Subject: Re: [linux-audio-dev] latest audioengine snapshot (laaga proposal)
From: Paul Davis (pbd_AT_Op.Net)
Date: Fri Jun 15 2001 - 04:58:57 EEST
>Yep, agreed it is more complicated, but, its less of a special case, and
>it allows for multiple timebases cleanly. On the downside it makes it more
we don't *want* to allow for more timebases cleanly, do we? isn't the
whole point to have everything running in sync? if there are two apps
that wish to cooperate, we could have "well know port names" for apps
that publicize a timebase, but thats not the same as "the system
timebase".
>I was imagining that the library would provide a timebase decoding
>function, or that the buffer pased to process could just be a c
>struct or whatever.
i wouldn't want anything other than audio frame time at this level of
the library, though i'm slightly partial to a "time since interrupt"
value, which helps synchronize visual display when the interrupt
interval from the h/w is long. other time specifications (SMTPE, MTC,
etc.) are all to domain-specific and low resolution to be usefully
defined inside of LAAGA. i think.
>Although MIDI Clock is not a proper timebase, applications often try to
>sync to it in the same way as they do MTC's bars:beats:frames mode, though
>I'm not sure what the low level representation of that is.
there is no MTC "bars:beats:frames" mode. thats a (G)UI construct
presented by an application.
MIDI itself defines song position pointer messages which includes a
beat concept, and these tend to be used by things like the Akai
MPC-XXX "MIDI production devices". They specify the number of beats (1
beat = 6 MIDI clocks) that have elapsed from the start of the "song",
and the MIDI spec says "is used to begin playback of a sequence from a
position other then the beginning of a song". Note: the smallest song
position change is 1/16 note. The MIDI spec has good instructions on
computing the time value given the song position pointer. Using SPP is
a little tricky, but commonly done. Some s/w sequencers may use it too.
--p
This archive was generated by hypermail 2b28 : Fri Jun 15 2001 - 06:42:56 EEST