Subject: Re: [linux-audio-dev] Question to developers of sound editors.
From: Tom Pincince (stillone_AT_snowcrest.net)
Date: Mon Jun 04 2001 - 22:07:20 EEST
> So my question is... Is any kind of cohesiveness developing on the sound
> editor front
Much of what we have known regarding the traditional functions of a
sound editor have been made obsolete and have been replaced by the
functions found within an edl based editor/arranger. With very few
exceptions, such as noise reduction and dc offset removal, soundfiles
are simply not edited any more. If the arranger is integrated with a
mixer, then the common dsp treatments, such as compression and eq, are
applied in rt as mixer plugins. Also with the edl approach there is
little or no difference between an editor and an arranger, so a
multitrack app that offers arranging will have no need for a separate
editor. The method of assembling tracks by referring to sections of
multiple soundfiles has many subtle implications. While I don't code, I
do have countless hours assembling playlists and sequences using a
number of edl based apps. This approach to audio manipulation resembles
midi sequencing much more than it resembles traditional soundfile
editing. This suggests that there may not be one definitive approach,
since editing by sequencing can be implemented flexibly to accommodate
working style preferences. Essentially it is about defining
inter-related building blocks such as regions, sequences, sub-sequences
(mirrors and clones), playlists, and tracks, and allowing as much re-use
of constructions as possible. Any developer who is not familiar with
this approach will have to "do his own thing" first to become familiar
with this method. Probably the only place for cohesive development
right now is in establishing naming conventions, and this is not as easy
as it may seem. I have been privately sending Paul a stream of
(occasionally useful) insights and suggestions on this topic. One thing
that we agreed upon was that, in instances where a particular aspect of
ardour is implemented *exactly* the way that it is in protools, the
protools naming convention should be used. We both downloaded the
protools manual, and came up with *completely* different definitions for
what protools refers to as a "region".
It seems to me that the laaga discussions will result in a model that
allows multiple independent editor/arrangers to function as a source for
a compliant mixer, so establishing naming conventions for standard edl
techniques seems to be a wise move.
Ideas?
Tom
This archive was generated by hypermail 2b28 : Mon Jun 04 2001 - 23:53:14 EEST