Multipass VBR - how does it work and how much would be overkill?

[Chetwood]

I'm currently ripping some Buffy episodes with 2audio tracks and 2 subtitles to fit on 1 ep/800 MB and I'm using CCE 2.50 with 5 pass VBR. Does this make sense? I mean, my computer is in the basement so it can run all night without bugging me at all. Still I'd like to know if I should use more/less passes and how's it done anyway.

After the first encode what does the second pass do? Compare the results to the new pass and check which frame should get more bits or something? Would I get better results on encoding with 7 pass? Or would it make sense to use more bits for the video and less passes?

Thanks for any info.

[Pyre8]

I am no guru but it seems to me that for a given bit rate, you set the number of passes to get the quality you desire for that bit rate. Each pass seeks to optimize the distribution of the bits and unfortunately it is a iterative process (i.e. CCE can't figure out the optimal distribution without doing the distributions - wouldn't that be cool - one pass with the optimal distribution ). Beyond the 3rd pass, it is marginal improvements, not enough to get me worked up over getting it through another pass. it may also depend on your source perhaps, a bad source, lots of action/movement may benefit from additional passes. I usually choose between 3 and 4 depending on content and source (and of course, the time I go to bed:) then leave the PC to chug away while I sleep).

[wmansir]

If your keeping within spec your capping your max bitrate at ~2700kbps including audio, and since each ep averages around 43 minutes your average bitrate, including audio, is around 2500kbps.

Your average is so close to the max, having multiple passes isn't going to get you much improvement since CCE doesn’t have much head room to add more bits. Even the difference between CBR and VBR would be very small with such a high average bitrate, so I don't think more passes is necessary.

I'm also doing Buffy S2, using 3 pass VBR. There’s nothing too difficult about this title, just make sure your using IVTC. Also, since S2 was shot on 16mm it's pretty grainy. To help it compress better I use a touch of temporal smoothing. But that’s a personal preference, some people don’t like it.