Lol JA, let's see how the wedding season ends and what the D800 prospects end up coming to...seriouslt though, nikons kicking the crap outta canon in image quality, excited to see your test shots jonny (1 month right?)

Drab's got it. If you want a more photographer-centric explanation, think of having a DX sensor as constantly using a 1.5x teleconverter with every lens you put on the camera. People talk about how putting a teleconverter on a lens brings out any flaws and it's for exactly the same reasons as these high-MP-count show up lens flaws—you have more photosites imaging a smaller area of the lens's imaging circle. (okay, this is assuming you could make an optically flawless TC) EDIT: It's also not exactly like using a teleconverter because you don't lose light. You get what I'm saying, though, right?

I like this discussion about physical resolution because it's forgotten most of the time. You forgot the D3x in FX, which has a pixel pitch of 5.9 microns. That means even the D300 and company out-resolve it (but with less than half the area of the image circle covered), not to mention the new sensors on the D3100 and the D7000!

I'm just glad they can keep getting better (or similar) pixels even while they add more pixels. I had a look at that chase jarvis shot at 3200 with the D7000, and I would have been thrilled with performance like that with my 6MP D100 at ISO 400. We've never had it so good. The D700 replacement is going to be spectacular—not that I'm going to get one. I've done some soul searching, thanks in large part to this forum, and I'm going to be completely content with my D7000...once I get my SB700 and 24-120 f/4. ;-)

But you'll have to let us live through you when you get your D800. Don't become a D800 snob, Shivas.

shivaswrath said:
I'm not math dude, so you're saying as things shrunk in pixel size 6.1 (D200)->5.5(D300)->4.78(D7k), the lens shortcomings become more prominent?

How does this reconcile with the D3 pixel pitch of 8.5?

I'm confused. . .

As the pixel size shrinks you're putting more and more pixels under the same area of glass.

Therefore the smallest feature one can resolve (angular resolution) improves.

Therefore softness / misfocus / aberrations / distortions which heretofore were not resolveable have become so.

Your problem with reconciling the D3 with this is you appear to not be taking into account the sensor size, and thus confusing the MP count with the number of pixels per mm of image circle.

The D3 should be quite forgiving of lenses, as it has quite the "fat" pixels and therefore doesn't ask much of the glass. With some of these DX sensors on the market now we're (theoretically) beyond the resolving power of common 35mm color film and approaching some of the lithographic films.

I don't know that math either, it's just something I've seen in practice. Images taken with the same lens on a D50, D80, D300 all have a different feel in terms of sharpness, and it isn't just the difference in sensors. Example, on the 6MP sensors the 18-70mm DX kit lens looks great, but on the higher resolution sensors, 10 and 12MP, it looks a little flat and generally unsharp.

I'm not math dude, so you're saying as things shrunk in pixel size 6.1 (D200)->5.5(D300)->4.78(D7k), the lens shortcomings become more prominent?

How does this reconcile with the D3 pixel pitch of 8.5?

I'm confused. . .

Same question is in my mind as well.

Other than low light I have seen many "real" world landscape test images with a D300 and D700 (shot at the same time) where I can see the larger microns also create sharper images in fine details. Now the D3x detail...that is in another league but then again those are the 5ish micron size. Micro-lenses, sensor wells, software, back lit/side lit sensors, who knows what all they can tweak to get more out of a sensor. One thing is for sure, hardware parts like micron size has trumped software in image quality.

Cannon seems to rely heavy on software and I hope Nikon does not do the same just to keep in line of the MP war.

I'd say SNR isn't as much of an issue as it was in the early days, as clearly displayed by the D300, D90 etc. I find image sharpness to be more of an issue as the pixel size gets smaller. The smaller the pixels, the more they expose shortcomings in the lenses, especially older ones, that were once thought to be very good.

I am excited about the D7k. One, because it's "pro" enough to be a backup to the D3-series, but not as expensive and heavy as a D300 series. Two, because it's got 1080p video (lol, haha!)

Quick sanity check on SNR and micron sizes (JA, your input here invaluable):
DX
D40 6mp - 7.9 micron CCD
D40x 10 mp- 6.1 micron CCD
D60 10 mp- 6.1 micron CCD
D80 10 mp- 6.1 micron CCD
D200 10 mp- 6.1 micron CCD
D300 12 mp- 5.5 micron CMOS
D3100 14mp- _._?? micron CMOS?
D7000 16mp-4.78 micron CMOS

FX
D3/D700 12 mp - 8.5 micron CMOS (largest pixel size of ALL current DSLRs)

So, ummm, isn't Nikon going in the WRONG direction here with SNR. . .since larger micron size is better?

Or are we at the cross roads of Physics and Computer Engineering, where they basically can make shit look less noisy than it physically is?