Subject: Re: [linux-audio-dev] rough rough ladspa docs
From: Benno Senoner (sbenno_AT_gardena.net)
Date: Mon May 15 2000 - 00:32:46 EEST
On Sun, 14 May 2000, David Benson wrote:
>
> For example, I am going to propose a semantical
> change for LADSPA_PROPERTY_HARD_RT_CAPABLE
> soon. The amended comment is merely:
>
> The plugin will not block on I/O.
>
> I'd also like to make LADSPA_PROPERTY_INPLACE_BROKEN
> the default and cut the flag.
No, no no !
HART_RT_CAPABLE is useful for potential hosts which would run this
plugin in an IRQ routine context, in order to achieve best latencies.
INPLACE_BROKEN is useful too, in order to minimize cache footprint
when you can.
most plugins will not set this flag, but some like the double <---> float
converters in my multi-datatype LADSPA proposal must set this flag
because at different input and output data sizes you are overwriting
the source data thus making inplace processing impossible with this plugin.
(if you look at my applyplugin.c , it tries allocates only
(1 + number of INPLACE_BROKEN plugins) buffers which optimal
in both terms of speed and mem usage.
Conclusion:
leave the HART_RT_CAPABLE and INPLACE_BROKEN flags as is !
I don't see any pratical reason to change this behaviour ?
(or are you really too lazy to set 2 flags in the initstage of your plug ? :-) )
Benno.
>
> As you are one of the few host authors, I'm quite
> interested in your feedback.
>
> Thanks,
> Dave
This archive was generated by hypermail 2b28 : Mon May 15 2000 - 00:12:13 EEST