Log in

View Full Version : Can I merge multiple clips after encoding with meGUI


LordHeinrich
23rd January 2009, 03:55
I have a Quad Core CPU and meGUI is able to use 4 workers to encode 4 jobs at parallel. This is great because I can encode videos 4x faster when I have four videos to encode.

I was wondering though, is it possible to split the source video into four parts and then encode them using all 4 workers and then merging them after the encode?

If not, is there any way to get all 4 Cores to work on encoding a single video?

linyx
23rd January 2009, 04:17
If you are using x264, it automatically detects and uses all of your cores.

LordHeinrich
23rd January 2009, 05:15
Ok, thanks! I would not have noticed it had you not pointed it out. I never bothered to look at utilization when it was encoding. It does go to 100% utilization when encoding, but about 50% when analyzing. On that note, is there a way to utilize the four cores at 100% when analyzing? I have a Q6600 at 3.2GHz.

Thanks

Dark Shikari
23rd January 2009, 05:27
Ok, thanks! I would not have noticed it had you not pointed it out. I never bothered to look at utilization when it was encoding. It does go to 100% utilization when encoding, but about 50% when analyzing. On that note, is there a way to utilize the four cores at 100% when analyzing? I have a Q6600 at 3.2GHz.

ThanksThe first pass normally uses less CPU because the encoding is much faster, and so it can get bottlenecked by more things. The most common culprits are decoding the input video and --b-adapt (--b-adapt 2 in particular can bottleneck very fast encodes quite heavily). Of course, using lower --b-adapt values will decrease quality.

LordHeinrich
23rd January 2009, 06:19
Thanks for the reply.

Does that mean that the computations/analyses are too complex for current CPUs and causing the bottleneck, or is another component responsible for this?

I do use "--b-adapt 2" btw. I don't mind sacrificing speed for quality.

Dark Shikari
23rd January 2009, 06:20
Thanks for the reply.

Does that mean that the computations/analyses are too complex for current CPUs and causing the bottleneck, or is another component responsible for this?A "bottleneck" means that something is singlethreaded and everything else is held up waiting for it.

Currently, the lookahead x264 does for frametype decision is singlethreaded (b-adapt).

LordHeinrich
23rd January 2009, 06:39
Thank you