I have gone through basically all of Nikon's 24MP DX line-up within less than a year and have found real world, meaningful differences in the IQ of each one I've owned. I started with a D7100 I bought last Thanksgiving and it currently holds the highest IQ watermark for me, but I found it cumbersome to use and carry, after all it was my first DSLR, coming straight from cell phones, point and shoots and a Nikon F3, so it was quite an intimidating camera for me as a DSLR beginner to use. I also didn't like how its JPEG engine handled artificial lighting, which tended to output greenish, making everybody in the pictures look like martians (I'm exaggerating a bit here, but you get my drift). I then got a D3200 and loved its simplicity and lightness, I also held high hopes for its purportedly in-house Nikon sensor, but in actual use I didn't like its IQ as it consistently outputs under exposed and warm, maybe because of its simple 420 pixel matrix sensor compared to the 2k one on the D7100 which should be better at nailing proper exposure and WB. Also its video output is very noisy compared to the D7100's and there was no 60 fps option when shooting 1080. So now I have a D5200, which has similar IQ to the D7100 in terms of exposure, WB and noise, but with sharpness more akin to the D3200 and, ugh, buttons all over the body and in the wrong places. So I was going to wait until the D5300's price came down lower to upgrade to 1080p60 and regain the D7100's sharpness but a great deal on the D3300 got in the way of my plan, so I couldn't resist the temptation. I was expecting the D3300 to get IQ and video noise akin to the D7100 as I previously thought that it used the same Toshiba AA-less sensor, but reading an article somewhere else (AFTER I placed my order on the D3300) lead me to Thom Hogan's reviews where he finds out that NONE of Nikon's 24MP DX cameras share the exact same sensor. Darn. So today I received my D3300 and after lifting up its mirror I see that, indeed as Thom Hogan found out, the sensor looks completely different to the one on the D5200. Different color, different size (D3300 is smaller), different etchings. I don't remember how the sensors on the D3200 and D7100 looked like, but I sure remember how noisy the D3200 was at low light/high ISO when taking videos and I am seeing almost the same noise level on my D3300.
All this means that Nikon actually goes the extra mile and spends 2 or 3 times the same resources to make sure the IQ of its cheapest DSLR isn't up to par with the next one in price. Let me rephrase that: Nikon actually spends money to make the D3*00 series' IQ worse than the D5*00's ON PURPOSE! I mean, to improve the IQ of the D7100 they had Toshiba take away the AA filter from the sensor used in the D5200 while at the same time they had Renesas making a completely different sensor for the D3200 even though all of them are 24MP and have more or less the same behavior when taking photos, but they make sure that none of them share the same behavior when taking videos. And different sensors mean different processors, data bus design, FW calibration, etc. And then Thom Hagan finds in his D5300 review that it uses yet another different sensor, this time apparently from Sony! I was mislead by almost all reviews out there and by Nikon's own posted specs that the D5300 and D3300 shared basically the same sensor as in the D7100 and thus behaved all very similarly. Caveat emptor: don't let yourself fall for it as I did. Yes, all reviews out there pretty much say that the D3300's videos look very much like D5300's, but that's REALLY NOT THE CASE! I am now just hoping that the D5300's video quality is close to the D5200's, and here's hope that Nikon will take advantage on having 3 different 24MP sensors in their DX lineup and finally give us one with on chip PD for the D7100 and Coolpix A replacements. Or maybe they should start buying Samsung's 28MP sensor as seen in Samsung's NX1.