Originally posted by MRRiley Frankly, the more crap they put in my still camera to make it do other things, the less I like it. I want a still camera to take still photos. If I want movies I will buy a camcorder or movie camera.
True, however I don't mind additional features as long as they are not in the way of the main functionality, and if the whole system is not compromised by adding them. If a video DSLR would behave the way some cell phones behave on startup that would not work for me at all. (The phone they gave me at work was asking me every single time "do you want to use me as a camera, or as a media player, or... wait, how about I do this for you, or that, or...". No! No! No! You are a PHONE, so behave like one I need to make a call!)
Video in DSLR is technically logical thing: with fast and direct access to pixels in CMOS sensors it's mainly just a matter of software to add some sort of video mode. There is really nothing to add from the system architecture perspective. Mirror goes up, live view activates, and once we press shutter release button an imaging engine and CPU start reading CMOS, processing and storing data. The only difference is that instead of one CMOS readout and packing data to JPEG, the software takes 24 or 30 readouts a second and packs data to MPEG (or any other video format). The only hardware improvement needed is faster hardware to accommodate for more data and that's basically all. (Of course, it is possible to add dedicated video processing hardware.)