It all comes down to bitrate. Bitrate is the standard unit-of-measure for setting quality in an encoded MPEG. For SVCD we have a max bitrate of 2600kbits, but for most encodes, our average bitrate will be somewhere around 1500-1700kbits. What most people don’t realize is that bitrate is applied per second not per frame. This means that a 23.976fps 1 second clip encoded at 1700kbits will have 6 fewer frames to allocate bits to than a 29.97fps 1 second clip encoded at the same bitrate. More simply:

23.976fps @ 1700kbits = 70904 bits per frame, while

29.97fps @ 1700kbits = 56723 bits per frame.

To put this in perspective, let’s say that you encode a movie at a bitrate of 1700kbits @ 23.976fps. If you encode the exact same movie at 29.97fps, you’d have to use a bitrate of 2125kbits in order to get the same kind of quality as the 23.976fps encode. That’s an extra 425000 extra bits! Think of how much more quality you can pack into your 23.976 encode!

"Okay, but I thought you said that I can only play back my SVCD at 29.97fps?"

Right. But you can still encode your SVCD at 23.976fps. All you need to do is set the 3:2 pulldown flag in your encoder. This way, your video is still stored at 23.976fps goodness, but it has special flags embedded into it force playback at 29.97fps.

Another advantage of converting back to progressive frames is that it makes resizing much easier. Even if you’re still encoding at 480 lines, you’ll need to resize if you’re converting a 16:9 anamorphic encode to the 4:3 aspect ratio required by SVCD players. If you try to do this without undoing the pulldown, you’ll end up with a very badly-interlaced mess.

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

這是DOOM9連到的文章Myths About Interlacing
不過我還在研究中...........