[mythtv] Corrupt analog recordings

Tyson Harding tharding at cs.utah.edu
Mon Aug 14 15:27:31 UTC 2006


Torbjörn Jansson wrote:
> Hmm...
> Intresting, i'm prety sure i saw 100% cpu utilization too when one of the
> bad recordings was being recorded.
> 
> My normal cpu utilization for analog recordins is also around 80-85% if i
> remeber correctly.
> Do you know what change it was that started this?
> 
> And do you also have the message "v4l2: unable to handle these sampling
> parameters" in the backend log?
> 

I don't know, I can check.

> 
> The problem for me is that i don't have a reliable way of reproducing this
> so testing out different encoding settings is going to be problematic.
> 
> But i'll check the settings and see if i have some of those options on.
> I haven't changed messed with the encoding options much, must be a year (or
> two) or so maybe since i changed anything there.
> 

You don't have to get a corrupted recording to have problems. If the CPU 
can't keep up, the encoder may just drop a frame every so often to keep 
up. The thing that tipped me off to this was all of a sudden my analog 
recordings went from about 1 Gig an hour to about 600 Meg an hour. Then 
I also got a few corrupted recordings.

When the backend is doing software encoding if the CPU is at 100% you 
are most likely at least dropping frames, and possibly getting blocky 
corrupt recordings.

In the MythTV HowTo sections 23.3 it explains setting up software 
encoding, and the implications of too high CPU use.

I am going to attempt to pinpoint the revision that I started seen the 
problems on and open a ticket once I have an idea what is causing it. If 
you know the previous revision you were using, it might help me to home 
in on the revision.

Tyson


More information about the mythtv-dev mailing list