[LAD] Announcing: ecaedl.pl Edit Decision List Capable Audio Playing

Robin Gareus robin at linuxaudio.org
Tue May 25 20:30:13 UTC 2010


On May 25, 2010, at 8:02 PM, Niels Mayer wrote:

> What is the particular advantage of EDL, versus something more standard:
> http://www.w3.org/TR/SMIL/
> http://en.wikipedia.org/wiki/Synchronized_Multimedia_Integration_Language ??

"more standard" is relative and depends on the point of view.

EDL (as in formats such as CMX3600) is certainly dated, but still common in Film-production (it can be read by a human with a scissor & tape).

Basically that's why I asked about the use-cases or the goal of this 'exercise'.

From what I can tell about looking at Drew's perl script he's not trying to interpret 'classic' EDL; and just used the term to get the idea across. SMIL would certainly be the better option, though it may be overkill as well..

> Mplayer (and thus http://en.wikipedia.org/wiki/K-Multimedia_Player ,
> smplayer, and gnome-mplayer) supports SMIL although the support is
> rudimentary. SMIL and other means of setting clips&playlists is
> available in the flash-based 'jwplayer' http://www.longtailvideo.com
> "Attribution-NonCommercial-ShareAlike 3.0 Unported" ( in action:
> http://nielsmayer.com/xwiki/bin/view/Macros/JWPlayerJSPL_Test and
> http://nielsmayer.com/xwiki/bin/view/Exhibit/NPRpods3 (work in progress) ).
> 
> The other issue w/ doing it "in the player" is how well do the given
> players handle a playlist. Most of them, even if accessing the same
> media in different cue-locations, do not do a very good job with going
> through "clips" with seamless transitions between the clips. However,
> it's readily possible , with a bit of pre-buffering and pre-fetching
> that allows seamless synchronized playback .... just a small matter of
> programming...

been there, done that and using libsndfile this just works and is pretty straight-forward.

However other decoder libraries are not very predictable when it comes to accurate seeking.
In particular ffmpeg's av_seek_frame() and libqt's  quicktime_set_audio_position() are not very precise.
The only reliable way to do this correctly with said libs would be to always start decoding from the beginning or to stick to a limited set of codecs. OTOH you may or may not care about +-1920 audio-samples.

> (i waste way too much time abusing my own software for
> unintended purpose -- "internet sampler and looper"
> --http://nielsmayer.com/ts-episode-timeline.png  )
> 
> Niels
> http://nielsmayer.com
> _______________________________________________
> Linux-audio-dev mailing list
> Linux-audio-dev at lists.linuxaudio.org
> http://lists.linuxaudio.org/listinfo/linux-audio-dev




More information about the Linux-audio-dev mailing list