[mythtv] mythtranscode: potential optimization
Mark Steckel
mjs at qwsa.com
Tue Nov 27 17:49:36 UTC 2007
At 12:30 PM 11/27/2007, you wrote:
>mythtranscode behaves differently depending on whether you told it to do
>a normal transcode which is what Chris's example is doing or a lossless
>transcode with the --mpeg2 option which it looks like you are doing?
Ah... You are correct. I am specifying --mpeg2 whereas Chris' example does not.
>The
>last time I looked the lossless transcoder scans though the entire file
>at the end to create some sort of map file that as far as I know isn't
>actually used by anything. It may also scan though the entire file
>during the actual processing as well. The lossless transcoder is usually
>pretty quick though?
There is a *big* time difference. Using the same 2.2GB mpeg input file
with --mpeg2: 45 seconds
without --mpeg2: 0.5 seconds
Since I am interested in the video quality of the output file I
presumed that --mpeg2 was necessary. However, I ran 'cmp' on the
resulting files produced with and w/o --mpeg2, and it showed no
difference. So, at least in this case I can run mythtranscode w/o --mpeg2.
However, while I have a solution to my particular issue, it does
appear that using --mpeg2 causes extra work to be performed.
Thanks for everyone's assistance!
Mark
More information about the mythtv-dev
mailing list