Subject: Re: [linux-audio-dev] LADSPA and run_adding()
From: David Olofson (david_AT_gardena.net)
Date: Sun Nov 19 2000 - 20:21:07 EET
On Sun, 19 Nov 2000, Kai Vehmanen wrote:
> On Sun, 19 Nov 2000, David Olofson wrote:
>
> > Yep. And we actually have three situations to take into consideration
> > 1) plugin taking data from one buffer, mixing into another
> > 2) plugin taking data from one buffer, overwriting another
> > 3) plugin taking data from a buffer, while overwriting it
>
> This is a good analysis of the current situation. Ecasound only
> uses (3) at the moment, while Paul would prefer (1) in Ardour.
>
> Only (2) is required by the current LADSPA API. If (3) is not supported,
> plugin must define INPLACE_BROKEN. We could use similar policy to provide
> (1). If plugins don't provide (1), they'd have to define RUNADDING_BROKEN
> or something. How does this sound?
Well, 2) and 3) are implemented by the same code (run()) in the
normal case, and NPLACE_BROKEN means only that 3) doesn't work even
though the run() function being present suggests it would work.
As to RUNADDING_BROKEN, it actually already exists, in the form of a
pointer being NULL, or pointing at some code. Likewise for
run_adding_gain() - if it aint' there, you see a NULL pointer.
Now for some marketting style jibberish:
"Implement both run() and run_adding(),
and optimize THREE cases using TWO functions!" :-)
At least, the extra case we found makes it sound a little better,
doesn't it? ;-)
//David
.- M u C o S -------------------------. .- David Olofson --------.
| A Free/Open Source | | Audio Hacker |
| Plugin and Integration Standard | | Linux Advocate |
| for | | Open Source Advocate |
| Professional and Consumer | | Singer |
| Multimedia | | Songwriter |
`-----> http://www.linuxdj.com/mucos -' `---> david_AT_linuxdj.com -'
This archive was generated by hypermail 2b28 : Fri Nov 24 2000 - 22:23:28 EET