Well first of all, thanks for the plug
I'm a huge fan of Sony Vegas Pro, probably because it started life as a DAW that grew video capabilities. It's great to be able to use VST and DirectX plug-ins, ASIO, etc. Also, it lets me "think like a musician."
As a result, I tend to use SONAR and Vegas as a team, but not always. Also, it depends on the video...
Tutorial VideosFor my advanced workshop videos, I did the narration first in SONAR (with a lot of editing and sweetening), then loaded it in Vegas. While Vegas played back the narration, I listened and did the moves required for the video in SONAR (while capturing the moves with Hypercam). I then took the Hypercam video, which also contained what I was doing with the audio in SONAR, loaded it into Vegas, and made it conform to the narration. I then rendered separate audio and video streams from Vegas, which was what Streamworks wanted in order to assemble the videos.
This requires using two interfaces and frankly, the patching and preferences were sufficiently complex that I ended up writing documentation for how to do it! That way I just had to follow the steps to set it up.
Soundtracks for Existing VideosThis is what I did for trade shows like for Harmony Central. I'd edit the footage first in Vegas, then load the video into SONAR. I'd then create the sound track in SONAR. In some cases I also worked on the audio part of the video in SONAR, or from audio "flown in" from a separate hand-held recorder. Once the audio was done, I'd export, load it into Vegas, and do the rendering from there.
Much depends on whether you're cutting video to audio, or audio to video. You can render the video/audio combination from SONAR, which is fine if you need AVI or Windows Media Video (or Quicktime, if you're running a 32-bit system...c'mon, Apple, 64 bits has been around for a while!). Most of the time I need to do MP4, which is why I render from Vegas.