> -----Original Message-----
> From: Michael Cheng [SMTP:[EMAIL PROTECTED]]
> Sent: den 12 september 1998 09:43
> To: [EMAIL PROTECTED]
> Subject: [MP3 ENCODER] mp3enc idea
>
> I'm a bit shirty at fraunhofer at the moment; all we've done is
> clean up
> the ISO source, made our changes freely available, and
> redistributed it.
>
[Jan Peman] I think most of us are a bit mad at
fraunhofer.. :/
> I've got an idea. It's a little evil and underhanded, but it's
> probably
> quite do-able:-
> Given that: mp3enc (the demo version) only encodes 30second
> samples.
>
> Do: - Split a soundfile up into 30 second samples
> - encode each of the samples
> - join the 30 second mp3s back together into one mp3.
>
> I can't see any reason why this isn't do-able.
> The hard part will be fixing the final header of the final frame
> of one
> mp3 to match the start position of the first frame of the second
> mp3.
> Other things that may need thinking about:
> - bit reservoir may need to be purged.
> - is there anything special about the first frame of an
> mp3?
> - is there anything special about the last frame of an
> mp3? eg
> terminating bits.
>
> Can anyone think of why this would not work?
>
[Jan Peman]
I can't see why it shouldn't work.. I just tried to split
a file in 20secs parts,
encoded them and just joined them together (using cat >>
test.mp3) and
it was playable.. some glitchs every 20 sec:s though :)
But splitting the
input-file in even frames (1152 samples?) , with an
overlapping 1152
samples from the previous splitted part, then encode the
files, removing
the overlapping frame and just joining them again to a
whole mp3 should
remove the glitches.. This might work.. if so, no need to
worry about the
bitreservoir..
> suggestions welcome
> mikecheng
> (who gets holidays in 1 week)
[Jan Peman]
/Jan Peman (who needs to configure his outlook mailing
program)
