I would agree that anything less than 1080p just isn't worth it. I've mostly flown Mobius ActionCams, flew a U8 (similar to the 808, just a different form-factor, it's shaped like a USB thumb-drive rather than a keyfob) once and was so unhappy with its quality compared to the Mobius that I never flew it again. Beyond video resolution (the 1080 means it's 1080 horizontal lines [vertical resolution], if it's HD the vertical lines will probably be 16/9 higher, i.e. 1920 [horizontal resolution]) framerate is the next thing that matters. The Mobius is 1080p/30(fps), there are newer cameras like the Mobius Mini or Firefly Q6 (my current favorite) that are 1080p/60, so you get twice as many frames per second and therefore smoother video, or ability to do smoother slow-motion video, etc. The cameras that offer higher framerates tend to also offer higher resolution options at the lower framerates, like 2K or even 4K video, personally I prefer >30fps over >1080 resolution). The "p" suffix also means progressive, where the entire frame is taken at once, I'd suggest avoiding anything that ends in "i" as interlaced video shoots half of the lines (even vs. odd) in one frame and the other half in another, so for fast motion it can look pretty terrible, I'm not aware of any of these style cameras that shoot interlaced video. So I wouldn't ignore this suffix but I think you're unlikely to see anything other than "p".
I haven't found a shroud to affect the spin on any of my rockets, having flown Landru's shrouds on both 54mm MD and 4" rockets. If you just go and strap it on with tape (have done this a few times with electrical tape) I'd think that might have a bigger effect just because it's probably going to be bigger, and harder to make smooth/symmetric (and probably flaps about in the wind also). My 4" FG Nike Smoke rotates naturally (I tried to create my own diamond profile shape in the flat FG fins, and clearly wasn't perfectly symmetrical with it), as I'm working on a shroud for the Firefly Q6 for this rocket I'm tempted to see if I can put some features on the shroud to try to offset the spin imparted by the fins, but given the different positions on the rocket I suspect this might turn things into more of a spiral and probably not help as much as I'd hope.
But I definitely prefer down-facing. Most of the shrouds point the camera straight down, it minimizes the profile of the shroud, but also occupies more of the frame with the rocket body. As I design my own shrouds I'm going to try angling the camera away from the body at least a bit, I definitely want to ensure that part of the airframe is still in-frame. And in a DD rocket I'd certainly suggest having the shroud above the drogue separation point, my favorite videos are the ones where you can see the lower part separate at apogee. If you put it on the mid-section the camera will end up pointing upwards during the drogue descent (might even get some good shots of the drogue as well, and gets you a better view once the lower part of the rocket is "out of the way"), then will end up facing towards the ground again once the main deploys, so you can probably see touchdown.
Two example videos, first one is a 4" Excel DD, Mobius ActionCam was attached to the avionics bay (had the vent band lengthened 1" for the Landru shroud):
https://youtu.be/ax9BcakBrmc?t=28s
Second one is a Firefly Q6 that was taped-on to my 4" FG Nike Smoke with electrical tape, I particularly liked this flight both because it was a Vmax motor and it didn't arc-over at apogee, it just started to descend into its own smoke trail, which wasn't obvious from the ground video so it was a pleasant surprise when I reviewed the on-board video afterwards:
https://youtu.be/O6K3ks6h6d4?t=19s
