Hi,

On Tue, 23 Oct 2001, Alexander R Angas wrote:

> The event sound server prototype sounds OK now. It's using the DirectMusic sound

Cool! Nice work :-)

> functions which seem to work well except for one thing. It measures time in
> units of 100-nanoseconds, so I worked out 60 ticks/second to be 166667 100-ns.

That sounds about right.

> Thing is, that sounds pretty slow. I'm not sure if that's my machine or what.
> Sounds OK at intervals of 100000 100-ns though.

This is strange. Are you certain that it's 0.1us intervals you specify
there?

> Also, CPU usage is high which is
> something I'd like to investigate further.
> 
> So I'll put both a fast and slow compiled copy of the prototype at
> http://www.users.on.net/wgd/freesci/. I'd like to hear feedback from anyone who
> can try it on Win32, but remember I've only ever tested it under NT on the intro
> tune in LSL3! It's not complete.

I guess feedback is needed for those DirectMusic timer functions...

> I'd also like to hear how Christoph would like to incorporate it. Most of the
> functions should be able to be reused between the polled and event sound servers.

Well, pretty much the way outlined in the previous mails: by abstracting
functionality in implementation independant functions:

- handle_command()
- play_event()

while using parameters and return values as appropriate (play_event would
have to provide the next time it would like to be called as a return
value, for example, but it should also check whether the call happened
at the correct time- this would also help with fine-tuning your
DirectMusic implementation).

llap,
 Christoph


Reply via email to