Whilst making a sample of a video I find that there are thin black borders on the sides of the video as well as at the top and bottom.

Any help on rectifying this would be helpful.

Btw, I'm converting a 640x304 video using 352x240 template.

Yes that's normal. You should have a script using gripcrop, and there is a parameter called "overscan". If set to one, you have 8 pixels of black border on 4 sides. with 2, 16 pixels....
But don't worry, they won't be on your TV.

Well, the idea is, ... why using valuable bitrates to encode a part of a movie you will not see, so we put these black borders which give us a little, little advantage to receive more CQ (like 1-3%). And also to receive best mathematical conditions for the mpeg encoding process.

In case of the Tv-advantage it referes to the usage of a "cache" in Tv-Productions, thats also called "Tv safe area" you shurely noticed in some DVD Authoring apps when creating Menues for example.
The cache normally takes about 10% oft the screen at all sides.

Well to explain why we use 16pixels black borders on the left and the right I'll translate a part of a well done german written page.

NOTE: The copyrighter of this explanation is Angel an the original URL is
http://www.uni-kassel.de/~eckhardm/hq.htm
So I did only translate this section! Maybe you can use google's translation engine to translate the other contents of his page

So here it goes:

---------------------- http://www.uni-kassel.de/~eckhardm/hq.htm ----------

Explained in a case of 480x576 encoding but mpeg encoding is MACROBLOCK based and so you can use this example also on other sizes.

One Macroblock is exactly 16x16 pixels and if we fill out one Macroblock with black contents, we shurely will get better compression!
So if you just add an 8pixel boarder at the sides the peg routine will still encode the WHOLE 480 horizontal information!

Here an example based on an "P" Frame:
(the green macroblock will be missed out (motion detection) and the black macroblocks will NOT be encoded.

Size 480x576 (464x320) which will be letterboxed by using 8 pixels on each side afterwards by adding the black borders.




Size 480x576 (448x320) which will be letterboxed by using 16 pixels on each side afterwards by adding the black borders.

You can see very well that the border hast to be ONE Macroblock (16x16px) so that it will NOT be encoded, which lets us enjoy more compression





And why we crop the black boarders from the source exactly as we can?

Example:
2.35:1 Letterbox (Resize : 448 x 320 / AddBorder 16/128/16/12




As we see the macroblock encoding advantage will not be at these exat "vertical" cropping settings so we do an optimization:

Optimized:
2.35:1 Letterbox (Resize : 448 x 318 / AddBoarder 16/129/16/129)



You see that by doing this the last macroblock line between the black vertical borders and the movie is handled optimal. Ok, there will be a minimal aspect ratio error, but you won't recognise it.

------------------------------------------------------------------------------------

Well these optimizations are very detailed, but if we use all advantages including GOP optimization and max bitrate optimization, we'll receive the best compromise between quality and compression.

So please visit
http://www.uni-kassel.de/~eckhardm/hq.htm
he's the author
and use Google's translation engine to check out his very interesting guides. Cause this is only a translated part of them!