A newbie's guide to MPEG fundamentals

[doug1937@hotmail.com]

In my ATI AIW digital VCR I can customize my MPEG capture settings. I've always used the defaults, but I'm wondering:

1) What use are the B and P frames? The default uses 4 P frames and 2 B frames, and of course the I frames. Would I get better capture quality by using more P and B frames? Or less? I've seen suggestions to use only I frames...

2) There's a deinterlacing option...I've always heard interlacing is bad...should I use it?

3) What about Motion Estimation Search ranges? Are they worth manipulating? What effect do they have?

Thanks!

[Sulik]

1)
I frames are like JPEG images
P frames can use forward prediction (more compression)
B frames can use forward AND backward prediction (even more compression)

The more B frames, the better the quality at the same bitrate, but past the 4:2 setting it doesn't make much difference in quality.
The drawback of P and B frames is that it requires much more CPU power to compress (ie you may drop frames), especially with the highest quality of motion estimation.

2)
ALWAYS deinterlace with MMC, since MMC only does progressive capture
(Unless your source is VCD or progressive, which is rarely the case)

3)
Motion Estimation search range -> the larger the window, the better, but it doesn't seem to make much difference in MMC (leave it as it is).
The motion estimation level (0-100%) is what has the most effect on the capture quality (besides the bitrate, or course), at the cost of extra cpu usage.