DxO T1i/500D vs D5000

Ditto in some spots, not so mutch in others.

Dxo is interesting: applying a numerical rating to sensor performance. I think it is worthwhile to qualitatively compare sensor characteristics, but the numerical ratings are pointless. Their findings should be backed up with real life examples where you can see a difference, otherwise it’s purely academic.

On a separate note: Canon, get your act together.

Actually this is bad news for the Nikon, according to DXOMark itself:

http://www.dxomark.com/index.php/eng/Insights/More-pixels-offsets-noise!

Noise measurements in RAW format are indicative of pixel performance, but the viewing conditions of the image must also be taken into account. To compare prints on a given format of cameras with different resolutions, it is more suitable to compute the equivalent SNR of a camera with a given reference resolution. For equal, normalized SNR, a high-resolution camera is still better than a low-resolution camera. While it is always possible to simulate a low-resolution camera using a higher resolution camera (since downsampling is easy), it is not possible to simulate a high-resolution camera from a lower resolution one other than by interpolating or inventing data.

Bottomline, the 500D sensor is better since it is either equal or close to the D5000 sensor even with smaller photosites plus 3 more MP.