Ok, here's my quick analysis, Bob.
On the capture:1. I love the action, plus the incoming bird (gull sp.?) makes the footage even more interesting.
2. The camera is not level during the capture (seen in water splashes). This can be corrected in post, but it will take away valuable pixels.
3. I can't seem to hear an audio. Audio (even ambient sound if the bird is not calling) makes a great contribution to the viewability of a bird video. One thing I do when the original ambient audio of the footage is ruined by noises that don't belong to the scene is to record another footage right after the main one. I use the audio of the second footage for the main clip. When such re-recording isn't possible, then a BG music becomes useful. ;D
4. I suggest you practice with the manual exposure mode. In the old auto mode, the camera defaults to f/45 in brighter scenes. This results into a very deep DOF (actually beneficial in this footage because the subject is moving), but takes away lots of sharpness and contrast, and makes sensor dust very visible. You can use auto-ISO in manual mode, and the camera adjusts the exposure automatically with up to 6 stops leeway (ISO 100-6400).
5. For natural looking motion, it's best to keep within a shutter speed range of 1/30 - 1/125 sec. Anything faster than this will result into a strobe-like effect. Use an Av with DOF deep enough to cover the whole bird.
6. TCs are much more effective in HD video than in stills. Because the 5D2 undersamples (reads every third row of pixels) to reduce 21 MP to 2 MP, the sharpness hit due to TCs is not as bad as when using these in stills. The TC also mitigates the the aliasing that might result from undersampling. On a 500 f4 IS, I've used up to a 4x TC (Sigma 2x + Canon 2x) and the IQ is still decent, see the footage below.
exposureroom.com/romyocon_barnswallow7. For smooth pans, you need a proper fluid head to dampen the panning force. IS mode 2 also helps in panning. I'm using the enormous Manfrotto 516 fluid head because I've to mount also the heavier Sigmonster, but for a 500 f4 + TCs I think a Manfrotto 503 will be good enough.
8. Shoot as much footage of a scene as you possibly can, using a variety of angles and focal lengths. This will allow you more editing choices later. CF cards and extra batteries are now very cheap. On TV productions, they normally use 3 cameras at various angles running at once, to make the scenes very dynamic and interesting after editing.
On the Post Processing1. The clip is quite short so we can't do much here, as cutting out the scenes with rough panning will leave us with very little footage.
2. I notice that you have a black space above and below the footage - this can be a result of improper settings of video output. I'm not familiar with video editors other than Premiere, but check if your setting is "square pixels" and the output resolution is proportional to the capture res, i.e. 1920x1080, 1280x720 or 640x360.
3. I use *.wmv for web postings in XR. IIRC, XR uses a PC/Windows-based transcoding engine and a *.wmv file approximates best the colors and brightness I see in my video when compared to the transcoded one. For a 1280x720 *.wmv, I render to a bit rate of 6 Kbps.
4. When editing multiple scenes later, try to keep each clip to a maximum of 10-15 sec long, unless the action is continuously interesting. This is to prevent the viewer from getting bored. This is easier said than done though if the editor is also the filmer, particularly when the filmer had to work very hard to get the footage. I myself am always guilty in breaking this rule, as I can't seem to cut out hard-filmed footage.
5. For transitions, the general rules are:
Clean cut or short dissolve - indicates the scenes are happening after each other in real life
Longer dissolves - indicates passage of time
Fade to black - indicates the end or even longer passage of time.
That's it, Bob..... feel free to post follow up questions/comments.