[MEncoder-users] Backup policies

[infernix]

James Hastings-Trew wrote:
> I've done some playing with a variation of this command line and I do 
> have to say, in situations where you don't care about the final file 
> size or bitrate, this does produce a very high quality encode in about 
> half the time of a two-pass.

The speed should be the same for a CRF run and the second pass of a 
2pass run; the speed benefit is gained by not running a first pass at all.

Now that you've experimented a bit more, you'll want to look at some 
possible patches to x264 that might be of further help. I'm using the 
following patches:

x264_aq.5-654.diff
x264_thread_pool.04.diff
x264_faster-dia.r680.diff
x264_cachesplit.02.diff

You'll find them here: http://forum.doom9.org/showthread.php?t=130364
and here: http://akuvian.org/src/x264/

Thread pool is only relevant with multicore (4+). cachesplit is fairly 
new and is about code optimisation (same for faster-dia). the AQ patch 
is adaptive quantizing and is supposed to reduce blocking/banding in 
background scenes (it spends more bits on it depending on the used 
aq-strenght; I use 0.3 mostly).

There's more patches out there as well. If you're conservative you can 
just stick with SVN, it'll do fine.