>>>>I think FX files are generally of higher quality >as >>they >>Ultimately you're magnifying your output less from an FX >sensor... > >Perhaps we should define what we mean by >"enlargement" before we go any further... > >Once a digital image is stored, it has no real physical >dimensions (length, height) - only pixel dimensions. The >traditional view on enlargement was the ratio of negative size >to print size - both had physical dimensions, so it was easy >to say that a 35mm negative (approximately 1.5 x 1 inch) >printed at 6 x 4 inches involved a linear enlargement of 4x, >or 16x in area.
Ultimately though, the lens is projecting an image onto a light-sensitive medium of a certain size. That projection is recorded somehow and then enlarged to an output medium. It's the size of the image projected onto the sensing medium vs. the output size that I'm talking about when I refer to magnification in this context.
>Now in the digital age, the thing that matters most for >resolution (the original question here) is pixels per inch. A >digital file of 3000 x 2000 pixels (6 MP) can be printed to 10 >x 6.7 inches at 300 ppi - and that is true whether the >original sensor was DX or FX.
Yes, but I don't believe the OP's question was meant to isolate image quality. In that case, the quality of the pixels matters too, and this relates to sensor size *unless* you crop both sensors to the same size (crop FX 2/3 on both dimensions gives you DX).
>>...so forget about the idea that larger photosites are >what >>made D700 better. > >What makes the D800 equal to or better than the D700 in image >quality is the newer technology of the sensor and firmware. >If you're comparing the same generation of sensor and >firmware, larger photosites do help. The D700 outperformed >the D300 (both 12MP) in low light chiefly because of its >larger photosites.
Are you sure about that Brian?
That was certainly the prevailing wisdom of the time. That same wisdom caused many to cry foul on the D800 announcement, believing that it couldn't challenge the D700's high ISO performance with those small pixels. Then, samples came out and people had to reconcile that D800 could have the same pixel-level noise as D7000 (similarly sized pixels) but still beat it by a stop on high-ISO and dynamic range. This confused many, until we started to realize that the sampling frequency didn't really have any effect noise performance, but the larger sensor always does. At 2.25x larger, and FX sensor collects 2.25x more light, regardless of pixel pitch.
The D300 & D700 had equiv-tech on the processing side, but the D700's sensor was also 2.25x larger than the D300's. You'd expect that to provide somewhat more than a 1-stop advantage, and it turns out, that's exactly what it did. We don't have a 27MP D700x with D300-sized pixels to make the comparison, but you can extrapolate from the D800/D7000 comparison, which is equiv-tech with the same size pixels, and you have your result.
> >>Shoot the same subject with the same lens on an FX camera, >and >>now you need to crop deeper to get the same framing. > >You have to crop more of the scene away, yes, but which will >turn out better depends on the pixel density. Comparing a >D800 and D7000 in your scenario, for example, you'll have >roughly the same amount of pixels on the subject - or the same >"resolution". Both could be printed the same size >and show about the same level of detail.
Once you crop both sensors to smaller than DX, you negate the FX sensor advantage, and yes the pixel density (and pixel quality) now matters. This becomes very difficult to logically compare D600 to D300 because the 5-year-old tech throws a wrench into the works. You have (slightly) greater pixel density with D300 but newer tech with D600.
I can tell you first-hand that at the pixel level, the D7000 blows away the D300s, but I still use D300s for shooting things like birds, for its frame rate, AF system, and very deep buffer. Which is what I was trying to say in my other post, that ultimate potential image quality isn't really the whole story when it comes to "getting the shot". The other features of the body leave the semi-pro DX camera as an attractive option, and given my D7000 vs. D300s experience, I'd be very surprised if D600 makes a better DX camera than D300s, since in DX mode it's probably very close to being a 10.3MP D7000.
>This is all very complex, and there are many ways to look at >the issue...