Re: [sc-dev] new zombies on the block

[Julian Rohrhuber <rohrhuber@xxxxxxxxxxxxxx>]

> On Nov 30, 2003, at 11:48 AM, Julian Rohrhuber wrote:
>
> >  when trying to release a synth within less than
> >  blocksize/samplerate they do not get released reliably.
>
> Anytime a gate off message is scheduled in the same control period 
> as the s_new, the EnvGen will never see the trigger.
> possible solutions:
>
> initialize gate to -1 < gate < 0. Then if EnvGen sees 0 on starting 
> up it can assume that the EnvGen has been released by explicitly 
> setting it to zero. gate must be > -1 because that is already used 
> for the EnvGen voice stealing mode.
> 	pro: seems like the most minimal change.
> 	cons: This would break existing patches that wait for a 
> trigger, because all patches that start with 0 waiting for a trigger 
> will go immediately into a done state. These patches would have to 
> be changed to initialize gate < 0 and recompiled. Notationally 
> inconvenient: would require explicitly initializing the gate 
> argument of a SynthDef.
>
> have a separate input for start trigger and release trigger.
> 	cons: breaks all existing patches that use EnvGen. not as 
> simple as a single gate. seems inelegant.


first solution seems the better. If we decide not to change it and make it a
user responsibility, there should be a check in NotePlayer for sustain time.
But again this might be problematic when there is tempo changes.
--








.