Calling for certain test

[Snapper123]

Hey guys, are there any tests out there comparing a still from a DSLR video to a resized-to-1080P 18Mp photo, with then a video still from that same video once the video has been enhanced with the appropriate post-sharpening or depth effects that are probably used for official productions?

So far, tests like this seem to show an extreme difference from a 1080P still, and a 18MP-to-1080P still, basically in my opinion the difference between something looking entirely cinematic/right for documentary shots, than not. But I'd truly like to know if the captured footage from these videos is pushable in post to look much different.

 

[robotsound]

The still image will look noticeably sharper every time. This is because the still image started out with all 18MP, and would be resized in Photoshop or similar using bicubic interpolation. The video frame grab would be starting with an image comprised of only 2MP spread out with gaps (also the source of the aliasing and moire issues), and then run through the h.264 codec.

I don't have a link to any tests, but I know I've seen them before and done them myself.

 

[robotsound]

Here is an example of some still vs. video tests. You have to scroll down a little farther on the page to see them, but they have side by side pictures.

http://prolost.com/blog/2010/9/2/ha-ha-very-funny-canon-now-get-back-to-work.html