for. The parts (Buttons, rubber, Mag alloy, plastic, etc. have been are being spread across 12-20 different bodies for years. FX is new, and only has had 4 production bodies"
The D700 is the D300 is the D200 is the F100 in this regard.

FWIW I think D9000=D7000+(D700-D300) is a safe equation for the minimal expected price.

Testing123 slipped one in there on me.

R&D works into everything. And yes, I would bet my car Nikon is loosing money on FX cameras- they make a ton on the 3100 and p&s. What is working against our pocket books are the basics of manufacturing: Fixed cost (R&D, creating the manufacturing machines, dies, assembly line etc.) + Variable costs (chips, memory, motors, parts, rubber, glass, shutters, etc.) To drive costs down you need to produce a ton so you can buy the variable cost items cheaper.

What drives DX down, is that the assembly lines have been in place for years and paid for. The parts (Buttons, rubber, Mag alloy, plastic, etc. have been are being spread across 12-20 different bodies for years. FX is new, and only has had 4 production bodies. DX probably sells 1000 times more bodies than FX and in the last 5 years has sold millions more.

Testing123 has got it right - I was trying to keep things to layman's terms.

D7000 is a new sensor, but the sensors for the 300, 90, 5000, and maybe the 5500 and 3100 are essentially the same. Then you get into the backbone of cameras, the boards, processing power, cash, meters, on board memory, motors, shutter, etc. The 300 and 700 are entirely different cameras (with just a couple of shared parts)- D3 and D700 are very close if not the same insides.

Sorry Hearty you are missing the boat. The price difference is exponentially higher. Think of sensors like light-bulbs. Your argument is like trying to take a light-bulb from a light house and put it in a night-light in your hallway. Then you are trying to say somehow the light house light-bulb is somehow cheaper to make.

I don't believe the R+D costs of the D700/D3 can be determined, but I don't believe they affect your calculation either.

The rest of your logic and calculations appear sound. I was just attempting to point out the single place where I thought you were out on a limb. And, yes, it doesn't change the price much.

Though the other cost differences between the D300 and D700 (mirror assembly, prism box) are likely fixed offsets and not percentages. And I believe we'll expect a new AF package on the next FX camera bodies, not the hack of shoehorning the D300's AF module into the D700 and having inferior coverage at the edges. That would be another cost. But maybe the D9000 (D7000 FX) won't do that upgrade?

so you are saying the sensor cost is probably virtually the same(0) for the D90/D300/D5000 and the D7000? So what are you able to deduce re the D700/D3 sensors cost in terms of R&D? or does it matter at all?

Hmm If we accept that logic then the price diff is not $100 more which still puts the D7000FX at $1850. Still a great price !! I'll have one please !

heartyfisher said:I would say that the D7000 sensor is substantially more expensive than a D300S sensor

Sensor cost is almost entirely two things: effective die space (amount of silicon wafer) and R+D.
R+D on a mass-market device such as the D7000 will be amortized to near zero. That means you're left with effective die size, which is actual die size divided by yield percentage. We do not know what the yield rate is for the D7000, D700, or D300s, but in lack of better evidence we must use the same number.

Point being the D7000 and the D300s have the same cost sensor, with a nod towards the D7000 being cheaper if anything (more favorable R+D amortization rate due to higher volume, the D5000 replacement getting the D7000's sensor will only exacerbate this. Though there is a good chance the D300s sensor is simply a higher bin of the D300/D90/D5000 sensor in which case the R+D cost is effectively zero.)

Hmmm numbers don't seem to add up... and you need to compare similar level models of fx to dx i think..

And I seriously doubt that Nikon is making the D700 at a loss !

so all things being equal the difference between the D300 and D700 should be the sensor. minus say 10 % for the smaller size and minis another 10% for the number of higher volume of the D300 and that should be the price difference of the sensor. so its 2600 - 1800 = 800
taking away the 20% makes its about $ 650 now I would say that the D7000 sensor is substantially more expensive than a D300S sensor but lets just say its 100 more so that means that if you putthe old D700 sensor into the D7000 FX it would be about 1200+550 = $1750.
Sounds right to me .. you?

I think the D700 was US $2999 when it was released, and has dropped to $2399 where it sits today.

Oh hell yeah. I'm guessing they actually loose some money on the D700 but helps cover the costs to get manufacturing up to FX frame speed for future units. It has to do with how many sensors are sold. Gobs more DX sensors are sold then FX. I would be willing to bet that there were more D40s sold than all of the FX sensors combined. It would be interesting to see the sales numbers per body from nikon. The more built and sold X, the cheaper you can produce X. Pure economics.

There is allot more to a camera than the sensor and the D7000 is built to less standards than the D300s body series. The D300s is still $400 more than the D7000. D700 was around $2500 when it was released which is almost $1400 more than the D7000. Go by release date prices, not current prices to compare models.

