would this get me better compression?

[lancer]

I have six files linked using an avisynth script. I ran it through virtualdub and created an interim file using fast recompress which was 35gb in size. after I'd converted it into a 4.2GB M2v @ 4600 average bitrate using TMPGenc I noticed that there were some dropped frames that I'd missed.

I grabbed the relevant segment and spliced it into the interim by putting it through virtualdub via a script so I again had an interim file 35GB in size. however this time when I ran TMPGenc at it, the file at 4600 average came out at 3.9GB in size which foxed me.

I've run TMPGenc again and this time I got a 4.2GB file using 4900 as the average which is good, but I'm struggling to figure why.

by recompressing the interim file when I spliced in the fixed segment, has this done something?

[digitalman]

This is probably way off base, but did you encode in TMPGEnc both times to and from the same exact partitions or drives?

[lancer]

yup. this is why it doesn't track because the redone interim file is the same size as the original right down to the last byte and yet I am getting 200 more per second average.

not that I'm complaining but I just thought it a bit odd. the only thing I can think is that it's made it more compressible, hence the question.

this was using faerydust by the way on both occasions.

[Acaila]

Encoding twice with a lossy codec causes it to throw information away twice also. Hence the smaller filesize (or higher bitrate required to get to the same filesize). This also reduces the quality btw although it might not be noticable if the bitrate is high enough to begin with.