3 (or more) passes in QuEnc, FreeEnc, FFMPEG, and Mencoder (would you use it)

[Peter Cheat]

I am considering implementing multiple passes for avcodec library. I am just interested to know if anyone would use it, if it was implemented.

[dragongodz]

3 could be advantages but beyond that would be questionable. again using CCE as an example , after 3-4 passes gain of quality is next to nothing(if at all, placebo effect ?). i would prefer a better working 1 and 2 pass first.

[jorel]

a particular oppinion of a simple user: time to encode is not problem if 4 or more steps can give quality is the target.(from time to time cpus are faster)

[EpheMeroN]

I would love a third or even fourth pass only if it would ensure even more quality on the final video. Otherwise, just keep making the 1-2 pass better =)

[Bogalvator]

I voted no for this. Being an XviD user as well as MPEG-1/2, I have seen what can be done with only 2 passes, so any more is unnecessary IMO.

[SILICON]

If the 3 pass mode give more quality, I'll use it.

But I think that the broken bitrate control is more interesting.

[Bogalvator]

I think I actually missed the point with the 3 passes thing. It seems the main improvement (so far) with Peter Cheat's new code is that the bitrate no longer spikes over the specified max so much. I don't know how often XviD does it, but since I only play back on a computer it's never mattered.

If three passes can guarantee never going over max bitrate, and hence guarantee smooth playback on a hardware player, then yes I would use it (though I'd prefer to see the same with only 2 passes ;))