soundcard access

[Patricio Bruna V]

hi,
   i have the following trouble. My apps can't access the soundcard at
the same time. I guess its probably because my soundcard does not have
this hw avility. So i look and do some research on the subject and find
dmix, but i need to configure the apps to use it.

So the question is, can i enable this software based option globally for
the system?

PS: Im posting here because im using fedora-devel, and i'm think this
ist an important feature.
-- 
Patricio Bruna                                  http://www.linuxcenterla.com
Ingeniero de Proyectos                  Mariano Sánchez Fontecilla 310
Red Hat Certified Engineer              Las Condes, Santiago - CHILE
Linux Center Latinoamerica              Fono: 2745000

[Colin Walters]

On Tue, 2005-01-04 at 15:28 -0300, Patricio Bruna V wrote:
>  hi,
>     i have the following trouble. My apps can't access the soundcard at
>  the same time. I guess its probably because my soundcard does not have
>  this hw avility. So i look and do some research on the subject and find
>  dmix, but i need to configure the apps to use it.
>  
>  So the question is, can i enable this software based option globally for
>  the system? 
Basically our OS is broken, you don't want an option to unbreak it.  You
want it to work automatically.

>  PS: Im posting here because im using fedora-devel, and i'm think
> this
>  ist an important feature. 
It is a very important bug to fix, I agree.  Here's the Bugzilla:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=130593

I hope to get some time soon to work on it.

[Mike Hearn]

On Tue, 04 Jan 2005 13:21:38 -0500, Colin Walters wrote:
>  It is a very important bug to fix, I agree.  Here's the
> Bugzilla:
>  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=130593
>  
>  I hope to get some time soon to work on it. 
There are some serious bugs in the FC3 version of dmix/alsalib, a new
version would need to be pulled from upstream and tested heavily before it
can be enabled for everyone IMHO.

It also leaves the problem of what to do about apps that don't use
alsalib. There is an "aoss" wrapper but the user has to know to use it,
which isn't very user friendly. That's a variation on a theme though: we
may need to develop some equivalent to the Windows shimming mechanism
in the future, ugly as it is :/ I believe the ALSA guys are looking at how
to enable kernel OSS emulation -> userspace routing actually so there may
be hope here still.

Final note is that it'd be wonderful if distro practice didn't splinter
over this. Ubuntu seems to be leaning towards not using dmix and
instead using polypaudio for everything. I don't really care what happens
here as long as it's fairly consistent ... ie dmix + polypaudio would be
fine by me :)

thanks -mike

[Ralf Ertzinger]

Hi.

Mike Hearn <mike(a)navi.cx&gt; wrote:

>  There are some serious bugs in the FC3 version of dmix/alsalib, a
> new
>  version would need to be pulled from upstream and tested heavily before
>  it can be enabled for everyone IMHO. 
Would some of these bugs explain crackling sound when the system becomes
busy? Raw processor load does not trigger it, switching applications or
scrolling in Mozilla/Firefox does...

-- 
"The fact that the Web now allows you to realise anything your
imagination can come up with, has simply served to highlight how many
people were lacking in imagination in the first place."
		-- Nick Fitch

[Alan Cox]

On Tue, Jan 04, 2005 at 09:09:20PM +0100, Ralf Ertzinger wrote:
>  Would some of these bugs explain crackling sound when the system
> becomes
>  busy? Raw processor load does not trigger it, switching applications or
>  scrolling in Mozilla/Firefox does... 
What video ?

[Ralf Ertzinger]

Hi.

Alan Cox <alan(a)redhat.com&gt; wrote:

>  On Tue, Jan 04, 2005 at 09:09:20PM +0100, Ralf Ertzinger wrote:
>  > Would some of these bugs explain crackling sound when the system
>  > becomes busy? Raw processor load does not trigger it, switching
>  > applications or scrolling in Mozilla/Firefox does...
>  
>  What video ? 
nvidia gf2mx, agp port, on a VIA KT133A chipset, plain xorg drivers,
system is running the latest rawhide.

