Standard 2-Pass vs. FFDshow version
 

The 'official' XviD version only creates a stats file during the first pass, then uses that stats file in generating the second pass. Not so for the FFDshow version...it creates a working video file on the first pass, and creates its second pass file from this somehow...not too sure how the FFDshow version works, TBH.

So my questions are:
1) how the FFDshow version really works.
2) What is the quality of the first pass file in an FFDshow XviD two-pass process...can it be used for a "low-loss" pass in cases where maintaining video quality isn't that important (ie: the source sucks anyways)?

Offical also creates video but you need to configure it to not discard it. Press more next to Encoding type (when it is set as First pass).

Didn't know that, thanks :)

for 2) The first pass is a constant quant 2 encode, iirc.

CU,
lamer_de