Bug #933
openSound is periodically scratchy
0%
Description
System Specs:
Motherboard: DFI LANParty-UF nf4 Ultra-D
CPU: 1.80GHz AMD Opteron 165 (Dual-core)
Memory: 4GB DDR 667MHz
Video: 256MB nVidia 8400GS (DVI and VGA ports in use)
Audio: nForce AC97 Audio Controller
NICs: Marvell Yukon 1000bT (onboard-disabled), nForce4 10/100 (onboard-disabled), DLink 538T PCI (RealTek 8139C chipset)
Ports: USB 1.1/2.0, Firewire, 9-pin Serial, PS/2 (keyboard and mouse)
Storage: Hitachi 500GB SATA, LG HL-DT-ST DVD-RAM GH24L550
This is OpenIndiana build 148. The installation CD installed the audio810 driver. The Wiki is reporting that the audiohd driver to be used.
Updated by Derek Witt about 12 years ago
Derek Witt wrote:
System Specs:
Motherboard: DFI LANParty-UF nf4 Ultra-D
CPU: 1.80GHz AMD Opteron 165 (Dual-core)
Memory: 4GB DDR 667MHz
Video: 256MB nVidia 8400GS (DVI and VGA ports in use)
Audio: nForce AC97 Audio Controller
NICs: Marvell Yukon 1000bT (onboard-disabled), nForce4 10/100 (onboard-disabled), DLink 538T PCI (RealTek 8139C chipset)
Ports: USB 1.1/2.0, Firewire, 9-pin Serial, PS/2 (keyboard and mouse)
Storage: Hitachi 500GB SATA, LG HL-DT-ST DVD-RAM GH24L550This is OpenIndiana build 148. The installation CD installed the audio810 driver. The Wiki is reporting that the audiohd driver to be used.
The wiki is apparently incorrect about the audiohd driver being compatible with the nForce4 audio codec. It causes OI to get stuck in a reboot loop unless i either disable or remove the audiohd driver.
Updated by Derek Witt about 12 years ago
As per this site, http://unixben.com/2010/11/upgrading-openindiana-to-oracle-solaris-11-express/, I tried upgrading to Oracle Solaris 11 Express to resolving my sound issues.
It does seem that the audio810 driver included with O11X does help with my scratchiness. When it does happen, the scratchiness isn't as bad and seems to recover faster.
It would appear that audio810 from 151 may need to be backported 148 or have OI updated to 151 (which is in the works as per the wiki).
Updated by Rich Lowe about 12 years ago
- Project changed from site to illumos gate
Updated by Derek Witt about 12 years ago
I may have found a workaround for this.
I downloaded OSS 4.2 via mercurial (http://opensound.hg.sourceforge.net/hgweb/opensound/), and built it. It installed without a hitch. Upon reboot, I am able to use my sound module without any scratchiness.
OSS 4.2 detected my nForce4 as ICH AC97 (ALC850).
I had to install sunstudio12u1.
I modified PATH in ~/.profile as follows in Italics:
export PATH=/usr/ccs/bin:/opt/sunstudio12.1/bin:/opt/csw/bin:/usr/gnu/bin:/usr/bin:/usr/sbin:/sbin
The key parts are /usr/ccs/bin and /opt/sunstudio12.1.
Then, I followed the said instructions at http://opensound.com/wiki/index.php/Building_OSSv4_from_source.
But, i had to slightly change the order of things. Had to run "make" and then "make build".
- make
- make build
- make package
- sudo pkgadd -d oss-solaris-v4.2-2004-i386.pkg
This package will automatically uninstall the existing audio drivers. A reboot is required. Then, upon restart and logging in, the login sound should be a dead giveaway. osstest runs without a hitch for me.
Updated by Richard PALO about 12 years ago
- Difficulty set to Medium
- Tags set to needs-triage
just to add to this report as it already exists,
as per http://openindiana.org/pipermail/oi-dev/2011-May/000395.html,
the following is reported in the openindiana project
https://www.illumos.org/issues/638
btw, don't understand garretts observation to configure gnome as it was already configured.... oss seems to help for now, but it would be nice to go back to a vanilla installation. we'll see how it fares with 151 and illumos when it comes about.
Updated by Ken Gunderson about 12 years ago
I can add that I'm seeing this behavior as well, both on dev148b and recently updated dev-151. System is Tyan K8E with onboard nvidia.
kvg@allakaket:~$ audioctl list-devices -v audio810#0 (/dev/sound/audio810:0mixer) kvg@allakaket:~$ audioctl show-device -v -d audio810#0 Device: /dev/sound/audio810:0mixer Name = audio810#0 Config = NVIDIA AC'97 (CK804) HW Info = AC'97 codec: Avance Logic ALC655
Updated by Richard PALO about 11 years ago
still here after upgrading to OpenIndiana Build oi_151a3 32-bit (illumos 022f1af9e21c)
always around ~45 seconds into a clip
~$ cat /dev/mixer0 SunOS Audio Framework Audio Devices: 0: audio810#0 NVIDIA AC'97, MCP2 (DUPLEX) Mixers: 0: audio810#0 NVIDIA AC'97, MCP2 AC'97 codec: Avance Logic ALC650
Thought I'd try what turned out to be a very bad idea, to truss rhythmbox .. totally hung my system and I had to hard reset. Also tried real player 11, seems the same.
I still think it resembles a buffering problem because a minute later it clears up ... for a while (sob)
Updated by Richard PALO about 11 years ago
what is strange, even though nothing particular happens using LD_PRELOAD=libumem.so with rhythmbox,
I did notice a difference with realplayer... instead of scratching at 43 seconds, it scratched around 1:30 and immediate afterwords cleared up. bizarre.
Updated by Richard PALO almost 11 years ago
Thought I'd finally try 64-bit on a board with the same audio electronics...
Using OpenIndiana Build oi_151a5 64-bit (illumos 13740:836bfdf31fc4)
Still no dice, roughly the same scratchiness after a few moments:
$ cat /dev/mixer SunOS Audio Framework Audio Devices: 0: usb_ac#2 USB Audio, 1.0 (INPUT) 1: audio810#1 NVIDIA AC'97, MCP51 (DUPLEX) Mixers: 0: usb_ac#2 USB Audio, 1.0 1: audio810#1 NVIDIA AC'97, MCP51 AC'97 codec: Avance Logic 0x414c4761 $ isainfo -v 64-bit amd64 applications amd_svm tscp ahf cx16 sse3 sse2 sse fxsr amd_3dnowx amd_3dnow amd_mmx mmx cmov amd_sysc cx8 tsc fpu 32-bit i386 applications amd_svm tscp ahf cx16 sse3 sse2 sse fxsr amd_3dnowx amd_3dnow amd_mmx mmx cmov amd_sysc cx8 tsc fpu