I'm currently using vmesquita's guide but sometimes the final file sizes is too small even when the sampler size is only 1% smaller than the calculated value. I see a lot of suggestions on this forum but mostly they use TMPGEnc, KVCD and PAL formats which I don't use. Is there an alternative to vmesquita's script for a more accurate file prediction? I'm trying to encode several episodes of Divx cartoons.

Do you really need to open several theads on the same subject ?