[Theora-dev] using theora

[John Kintree]

Having built libtheora-1.0alpha3, I tried it out on a 46 second long, 
161 MByte raw dv format video clip.

First, I copied these lines from ogg-theora-microhowto.html
<snip>
mplayer -ao pcm -aofile stream.wav -vo null file_to_be_encoded.avi &
 mplayer -vo yuv4mpeg -ao null -nosound file_to_be_encoded.avi &
 encoder_example -v 1 -a 1 stream.wav stream.yuv > theora_file.ogg
</snip>
into a file named "encode", substituted "file_to_be_encoded" and "theora_file" 
with my own filenames, and gave "encode" execute permission.

It took about 15 minutes to encode my dv format clip, the resulting .ogg file 
was 6.8 MBytes, and when I played it with mplayer, the audio sounded good, 
but the video looked more like a slide show.

When I used MainActor to encode that same dv format clip into mpeg2 format, it 
took about 2 minutes and 50 seconds, the resulting file was 25.3 MBytes, and 
both the audio and video were close to the quality of the dv format file.

When I used MainActor to encode that same dv format clip into mpeg1 format, it 
took about 2 minutes and 20 seconds, the resulting file was 10.9 MBytes, and 
the audio was good while the video has not as sharp as the mpeg2 version.

I've got a 750 MHz AMD Duron with 128 MBytes of RAM.  There was a message when 
running encode that said compiling mplayer without CPU detection was a 
possible optimization.

What might I try to get ogg theora output that looks at least as good as mpeg1 
instead of like a slide show?  Will optimizations reduce the encoding time 
from 15 minutes to less than three minutes?
Regards,
John