Date: Mon, 09 Dec 1996 19:11:51 -0800 From: Amancio Hasty <hasty@rah.star-gate.com> To: dwhite@resnet.uoregon.edu Cc: multimedia@freebsd.org Subject: Re: GUS & Mbone: more updates Message-ID: <199612100311.TAA00526@rah.star-gate.com> In-Reply-To: Your message of "Mon, 09 Dec 1996 16:41:37 PST." <Pine.BSI.3.94.961209162258.1592A-100000@gdi.uoregon.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
>From The Desk Of Doug White :
> Hello again...
>
> After I replaced my VAT with a full duplex version from rah, the looping
> and popping and whatnot has stopped. I've been watching the IETF stuff
> for several hours now and it works like a charm. The only annoying thing
> is that the audio cuts out every so often, but I think this is from
> problems at the gateway or transmitter at ietf2.bbnplanet.com so I'm not
> worried about it.
Play a little with vat if you <right> click mouse on the transmitting
source <the text region> you will be able to start an rtp stats window
which will let you know if vat is dropping packets or not.
> I've noticed that the volume bars lag quite a bit behind the audio, such
> that I've just turned them off. Any ideas behind that?
>
> The thing that does worry me is the excessive debugging output from the
> kernel. I get lots of these:
>
> isa_dmastart: channel 1 busy
> isa_dmastart: channel 3 busy
Just comment out in /sys/i386/isa/isa.c the diagnostic message.
> I know this is a silly stupid thing, but I can't remeber where this is at.
> Someone have the patch? (BTW: these pop up whenever I jump between vats
> (I'm watching both IETF channels currently), and just every so often. Any
> guesses why both the play and record DMAs bounce?
>
> Secondly, when I was using the half-duplex vat, I got a lot of these
> lines:
>
> Dec 8 23:55:26 gdi /kernel: write select 1
> Dec 8 23:55:27 gdi last message repeated 672 times
> Dec 8 23:55:27 gdi /kernel: isa_dmastart: channel 3 busy
> Dec 8 23:55:29 gdi /kernel: write select 0
> Dec 8 23:55:29 gdi last message repeated 4 times
> Dec 8 23:55:29 gdi /kernel: write select 1
> Dec 8 23:55:31 gdi /kernel: write select 0
> Dec 8 23:55:31 gdi last message repeated 4 times
Thats my fault just get rid of my debuggin message .
It is in audio.c:
audio.c: printf("write select %d\n", foo);
> Here is the interesting part. The audio was clear until the 'channel 3
> busy' message came up. Then I got the write select 0's interspersing
> thoughout the write select 1 messages. (I got these write select's
> regularly.) At the same time, the looping, popping, and general poor
> audio started up.
>
> I'm guessing that these status messages may be related to the poor audio
> output from the half duplex vat (which must be used on SoundBlasters).
> Note that i got the exact same poor output until I bumped my vat to the
> full duplex version. I still have the half-duplex, so I'd be glad to do
> some testing. I'll only have my net until midday tomorrow or so, so get
> back to me soon if you want to deal with this before January.
>
> I'm not experienced enough programming to track down this stuff. I'm
> hoping someone will pick this up (bill?) and work with this. I'd really
> like to see the half-duplex vat's audio quality improve, as many of us are
> stuck with them and don't want to spend the $xxx on a gus when (IMHO) the
> awe32 has nicer MIDI audio :), and paid $200-300 for the AWE originally.
>
> Finally: all is well at gdi. Now to bludgeon the qcam into shape again
> and play with increasing the fps...
>
Well , the meters they chew up cpu time so if you don't have a fast
computer I don't recommend that using them. My old system was
a P100 and vat's meters did have much of a problem by way of coming
up with the sound .
Have fun,
Amancio
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199612100311.TAA00526>