[Prev] Thread [Next]  |  [Prev] Date [Next]

Re: [ivtv-devel] bad audio on pvr500 input until set audio input Andy Walls Fri Aug 28 13:02:24 2009

On Fri, 2009-08-28 at 12:01 -0500, Chris Kennedy wrote:
> I had an odd experience with ivtv driver setup in
> kernel  This is not using myth but just reading
> from the capture cards and manual setup using v4l2-ctl
> and ivtv-tune.  The audio on one input of a pvr-500
> sounded really horrible, like far away in the background
> and ringing.  I saw on the v4l2-ctl --all output that
> it said 'Available subchannels: mono lang2', different
> than all the other inputs saying just mono.  Upon capture
> restart and normal setting up of the card (changing input
> to input 0 or Tuner 0) it didn't fix it.  It did fix it
> when I ran the command '/opt/ivtv/v4l2-ctl -d /dev/video2
> --set-audio-input 0'.   
> I ran kernel for months without ever seeing this problem, 
> so not sure if it's a very rare fluke or it'll happen more with this
> newer kernel (which I just started running a few days ago).
> I guess this is the 'tinny audio bug', although I've never
> seen this until now, and didn't ever see anyone notice the
> mono/lang2 oddity either.

It sure sounds like it.

My guesses on the tinny audio failure mechanism are 

a) the audio microcontroller in the CX25843 is misdetecting the standard
(or some parameters about it) and setting registers wrong.  Maybe this
is caused by SIF audio settings or audio AGC setting in the frontend not
being quite right.  Though last I checked, I thought the cx25840 driver
was setting things up properly.

b) Firmware is not getting loaded to the audio microcontroller properly
- one or more bytes may be in error.  Maybe this could be explain kernel
dependence of symptoms.

c) For 32 ksps tuner audio, the CX25843 AUX_PLL is being run out of spec
at about 196 MHz; 4 MHz below the 200 MHz specified as the lower end of
operation in the CX2584[0123] datasheet.  Avoid using 32 ksps audio.

With the death of OTA analog in the US, I have no opportunities to
reproduce this issue.

If you can reproduce again, maybe use v4l2-dbg to dump the registers of
the cx25840 and compare the good and the bad cases.  I suspect registers
in the 0x800-0x8ff range will jump out as being very different.  It
won't explain the failure mode however...

> It happened twice now, and the second time after setting the
> audio input to 0 on each capture start, so that doesn't fix
> it for good.  The difference between a good input and bad input
> is the following from the v4l2-ctl --all cmd...
> Good -       Available subchannels: mono
> Bad  +       Available subchannels: mono lang2

I wonder why stereo doesn't show up...


> Thanks,
> Chris

ivtv-devel mailing list