If you match your Project Preset to your source footage (in this case some of it down-rezzed), you will not get black bars, anywhere.
If you HD cameras will down-rez, in-camera, that would be my choice.
Now, you have some choices to make. I would look at my SD camera's footage, and if I shot in Standard (4:3) , or Widescreen (16:9). The HD material will have been shot Widescreen.
Let's just say that you shot the SD material as Widescreen also. For NTSC, that translates to 720 x 480 w/ a PAR (Pixel Aspect Ratio) = 1.2 (wider than tall). That would be both what I would set the Project Preset to, and also what I would set the down-rezzing in my HD cameras to.
If there is any sizing issues, it should be correctable with Interpret Footage, which allows one to force PrE to see footage a certain way. This can be useful when one has footage, where a flag is not seen/recognized by PrE. This happens with some odd MPEG CODEC's and also some MOV CODEC's, but almost never with DV-AVI Type II footage - PrE always gets that right.
Now, the total success of this will depend on how well your camera(s) down-rez and how they handle going from Square Pixels to 1.2 Pixels. If, for instance, the shots of Aunt Marge look a bit "fat," you will want to Interpret footage back to Square Pixels and then use the fixed Effect>Motion>Scale to eliminate the black bars. A little Scaling should not be noticeable - but you do not want to overdo it.
Also be aware that different cameras might create Widescreen differently. One way is to just "crop" the 4:3 to "look like" Widescreen. Another is to shoot Anamorphic to 4:3 squeezed and then the NLE un-squeezes it in processing. The last is to shoot true 16:9 w/ PAR = 1.2. So, it depends on how your camera(s) creates Widescreen. One needs to read the specs. in the manual carefully, as each needs to be handled differently.
As you have found, there are a lot of considerations, when mixing footage from different cameras. One needs to plan the editing ahead, and find a workflow that will allow one to get the footage into a common format first, then set a Project Preset that matches that.
The ideal is to as little Scaling, as is possible. As mentioned, this is best done in-camera during Export (from camera), or during Capture, if the camera has FireWire and PrE can see and connect to it.
Mixing sources can be done, but it's not a slam-dunk. The workflow will depend on several factors, on one must do the research to see which method works best.
Thanks Hunt. Setting up in 720 x 480 and lobbing in 720 x 408 videos results in no stretching or resizing at all. This software's smarter than I thought.
Alas, my HD cameras do not shoot SD - but the handy thing about that is that shooting a wide-angle shot in HD can, with a bit of clever editing/filtering be cropped into multiple fake SD shots and NOBODY WILL EVER KNOW.
PS: Don't tell the band about this post, they think I'm really clever ...
My lips are zipped!
Now, this TUTORIAL is written for PrPro CS4, but there are some great tips on going from HD to SD, if one's camera does not allow for in-camera down-rezzing. Might be something helpful for PrE in there too.