I don't have specific answers to your questions because you're overthinking the problem. The solution to your problem involves making tradeoffs. Once you understand the tradeoffs, it's fairly simple to decide what to do.
I don't upload to any Internet site for MA. I do my non-Internet work with either .AVI or .MPEG files depending on which camera I use. Since I acquire my video in person and don't need help doing analysis, I don't need to get my source video onto the Internet. For my Internet MA, I just deal with whatever gets posted. Quality of source video is one factor determining whether I will spend time doing MA. If possible, I will locally save Internet hosted videos and convert the flash based videos back into avi format for use with V1. V1 works best with AVI and MPEG. V1 has new support for youtube clips but I've not had any luck with that yet. V1 has trouble doing decent slow mo with WMV files (e.g. it's real annoying to step through videos and watch the skier jump backwards and forwards doing frame by frame advance). Once you've done MA work with high resolution source files, working with any Internet hosted video is like surfing the net with a 1200 baud modem. Having a high quality source cuts my time required for MA 1/2 to 1/3 because it's so much easier to work with. For doing analysis, you want the biggest picture with the highest frame rate and the least compression. The tradeoff is the clip is too big to share over the Internet. Then again, some of the gurus can see clips in any resolution and don't need slow mo of any kind to do their MA. V1 has set up their Internet coaching service to support high res avi format files but we're talking $$$$ for this.
Within the same format, the longer the clip, the smaller you need to make the frame size. Youtube supports 100MB max input filesize. For your 48 second clip, you're not too far off. You can simply try cutting the frame size down to meet the 100MB limit or break the clip into 2 pieces.
Audio is pretty much useless for MA. That's easy to drop out of what you upload, but usually does not save much. But dropping audio also makes the non-MA entertainment value of your clip pretty low.
Kids may be used to these tiny IPOD videos, but for MA, the bigger the frame dimensions, the better. Of course, the bigger the frame dimensions, the larger the clip. Go figure.
Cutting the frame rate is a bad trade off because most of the time you are losing movement details that can be revealed in slow motion. But some of the compression methods achieve their compression by dropping frames and some source video has a high enough frame rate to afford dropping frames. So sometimes monkeying with the frame rate won't do any harm.
With respect to formats, MPEG seems to have the best bang for the buck in my experience with respect to quality versus file size. Serious video techies can talk for hours about codecs, but as soon as you go with an esoteric high performance codec, some percentage of your viewing audience is going to have playback issues. If you know who you're sending it to, you can work things out. If you're going for a widespread audience, that's begging for trouble. Codecs can be a pain in the butt.
If the MA analyst is going to use software instead of the built in Internet browser viewers, then built in slow mo is generally unnecessary. In my experience training instructors to do MA, their skill level is inversely proportional to the slow mo rate required. Rookies may need to see a still frame for several seconds before looking at the next frame. The more experience you have, the better you can see things at faster speeds. One thing I've found is that a slow mo clip can counter act some of the frame rate compression when video gets converted into flash format. Otherwise slowing down to from 1/2 to 1/4 works ok. The faster the skiing is, the more you slowing it down helps.
There are some other basic tips:
- use your movie software to edit out useless pieces of your clip
- use a telephoto lens adapter to increase your optical zoom capability
- avoid using digital zoom (this takes a bit of experience to recognize where your camera's digital zoom kicks in, but it's really evident when doing MA)
- turn the image stabilization on before recording
- do practice recording of other skiers to learn how to hold the camera and film with the least amount of shake and the smoothest zooming.
- try to keep your skier filling 2/3 of the frame in the viewfinder - this results in the skier filling 1/2 the frame in the clip (anything smaller than that only masks movement details). The tradeoff here is that the faster the skier moves, the harder it is to keep the skier large and in the picture.
- snow and sky backgrounds make for better compression than tree backgrounds
- the better the light, the better the clip (save the cloudy day clips for home viewing)
- if possible, set your camera's capture frame rate to match the lighting (fastest possible for sunny days, backed off for cloudy days)
- bury the phone video camera (or the still photo camera that does video) in the snow and slowly back away
To the extent that you do the basic tips, you can afford to lose more detail in the compression used for Internet posting.
I have not done any work with Dartfish's video overlay capability, but most of the tradeoffs should apply equally.