Subject: Re: [linux-audio-dev] Davis votes for MP, news at 11
From: Paul Davis (pbd_AT_op.net)
Date: Thu May 17 2001 - 18:00:43 EEST
>Much as I like the MP model, it would be a shame to loose the rock solid
>ardour performance. You have ardour running inside AES now, so can't we
>specify the same conditions in a LAAGA SP mode? It would require all SP
>apps to use gtk-- and so on I presume? Doesn't seem that large a price to
>pay.
Actually, its slightly weaker than that: GTK. AES doesn't care which
language binding you use, as long as the underlying toolkit is GTK.
I don't view this as acceptable. I want Muse to work in this
framework, and I don't think that its author will be happy about
having to rewrite it in GTK!
I don't think that Ardour's performance will suffer in any noticeable
way, other than to have slightly fewer cycles available for plugins.
>Known nominal 0db value (e.g. 1.0f)
my (admittedly weak) understanding of "dB" suggests that this makes no
sense. you could say that 1.0f is N dBa (or is that dBFS or dBv?), but
there's no way to know what N is, since it will depend on the analog
chain beyond the computer. dB is a ratio, and so I'm fine with the
idea of 1.0f as a gain value meaning 0db (unity) gain, but thats about
all i can imagine as a link between the two.
>Minimum sample rate (e.g. 44.1k's)
Seems like quite a restriction. Why do you want this? I would prefer
to have plugins just discover the engine sample rate, and decide that
they can't be instantiated, or something like that, if they're really
that fussy about SR.
>Guaranteed 2^n block size
we can't guarantee that. if there ends up being a reason for the
engine to have "events", it may have to subdivide a 2^n block.
>A way for apps to indicate thier current latency would be good too.
This concept doesn't really exist. In Ardour, for example, is the
latency the figure for a particular signal routing? That could be very
different depending on what LADSPA plugins and physical i/o
send/return combinations are in effect.
--p
This archive was generated by hypermail 2b28 : Thu May 24 2001 - 01:16:11 EEST