[gmpi] Re: New Reqs 3.8 - Events

  • From: UV <u.wildner@xxxxxxxxxxxx>
  • To: gmpi@xxxxxxxxxxxxx
  • Date: Wed, 17 Dec 2003 22:39:51 +0100

Jeff McClintock wrote:

In a system like VST or LADSPA the parameter value semantics (at least by
convention) are that at the sitart of the block the parameter has the old
value and at the end of the block it has the new value.


here @ steinberg we already discussed this as an issue to be improved in the next version of VST....

do we need timestamped parameters?



 Sounds very messy, as block size can vary from host to host, or even when
you change your soundcard latency setting.
 I don't want my parameter automation to subtly change because I switch
sample-rate or latency.
Quantizing parameter changes to the block boundary would be a step
backward.

Best Regards,
Jeff


I really think there is no reason to talk about non timestamped anything in a real time system.
Best effort delivery semantics is just not what you want to make music.....
Moving your project to another system would make it sound different??
I just cannot see any reason to use anything non timestamped for a real time API.
Just to save a few bytes of data you wouldnt want to sacrifice timeliness of the whole system?
what for....
To make writing plugins easy?
Then use VST thats an easy interface but it slowly gets outdated.....

So please lets focus on what kind of time stamps we need and not on if we need them.
Otherwise there might be no improvement to the existing plugin standards :-(

Cheers
UV

Other related posts:

  1. Home
  2. gmpi
  3. Archive
  4. 12-2003