|
|
Posted by sbt on 11/16/97 11:36
In article <La_uf.108877$J11.2325232@wagner.videotron.net>, NRen2k5
<napsterneorenegade@hotmail.com> wrote:
> sbt wrote:
> > In article <cOYuf.108549$J11.2266844@wagner.videotron.net>, NRen2k5
> > <napsterneorenegade@hotmail.com> wrote:
> >
> >>
> >>Actually, I'd like to see how some Macs compare to my PC when it comes
> >>to encoding DivX movies. Purely out of curiosity.
> >>
> >
> > Having done a lot of that sort of thing, we would need to start with
> > the same content and have the same target framesize and framerate. Even
> > with those conditions set, it will depend largely on which DivX encoder
> > we each chose to use. If I use the one from the folks at DivX, I get a
> > different encoding speed than I do from the one that comes with Toast,
> > which is still a different speed from either the mencoder or ffmpeg
> > DivX encoding tools in ffmpegX. I'm reasonably certain that similar
> > differences between encoders exist on the Windows side (they do on the
> > Linux side, too).
> >
> > Starting with "DV Stream" content of 58 minutes duration (came from a
> > miniDV camcorder), I got a 1236 kbps 720x480 DivX (NTSC Film framerate
> > of 23.976) in just under 42 minutes using ffmpegX on a dual 2GHz G5.
>
> I'm impressed. What speed/quality setting (if there was any choice) was
> the DivX encoder was set to?
>
> In fps processed, looks like you're getting about 33? I get about 10, at
> the "insane" setting, anyway.
>
You're allowed to make modifications, but I took the default settings
(that's where the 1236 kbps was chosen) and, yes, that came to about
33. Of course, I'm on a dual processor and the software takes advantage
of that fact, running separate threads on each processor. I could tell
it to just use one processor, but I don't know what sort of rate that
would get. I generally have it do any encoding when I take the dogs for
a walk, or go downstairs to cook/eat, so I figure the encoding should
get as much processor access as possible since I'm not impacted.
--
Spenser
[Back to original message]
|