Re: [linux-audio-dev] MuCoS, Glame, API Issues

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

Subject: Re: [linux-audio-dev] MuCoS, Glame, API Issues
From: Paul Barton-Davis (pbd_AT_Op.Net)
Date: pe maalis 10 2000 - 08:15:31 EST


>> "host controlled allocation" doesn't specify any particular way of
>> allocating buffers. it describes the division of responsibility. if
>> plugins allocate their own buffers, their actions may not be optimal
>> given the current global situation. if the host does it, it can
>> (presumably) make the best decisions about how to allocate buffers and
>> pass them around. whether it allocates one per plugin, makes them all
>> share, or whatever else, is invisible to the plugins and should be
>> assumed to be "taken care of" by plugin writers.
>
>Umm, well who does know more about which kind of buffer it will need and
>how to allocate/forward it most efficiently but the plugin itself???

Please notice that I am talking *only* about buffers used to pass
audio data from plugin to plugin. If the plugin needs its own
specialized buffers (say, for an internal delay line), it can allocate
those during its instantiate routine(s) and do so in any way that it
wants.

For the "passing audio around" buffers, the plugin has *no* idea how
to forward it. It has no idea what it is connected to, or what kind of
flowgraph it is a part of.

The plugin may have some idea on how to allocate it. It might know,
for example, that it would be an exceedingly good idea to mlock the
buffer, or to ensure some wierd alignment, or whatever else. But too
bad. The requirements of the system overall mean that we optimize for
the flowgraph, not for the plugin. Since the host is the only place
that knows the flowgraph, the host is in charge.

--p


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

This archive was generated by hypermail 2b28 : su maalis 12 2000 - 09:14:06 EST