Would this work?

[sebazvideo]

I came up with this idea with the goal of cutting down my encoding time. Let's suppose I have a timeline in Vegas that is 90 minutes long, so instead of exporting the full thing to a lossless intermediate avi to open in MeGUI, I do this:

1) Export the first 45 minutes to one avi and the next 45 minutes to another avi.

2) Send the second file to a second computer through the network

3) Encode both files to x264, using the same exact builds of x264, Avisynth and MeGUI, and using the same exact x264 parameters.

4) Bring over to my computer the x264 encode from the second computer

5) Join them in TS Muxer

If this worked, it would cut down the encoding time in half (not counting the time it takes to transfer the big lossless file to the second computer and starting the encode), but I wonder if it would still produce a legit file for blu-ray, because of the cut in the middle. For starters I would assume that I would have to encode both using closed GOPs, but how do I do that? I don't remember seeing any parameter in MeGUI or x264 command line to set this.

But even with a closed GOP, would this work? I will do a test later with two 5 minute files, but I wanted to ask here to see if I'm having a great idea or a totally impossible one and this wouldn't work in a million years.

[poisondeathray]

One potential problem I can see with this is tsmuxer - which overwrites HRD values . So it might work for some sources, not for others

[sebazvideo]

One potential problem I can see with this is tsmuxer - which overwrites HRD values . So it might work for some sources, not for others

What are HRD values? Are you referring to those two items that are checked by default for h.264 streams, "Add picture timing info" and "Continually insert SPS/PPS"?

[poisondeathray]

Yes, it might work if you uncheckmark? not sure. I know the search function sucks here, but shon3i has written a few times about tsmuxer + HRD issues.

But wouldn't joining 2 streams give different buffer values anyway (compared to 1) ? I am guessing that if you look at a buffer analyzer the values would be different for the joined case vs. single encode

[sebazvideo]

But wouldn't joining 2 streams give different buffer values anyway (compared to 1) ? I am guessing that if you look at a buffer analyzer the values would be different for the joined case vs. single encode

That's a little over my head too. Are you talking about VBV Buffer? If both files have the same VBV buffer, how would the two joined ones have a different buffer? I'm asking because I don't know, it might be perfectly possible.

[poisondeathray]

Yes I'm talking about buffer and hypothetical reference decoder = HRD.

I'm not an expert at this, but I'm guessing it would be different, because 2 encodes would have different initial occupancy values. A single encode would have it's buffer filled at that "midway" point. If you look at a single encode and buffer analyzer, it looks like a bell curve as the buffer fills and tapers off at the end. I can't see how joining them would give identical values. (Think of putting 2 curves end to end, instead of 1 bell curve). This is just my guess, I would wait for the official experts to chime in

[Guest]

See #9 here:

http://forum.doom9.org/forum-rules.htm