[xvid] Segment first pass only?

[jackiz]

Hi!
Is it (in win or linux) possible to encode the first pass in segments, then merge them and encode the 2nd pass as a whole?

Answers is really appreciated! :thanks:

[jackiz]

No, I do not think that is possible.
(To be honest, I even find it difficult to provide practical meaning of this. That is, you need to file A, B, C, make the first pass for each and combine them into a file D, and make a second pass?)

But if you write in more detail about your specific case, look for the best solution.

Lets say i want to record a TV-show and encode it as fast as possible, it would probably air kinda like this:
footage 1 | ad 1 | footage 2 | ad 2 | footage 3
I want all the footage, but no ads, so i could start encoding footage 1 1st pass when ad 1 is airing, and encode footage 2 1st pass when ad 2 is sending etc.
Then encode the 2nd pass after footage 3 has aired, that way i can chose the exact size of the result file without messing up the bitrate.

[ChaosKing]

Statsreader 2.1 kann merge statsfiles

[Didée]

In the past I've done such partial 1st-passes on a few unintended occasions (power failure during 1st pass). Basically you can take the different log files, and stitch-together the analysed data with a simple text editor. However that's walking on whacky ground, as one can easily create "off-by-one" (or by-two, or by-three) errors. If so, the later sections are borked, because the frametype decisions made during 1st pass don't fit anymore to the video data.

Considering that Xvid on a modern PC can do a 1st-pass at "hundreds of FPS", a not-segmented 1st pass of a 40-minutes-show takes only a few minutes - say, "the length of a cigarette". So, what's the point? Why such a hurry? Can't help, but it smells like scene releases ...

[jackiz]

Statsreader 2.1 kann merge statsfiles

In the past I've done such partial 1st-passes on a few unintended occasions (power failure during 1st pass). Basically you can take the different log files, and stitch-together the analysed data with a simple text editor. However that's walking on whacky ground, as one can easily create "off-by-one" (or by-two, or by-three) errors. If so, the later sections are borked, because the frametype decisions made during 1st pass don't fit anymore to the video data.

Considering that Xvid on a modern PC can do a 1st-pass at "hundreds of FPS", a not-segmented 1st pass of a 40-minutes-show takes only a few minutes - say, "the length of a cigarette". So, what's the point? Why such a hurry? Can't help, but it smells like scene releases ...

Thanks for replys!
How do i use the statfile after stitching them together? i have never encoded passes individually before.
I just like fast encoding