So you guys are saying that the D700 fx sensor costs nikon at least $800 MORE than the D7000 DX sensor? hmm I find that improbable. :-) ... still could be true.

heartyfisher said:

@PBPM : From the samples I have seen the D7000 is close to the D700 but the D700 is still better by maybe half to 1 stop at the high iso settings. My proposed D7000 FX will probably outperform the D700 by half a stop or so due to the new expeed2 processor. Do you really think there wont be a market for a D7000 FX that out performs the D700 but say at $1500 or even $1800 ? I would think that it would be as back logged as the D7000!

Sure, there are tons of people who would love to have that. Doesn't mean Nikon will make one. Sony was losing money on the A850 at $2000, what makes you think Nikon could sell a FX body for less and not bleed tons of cash?

gd said:
There will no doubt be a market for an FX camera at under $2000 but that doesn't make it profitable for Nikon to supply it. With what appears to be limited production facilities, I would expect Nikon will struggle to fulfill D4 \ D400 orders when the next generation of cameras come along and as such I think even a D800 won't be along any time soon after the D4 \ D400 (and I'd be equally surprised to see it come along earlier). A sub $2000 FX camera is a far away dream.

AMEN! Someone understands economics!

"Do you really think there wont be a market for a D7000 FX that out performs the D700 but say at $1500 or even $1800 ? I would think that it would be as back logged as the D7000!"

There will no doubt be a market for an FX camera at under $2000 but that doesn't make it profitable for Nikon to supply it. With what appear to be limited production facilities, I would expect Nikon will struggle to fulfill D4 \ D400 orders when the next generation of cameras come along and as such I think even a D800 won't be along any time soon after the D4 \ D400 (and I'd be equally surprised to see it come along earlier). A sub $2000 FX camera is a far away dream. Happy thoughts ... just don't be too upset if Nikon aren't listening!

@IndyGeoff : I am not saying that there wont be a 18-24M camera in a D700 body. In fact I am sure there will probably be one. see point 3 in my original post.

@tsy : I agree that the D700 body is better than the D7000. as I mentioned to IndiGeoff I am not talking about a D700 class camera or replacement. I can see that there are probably 3 possible replacement for the D700. I am not talking about the D700 and its replacement in this thread.

@PBPM : From the samples I have seen the D7000 is close to the D700 but the D700 is still better by maybe half to 1 stop at the high iso settings. My proposed D7000 FX will probably outperform the D700 by half a stop or so due to the new expeed2 processor. Do you really think there wont be a market for a D7000 FX that out performs the D700 but say at $1500 or even $1800 ? I would think that it would be as back logged as the D7000!

@ TaoTeJared : Sony is not nikon.. just because sony failed at that doesnt mean that nikon will as well. Sony just doesn't have enough "X" factor in that level I think.
eww funny cherios! 4 years huh... gives me time to save :-)

@ LessThanThreeLeo : D800 ? as I mentioned above that D800 could be 1 of 3 things all of which will be equivalent to the D3S . Its a different market segment. No, those people they wont be getting the D7000FX when they can or want a D800!.

I don't think a low cost FX camera is in the cards. Considering the high ISO performance of recent DX bodies, I don't think it matters that much anyway unless you want to shoot at ultra wide angels at lot.

honestly that wouldn't make a lot of sense to me... the d700 body is still better than the d7000. Im mainly referring to the button lay out/ergos.

Basically what you are saying is you want a d700 with video...

I honestly would hope that the d700 successor has more than just video as the upgrade.

I am hoping for 18-24MP ff d700 type body. If that shows for under 2500$ I am likely to upgrade from my D90.

I would not be interested in a new camera with only 12MP, I'd just pick up a used D700 if all Nikon put out was a 12MP replacement.

D7000 is the answer to Canon's 60D just as Canon's 7D was the answer to the D300/s. They are just matching each other's line-ups.

The D300/s bodies are tanks already and I don't see where carbon fiber gets anyone anywhere but adding more cost for very little gain. It will still have magnesium alloy for the frame since it is still robust and fairly cheap. More weather sealing possibly but I have never had a problem with it getting wet.

I love the thought of a Sub $1500 FX camera but at this time honestly, it is just a pipe dream. Everyone saw the affordable Sony FX and they didn't do terribly well with two $2000-$2500 cameras and I doubt anyone else can adsorb the losses from manufacturing them. FX is still very expensive to create and the cost vs demand just isn't there. Everyone seems to underestimate how much it truly costs to create pro bodies and FX sensors. If Nikon or Canon thought they could do it, they would have already. I don't think anyone has ever understood Sony's Camera division.

I don't mean to pee in your Cheerios but the "affordable" FX is probably another 4 years away. The D7000 is the first of the major updates that will come down the line which makes me very excited to see what will come.

I don't see it to be honest. Wont most people just buy this camera instead of the new D800.