In message <[EMAIL PROTECTED]>you write:
>Paul Davis wrote:
>>
>> >> This creates a very simple audio application that accepts a mono
>> >> input audio data stream and copies it to a single mono output audio
>> >> data stream. This is actually an inefficient way to do this, but it
>> >> will illustrate the basic idea of the API.
>> >
>> >Hardly...
>> >
>> >Please try again.
>>
>> the application calls audioengine_open(), registers some ports, and
>> provides callbacks to handle data processing. what's missing?
>
>??? I guess that we're talking about design of different things, I think
>it's better I close my mouth and stay behind the window.
i don't want you to be behind the window.
what I'm talking about is an API that (correctly) removes interaction
with h/w from the work done by 95%+ of all audio applications. It
provides an abstraction for dealing with audio that massively
simplifies the work of the application *and* allows for
inter-application data sharing and cooperation. this is provided for
in a way that supports real-time low latency systems as well as more
traditional applications. fundamentally, its based on what some people
call a "pull" model but i prefer to call a "callback" mode.
what i presented was the API used by clients. thats all.
so, i ask again, what's missing?
--p
