Idea for DivX multiple CPUs

[tritium]

Anyone have an idea on what the output size/quality would be like, if i set a job to encode half of the movie on one CPU, and the other half on the other. Eg.

Say we have a video MyChristmasBreak2004.
Its an hour long for convenience.

I have two copys on disk, one on C:, and the other on D:
I setup a job, and get tell it i only want the first 30 mins encoded(c:\).
I setup a second job, and tell it i want the last 30 mins encoded(d:\).

With a bit of luck, this should allow the box to run at 100% CPU.
At the end, join the two manually, and mux in the audio.

Questions:
Will this work? I cant see a reason why not.
What will the output be like. I would suspect that it would be slightly higher filesize / lower quality, as the codec would not be able to allocate the same number of bpf from a 'half encode' rather than the full lot.
I'm going to have to get down to the basics, and use VDub for this, which i'm not all that used to tbh. Before i start to play with it, can someone lend thoughts to whether they think its worth it?

Thanks.

[Brother John]

I would suspect that it would be slightly higher filesize / lower quality, as the codec would not be able to allocate the same number of bpf from a 'half encode' rather than the full lot.
If you split to equally long segments, BPF should be the same because you have half the space and half the length for each segment.
But BPF values are not that important. Quality is much more determined by the distribution of high and low motion scenes throughout the movie. If the high/low motion distribution is similar in each segment, quality difference will be small because even in a full length encoding each half would get an equal share of bitrate.
If – let’s say – the first half has 70% high motion scenes and the second half only 20% splitting in half is not a good idea. The first segment should be shorter because it’ll require more bits.

I think if you choose your splitting point carefully you should be able to compensate quite well for the quality disadvantage the splitting involves.

Finally, to reach a conclusion you need to do the encoding twice (segmented and on the whole), look at the quality, measure the speeds and decide if the difference really is worth all the hassle.