Re: [linux-audio-dev] XAP and these <MEEP> timestamps...

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] XAP and these timestamps...
From: Tim Hockin (thockin_AT_hockin.org)
Date: Fri Dec 13 2002 - 11:11:34 EET


Maybe I don't get all the issues, but the following thoughts occur to me wrt
time.

* A tempo delay that starts a delayed sound just before the host loops the
play back in (musical) time does not really want to know about absolute
musical time, but DIFFERENCES in musical time. More precisely, it wants to
be aware of the passing of ticks (where the duration of a tick is defined by
the host).

* What if plugins that care about musical time have a TICKS control.
Whenever the host crosses a tick boundary, it sends a (sample-accurate)
event to the TICKS control, with the value being the # of ticks
(ever-running time or track-time?). Plugins can look at
host->ticks_per_beat. That, combined with a TEMPO control should
allow the plugins that care to access musical time without host callbacks
and without impacting the fast path.

* Pre-queuing becomes:
  - Plugin: I have an event 100 ticks from now
  - Host sends ticks as they happen, even if we loop in time
  - Plugin decrements the counter on future events
  - If time jumps plugin can discard future events or not, as it wishes
  - when the counter hits 0, plugin does whatever

* We can provide SDK code to make this pre-queuing easy.

Just late night blathering. Tell me if I am an idiot (and preferably what
about :)...

I'm sure I'll have a slew of emails when I wake up, I always do.

Tim


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Fri Dec 13 2002 - 11:13:01 EET