[vorbis-dev] Idea for gapless tracks

[Beni Cherniavsky]

Gerry wrote:
> 
> Why not insert some extra realtime-calculated ADU's between the last of
> track 1 and first of track 2 ? Then nothing of the music would get "lost",
> and if the crossfading need more time to sound good, just insert extra
> ones.. These would not be part of the vorbis files, but calculated by the
> decoder.
> 
> I know pretty much nothing about this stuff, so if this is a really, really
> bad idea that doesn't make much sense at all, just ignore it :)
> 

Maybe I wasn't formulating this well but that's exactly what I kept
proposing: somebody mentionaed that he wants sample accuracy, so I
proposed extrapolating both tracks by the length of the overlap,
so that the end of track 1 is mixed with pre-extrapolation of track 2
end the start of track 2 is mixed with post-extrapolation of track 1.
The mixing transitions smoothly from track 1 to track 2:

+-------------+--------------
| ... Track 1 | extrapolation
+-------------+--------------
___________                     100% track 1
           \_
             \
              \
               \_
                 \___________   100% track 2

--------------+-------------+
extrapolation | Track 2 ... |
--------------+-------------+

The shape and lenght of the transition is a matter of taste, with the
objective being making it seemless with minimal artifacts.  For best
results you might try to make the length (and shape?) different at
different frequencies.

This all belongs probably in the player.  If you want continous sound
from the same recording which was split into separate Ogg Vorbis files,
you should inplement some metadata information in the playlist/whatever
which will instruct the player to enable this feature.  This seems
outside of the Ogg Vorbis specification and can't be fit in probably.
The only thing I can think about is to enable this feature automatically
if the serial number of track 2 is exactly 1 higher than the serial
number of track 2 - and make the encoder put such serial numbers.

-- 
Beni Cherniavsky <cben at crosswinds.net>
                 (also scben at t2,cben at tx in Technion)

  No, No! You're not thinking; you're just being logical.
                                             - Niels Bohr

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.