[mythtv-users] PVR500 Composite In Picture Quality

Brian Wood beww at beww.org
Tue Feb 6 04:33:02 UTC 2007


On Feb 5, 2007, at 8:55 PM, Jeffrey Born wrote:

> The video camera has a view finder that plays at the same time as  
> the output to the composite in of the PVR500.  I can watch both and  
> never see video corruption in the view finder, but will on the  
> capture.  The capture is 98% acceptable.  The 2% that it isn't, it  
> is really bad.  It is not a 5 minutes good and then some bad, it  
> seems to be completely random.
>
> I will move my future findings over the the thread listed above.  I  
> will also wait patiently for part 2, it will hopefully help me get  
> a clean capture.
>
> I tried playing the capture via mplayer and it had the same  
> artifacts in the video.  I did the capture via the myth interface.   
> Should I remove myth from the capture process.  If so what  
> application would you recommend?

You could just set the PVR up manually and then just cat the output  
to a file.

>
> My other option seems to be, use the VCR to play the tape, and  
> output via coax, since I've never seen an issue capturing standard  
> def via coax.

So your captures are OK using the PVR's tuner and you only see the  
"corruption" when capturing composite video ?

First, make sure that this is totally accurate information, because  
if it is you have virtually eliminated a lot of possibilities, and if  
it is bad info we will be chasing a large Red Herring.

The video output of camcorders is often not very good as far as video  
standards, but most of the timing problems will also show up in the  
RF output.

You could have a level problem. Can you relate the problem to the  
average brightness of the scene? (APL or "average picture level").  
Very bright objects especially tend to be too high in level and this  
sort of problem might not show up in the RF as they might have some  
sort of ALC in the modulator.

If you had a waveform monitor, or even a general purpose  
oscilloscope, you could measure the level, which should nominally be  
1 volt peak-to-peak. But you probably don't have either of those tools.

If you record the tape with the PVR twice, are the problems in the  
precise same places or are they in different places each time?

I'm assuming this is a tape you recorded with the camcorder. If you  
are playing commercial tapes they might have Macrovision or other  
encoding designed specifically to make recording difficult.

Can you try recording a known good (or at least thought to be better)  
composite video signal with your PVR? Say for example the composite  
output of a TV set or a cable box or satellite receiver. This might  
help pin it down to something specific to the camcorder's output  
signal. I'm assuming you don't have a color-bar generator which would  
be the best for this.

The fact that you don't see the problem in the viewfinder may be  
meaningless as that display is driven before the NTSC encoder, but  
the fact that the RF signal doesn't show the problem is noteworthy as  
it should be modulated by the same composite signal as you are  
feeding the PVR.

If you get a chance, send me a private mail with the results of any  
of these tests you can do and I'll be sure to post a summary when the  
problem gets fixed (aren't I the eternal optimist?).

Meanwhile we all of course welcome any other ideas :-)

Brian Wood
beww at beww.org







More information about the mythtv-users mailing list