This topic has been talked about a lot so I thought I'd share my observations - you are free to draw your own conclusions though...
First up - all images were taken on a PAL HV30 camera. All SD video was captured in PE4 and all HD using HDVSplit. All image exports were done from PE4 as .bmp files with de-interlace off and were converted to .png (lossless compression) using MSPaint. I took two shots one immediately after the other with SD second. All exports are anamorphic widescreen at native resolution so appear squashed horizontally.
Now for the images...
1) An SD video exported from an SD project.
2) HD footage loaded into an SD project and exported directly.
3) HD footage downconverted in camera into an SD project.
4) HD footage exported ad DV/AVI from an SD project then reloaded into the SD project.
5) HD footage exported as DV/AVI from an HD project then loaded into an SD project.
6) HD footage exported from an HD project. (1080i)
I note that there are differing amounts of "blurring" with differing detail available on the bricks. However more interestingly is a variation is the colour of the bricks.
Thoughts anyone?