[SunRay-Users] Re: Two-channel audio recording problems

Peter Åstrand astrand at cendio.se
Wed Apr 23 22:18:39 EEST 2008


We have now confirmed this problem on another x86 system. Our plan is to 
switch to open the device in one channel mode and perform the 
"duplication" ourselves in our application. How can we check if the audio 
device is indeed a Sun Ray utaudio device, the best way? The utaudio man 
page says:

     ...
     use the AUDIO_GETDEV ioctl to determine which  audio  device
     is  being  used.   The  utaudio  driver  returns  the string
     "SUNW,CS4231" in the name field of the  audio_device  struc-
     ture  to  indicate compatibility with other Ultra platforms.
     The version field contains "a" and the config field contains
     "pseudo."

However, on our system, the config field contains "onboard1" rather than 
"pseudo". Any ideas?

Best regards, 
Peter Åstrand

On Thu, 10 Apr 2008, Peter Åstrand wrote:

> We are facing problems with two-channel audio recording:
> 
> audiorecord -c 1 foo.wav
> 
> ...works great, but:
> 
> audiorecord -c 2 foo.wav
> 
> ...gives basically just noise. (The noise is reduced when you talk in the 
> microphone, though.) This problem has been experienced on a Solaris 10 x86 
> machine, with all updates installed, using two different DTUs (generation 
> 1 and 2). On another system, based on the SPARC platform, we don't see 
> this problem. I guess the problem might be that the DTUs does not support 
> two channels, but: According to the utaudio man page, two channels should 
> be simulated:
> 
> "Some input devices do not support 2-channel capture, but two channels 
> will be reproduced by duplication in the case where two channels are 
> requested and the device supports only one."
> 
> At this point, I think it looks like utaudio does not handle two channel 
> simulation on the x86 platform properly. 
> 
> I've found this posting as well: 
> http://article.gmane.org/gmane.os.solaris.sunray/2288/. It doesn't mention 
> if the Solaris 10 host is SPARC or x86, but the problem description seems 
> to match we are are seeing. 
> 
> Can perhaps someone out there try "audiorecord -c 2 foo.wav" on x86 and 
> report the result?

---
Peter Åstrand		ThinLinc Chief Developer
Cendio AB		http://www.cendio.se
Wallenbergs gata 4
583 30 Linköping	Phone: +46-13-21 46 00


More information about the SunRay-Users mailing list