If you are changing the aspect of the pixels in the image, supersampling generally helps, but not in all cases. You can experiment as Laurence has done, and what he sees may be different from what you see.
Work with small clips, shorter than 5 seconds, put them on the monitor, and choose what works best for you.
I find that when going up in size and shifting framerates, plus when the pixels are going from square to non-square, that supersampling is a big benefit. I've not used it much going down, because it's not been necessary for anything we do. I'll play with it in the next day or so to see if I can find some tweaks.

Spot, if you could post an example of some good settings it would be helpful. I'm shooting 16:9 with my A1, but often have to mix in 4:3 footage. I am always trying to get the best quality possible and any tweaks or information on how to do this better is alway put to good use.

Thanks to both of you.

Laurence, what you say is that the picture actually got worse when you applied Super Sampling?

Spot, I’m looking forward to reading your results / conclusions.

By the way, there is one thing I don’t understand: Super Sampling creates additional frames, for example 2 extra / intermediate frames for each of the original frames. But my frame rate is 25 fps. So how can these extra / intermediate frames enhance my video?

Thank you,
Joran

I might be awefully wrong here but as I understand it Supersampling only works when the media is Vegas generated.
For example, generate some text and using event pan/crop rotate it. It'll look not very realistic as there's no motion blur, the 'shutter' speed is infitisemal. By applying Supersampling Vegas will generate say 4 intermediate frames between the real frames, simulating the effect of a slow shutter speed.
Bob.

Also, super sampling is a temporal effect only and has nothing to do with resolution.

Yes supersampling only works on generated material. That is why it is used in conjunction with motion blur which is a generated effect. The idea on using it for resizing is that motion blur generated between frames will smooth out the resizing more than simply doubled or interpolated pixels. It makes sense to me in theory, I just haven't been able to make it work in practice. If you're not combining supersampling with motion blur however, you really aren't doing anything other than increasing your rendering time.

download some Quicktime media from one of the historical sites that is 12fps. Jump this to 24 or 30 fps both with and without SS. You'll see a difference between the two. And while it isn't supposed to have spatial characteristics, going from square pixels to non-square pixels also shows improvement with SS when shifting these framerates.

Again, changing the pixel shape (or aspect ratio) involves generating new data in the frames. So does increasing the frame rate. That is why supersampling works in these cases.

There's really only one party who can explain this and to date they haven't been very forthcoming. Having to work out just what something does by almost reverse engineering it seems kind of silly to me when all it'd take is for the guys who wrote the code to document what it does.

One good way to check any results is to use the Invert and Add method which Vegas does very nicely. This avoids the problems of subjective evaluation.

Bob.

... and the functionality of this less common feature should just be in the manual!!!

Actually, going from 16:9 to 4:3 I'd just crop the sides and do a 4:3 rerender.

Actually the Madison guys have posted in here exactly what the function is intended to do.

Here are some of the Sony posts on the subject:

SuperSample 1

SuperSample 2

Seems a clear few paragraphs explaining this in the Vegas manual would save everybody a lot of reading, especially having to read these past pus-filled posts.

Hey, who you callin' pus-filled??

Make me think of the John Belushi cafeteria scene in "Animal House."

"What am I?"

"A Zit!!"