MANY D300 users have reported obtrusive shadow noise showing up at ISOs as low as 200. The color rendering of the D300 vs. the D200 has also been hotly debated in some of the forums I frequent, with many D300 users downloading the tone curves from the D2x/D2xs, infinitely preferring them to the D300s default curves and their “high gamma.”

Thom Hogan, whose opinions and reviews I respect and trust, has concluded that the “old” D2x is the better camera of the two at ISO 400 and below. Considering that the D200 has long been regarded as a “baby D2x,” it’s not all that surprising then that the D200 outperforms the D300 in many respects at ISO 400 and below.

The proof is in the using: I rented a D300 for a week before purchasing and compared it directly to my D200 using a Zeiss manual focus lens. I kept my D200.

Everything comes down to image quality at the end, all the bells and whistles of the D300 are finally not so important to the average shooter. Let’s start with the noise. Most reviews claim one full stop more for the D300. The measured noise (DxO lab) is the same for both cameras until 1600 ISO, towards 3200 ISO the D300 is slightly better, but both cameras have unacceptable noise here anyway. At 1600 ISO both cameras do not expose at 1600 ISO but at 1145 ISO (D300) and 1494 ISO (D200), this means we are comparing approx. 1150 ISO for the D300 to 1500 ISO for the D200 (data from DxO lab), which looks like cheating in favor of the D300.
If we compare 800 ISO (D200) to 1600 ISO (D300) we are comparing really 759 ISO (D200) to 1145 ISO (D300) that’s a difference of only 386 ISO or roughly half a stop. No wonder the D300 is “apparently” one stop better, being just less than half a stop better.
My personal findings are same noise from 200 – 800 ISO for both cameras; at 1600 ISO the D300 is slightly better, which only confirms the above lab tests. I nevertheless find already noise at ISO 200 in light shadows with the D300, where the D200 stays clean.
Even the D70 stays a lot cleaner a t ISO 400, there is no noise in light or medium shadows whatsoever, color is also more accurate, specially in greens (e.g. grass) where the D300 is always oversaturated, too clean (missing magenta and/or black in greens) regardless it’s setting (neutral, standard or else). The D200 has a CCD sensor and so does the D70; the D300 has a CMOS sensor. CCD sensors by its nature have less noise than CMOS sensors, which may explain this, although one could expect after years of research and development CMOS sensors to be as good as CCD sensors. CMOS sensors are much cheaper and need less electricity. But buying a D300 for having less noise is a wrong idea.
By the way, if you expose the D300 at 3200 ISO you really do it at 2063 ISO only (!!) Compare this setting with 1494 of the D200 at supposed 1600 ISO. That’s why, many reviews recommend setting the D300 exposure compensation to -0.3 or -0.7 which gets you again to the right ISO values. No further comment here.
D300 vs. D90:
Most reviews claim, that the D300 quality is the same as the D90 quality. But the D90 came out almost a year later and something must have happened in the meantime – and it really has. The biggest difference we find with the low light ISO sensitivity, where the D90 is almost 50% better or in numbers 679 for the D300 vs. 977 for the D90 (DxO labs); the D90 also has a better dynamic range (12.5 vs. 12) and more color depth (22.7 vs. 22.1) compared to the D300; in one word it delivers better image quality. Nobody ever reported this, at least I haven’t found an reviews telling us here the truth.
All these numbers come from the DxO lab, a extremely serious and well equipped independent French lab of objective digital camera image quality measurements (http://www.dxomark.com/). Check it out yourself.
D300 Color: The basic setting shows an average of 8.9% over saturation (sRGB) and 14.7% with Adobe RGB (Imatest), whereas the D200 is pretty close to the truth. Any camera should have one basic setting with the colors as close as possible to the correct colors. Over saturation can easily be achieved with additional in cameras settings or with photo editing programs. The D200 has it, the D300 not. I was loading down the DX2MODE settings (I,II,III), which you can find on the D200 too, reducing the contrast and saturation below the Neutral or Standard setting of the D300, in order to get some normal colors; Skin tones on top are pink, too much Magenta and too little Yellow, so the Hue control has to be moved too. I am close now, but the D200 still gives me better skin colors straight out of the camera. Even so I still de-saturate frequently colors from D300 pictures in Photoshop slightly, which I never do with D200 pictures.
Over saturated colors are for beginners, people with no color knowledge whatsoever, for persons without taste like children who like strong colors. You should NOT find settings like this as a default on a semi professional camera with a high price tag. It may be justified for entry level DSLRs but even here there should one “true color setting” besides the over saturated default setting.

I bought the D300 after reading dozens of favorable reviews, a “no-brainer” as some of them put it – but I must say I am not overwhelmed by this camera compared to the D200.

Everything comes down to image quality at the end, all the bells and whistles of the D300 are finally not so important to the average shooter. Let’s start with the noise. Most reviews claim one full stop more for the D300. The measured noise (DxO lab) is the same for both cameras until 1600 ISO, towards 3200 ISO the D300 is slightly better, but both cameras have unacceptable noise here anyway. At 1600 ISO both cameras do not expose at 1600 ISO but at 1145 ISO (D300) and 1494 ISO (D200), this means we are comparing approx. 1150 ISO for the D300 to 1500 ISO for the D200 (data from DxO lab), which looks like cheating in favor of the D300.
If we compare 800 ISO (D200) to 1600 ISO (D300) we are comparing really 759 ISO (D200) to 1145 ISO (D300) that’s a difference of only 386 ISO or roughly half a stop. No wonder the D300 is “apparently” one stop better, being just less than half a stop better.