[pvrusb2] Occasional audio issue with recordings

[Jan Ceuleers]

On 04/11/2019 12:53, Jan Ceuleers wrote:
> Dear pvrusb2 list,
>
> Every so often one of my HVR1950 tuners, which records from a
> directly-connected set-top box, gets into a state where audio is damaged
> such that it cannot be played using either MythTV nor MPlayer. Video is
> fine, but audio is damaged.

I just realised that the above could be misinterpreted, so here is more
information:

  * This does not only happen to one of my tuners. Both tuners that are
    attached to set-top boxes occasionally suffer from this issue.
    "Attached to" means that the tuners record composite video and
    analogue audio through the line inputs. The S-video connector is not
    used.
  * Until a couple of days ago I also had three identical tuners that
    are directly connected to the cable, i.e. that record from the
    analogue cable signal rather than through a set-top box. I have
    decommissioned these tuners because I now have a 4-tuner DVB-C HDHR
    which can record the same channels at better quality and using less
    power, but during the time that I had them in service I don't think
    they ever exhibited the audio defect I described.

So I now further speculate that the issue only affects audio recorded
through the line inputs, although I could be wrong about that.

> A sample can be found here:
>
> https://gofile.io/?c=DQSKjQ
>
> Judging by the loud click when starting playback, along with the
> occasional crackling that can be heard it sounds as if a large DC offset
> has been added to the audio signal such that the audio decoder's output
> is saturated. But that's just an assumption.
>
> The only way to recover is to reboot -- when the driver once again
> becomes removable I'm sure an rmmod/modprobe cycle will be enough. That
> is: it is not necessary also to power cycle the tuner itself.
When the issue occurs all further recordings made by that tuner are
similarly afflicted, until the machine is rebooted.
> I have already excluded the possibility of these symptoms being caused
> by an iffy power supply.

That is: I favour large central power supplies over multiple sealed wall
warts (which tend to have been designed to cost). Each of my equipment
closets sports a good quality largish 12VDC power supply, which powers
everything that takes 12V, and devices that need different voltages are
still powered from the 12VDC power supply but through DC/DC converters
(battery eliminators) normally found in model aircraft. Of course with
proper fuses and ferrite cores.

> Any hints would be gratefully received.
>
> Best regards, Jan
>