Hi all!
Anyone interested in real time signal processing and
Digital Audio Workstations?
Well, I've had it with Windoze and audio recording.
Spending uncountable hours on finding the right
hardware, and then realize that most of the so
called "professional" applications actually are of
no better quality than the system they're running
on... It's a bit more than I can take after all
these years of seeing better and better hardware
just being wasted!
I feel I'm only wasting my time playing with it,
and working solutions are both ridiculously expensive
and quite limited. Also, I'm still a hacker. I like
to be in control of my machines, just as a guitar
player who likes to customize his/her equipment. I
like to be able to hack my own plug-ins without
having to pay a fortune for a proprietary SDK and
a DSP development system!
The Audiality project is back up, and other spare
time stuff will just have to wait until I start to
get this running. I've finally realized no one will
do this if we all just sit around waiting for it...
---------------------------------------------------
Home site at
www.angelfire.com/or/audiality/audiality.html
---------------------------------------------------
So, what am I doing all night long now? (Instead of
sleeping... ;-)
Well, yesterday I played around a bit with the
sound example, and modified it to use the periodic
scheduler instead of the RTC. Got quite impressive
frequencies before X started to be a little too
slow to use, and the jitter seems to keep quite a
bit below 20 us even if I stress the system a bit.
The figures are quite unbelievable, especially
considering that this is on a full blown X
workstation, running the dev tools, Communicator
and other things that are known to mean a certain
death for a Windoze or Mac digital audio
workstation. Will publish figures on the site soon.
I'll throw in some code from a sound card driver
and set it up with a circular DMA buffer (auto
restart - no messing with ports) and see what it
can do. So far it seems like buffering only one
sample (P-II-233) wouldn't be a problem if the sound
card can handle it. I'm not sure it will though...
(Get IRQ, read sample & write sample before the DMA
reads it? Depends a lot on the sound card and DMA
timing.)
Anyway, the buffer size I'll use for Audiality will
be somewhere between 16 and 256 samples, so I'm
pretty sure RTL will handle it reliably. I just want
some unbelievable (but real!) figures to blow away
the competition when it comes to users that care
about real time response and reliability... :-)
The facts we all know already about RTLinux should be
enough to tell that this is a lot more than "a little
bit better than Windoze and Mac." Even BeOS with it's
3 ms minimum audio latency will be blown away.
Will this be enough to render dedicated audio
processing hardware obsolete? I don't know about
Soundscape, but ProTools have more than 3 ms of
latency... I'm aiming for 1 ms. (Configurable, off
course, in case you want to push some more power out
of it before it starts to warn for near overload.)
Eh, well, sorry for this long post... *hehe* Hope
this project can be of interest to some people on
this list. :-)
//David
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