-- 
"Chocolatey space cadet reporting!  Space support down to zero
cocoa!  ... feeling spacey ..."	 -- Cocoapuffs commercial

[Max Kanat-Alexander]

On Tue, 2005-01-04 at 22:09 +0100, Ralf Ertzinger wrote:
>  > On Tue, Jan 04, 2005 at 09:09:20PM +0100, Ralf Ertzinger wrote:
>  > > Would some of these bugs explain crackling sound when the system
>  > > becomes busy? Raw processor load does not trigger it, switching
>  > > applications or scrolling in Mozilla/Firefox does...
>  > 
>  > What video ?
>  
>  nvidia gf2mx, agp port, on a VIA KT133A chipset,  
	I think that KT133A is your problem. I have a KT133, and they have a
famous interrupt problem. That is, at least with my M-Audio Delta 44,
unless I set the PCI Latency to 0 and turn off PCI Delayed Transaction
in the BIOS, I get crackle, crackle, crackle.

	And sometimes, even if I do all that, I still get the crackle. 

	It happened in Windows 2000, too.

	-Max

[Ralf Ertzinger]

Hi.

Max Kanat-Alexander <max_list(a)fedorafaq.org&gt; wrote:

>  	I think that KT133A is your problem. I have a KT133, and they have
>  	a famous interrupt problem. That is, at least with my M-Audio Delta 44,
>  unless I set the PCI Latency to 0 and turn off PCI Delayed Transaction
>  in the BIOS, I get crackle, crackle, crackle. 
Can these parameters be set from a running system in some way, or must this
be done on powerup/reset?

>  	And sometimes, even if I do all that, I still get the crackle. 
Yes, I had that too, with my SB16 (ISA! Yay!) under Win2k (and Linux), if
the card was using 16bit DMA. 8bit DMA fixed that.

The current card is an "Ensoniq 5880 AudioPCI" (Soundblaster 128 or
something like that), and it worked just fine with win2k (while I had
that), and with Linux, too, as long as the OSS drivers were used. ALSA
inevitably starts to crackle after a while.

-- 
The only substitute for good manners is fast reflexes.

[Ralf Ertzinger]

Hi.

Ralf Ertzinger <fedora-devel(a)camperquake.de&gt; wrote:

>  Can these parameters be set from a running system in some way, or
> must
>  this be done on powerup/reset? 
Seems like you can.
Anyway, I went into the BIOS and disabled "PCI delayed transactions". So
far the sound card seems happy. If not I found this:
http://www-106.ibm.com/developerworks/library/l-hw2.html

Max, maybe this helps you.

-- 
Confucius say, "A bird in hand makes hard to blow nose."

[Ralf Ertzinger]

Hi.

Ralf Ertzinger <fedora-devel(a)camperquake.de&gt; wrote:

>  Seems like you can.
>  Anyway, I went into the BIOS and disabled "PCI delayed transactions". So
>  far the sound card seems happy. If not I found this:
>  http://www-106.ibm.com/developerworks/library/l-hw2.html 
OK, neither of them solves the problem. Back to square 1 :)

-- 
"We turn the Cube and it twists us."  -- Erno Rubik

[Max Kanat-Alexander]

On Wed, 2005-01-05 at 12:22 +0100, Ralf Ertzinger wrote:
>  Hi.
>  
>  Max Kanat-Alexander <max_list(a)fedorafaq.org&gt; wrote:
>  
>  > 	I think that KT133A is your problem. I have a KT133, and they have
>  > 	a famous interrupt problem. That is, at least with my M-Audio Delta 44,
>  > unless I set the PCI Latency to 0 and turn off PCI Delayed Transaction
>  > in the BIOS, I get crackle, crackle, crackle.
>  
>  Can these parameters be set from a running system in some way, or must this
>  be done on powerup/reset? 
	You have to do it in the BIOS, usually.

>  Yes, I had that too, with my SB16 (ISA! Yay!) under Win2k (and
> Linux), if
>  the card was using 16bit DMA. 8bit DMA fixed that.
>  
>  The current card is an "Ensoniq 5880 AudioPCI" (Soundblaster 128 or
>  something like that), and it worked just fine with win2k (while I had
>  that), and with Linux, too, as long as the OSS drivers were used. ALSA
>  inevitably starts to crackle after a while. 
	Yep, sounds similar to my problem (except that I never could use OSS,
because it doesn't support my soundcard). 

	I find that forcing ESD to use 48,000 Hz also helps, sometimes, for
some odd reason.

	-Max