[mythtv-users] Channel Reception Issues

[Brian Wood]

On Feb 22, 2007, at 11:17 AM, ryanlists79 at gmail.com wrote:

> I've been trying to figure this out for the last few weeks, and am
> pretty frustrated.  On certain channels, mainly 3 (ABC), 4 (NBC), 43
> (Cartoon Network) & 69 (Spike), I am having what appear to be  
> reception
> issues.  Basically I have small "bars" rolling across the screen.  On
> channel 69, they are color, red and a semi-green, and sometimes  
> roll up
> and down, and sometimes they don't roll at all.  On channel 3 they are
> just darker & lighter bars, almost a change in brightness, and scroll
> from the bottom left corner to the upper right corner.  On channel 4,
> they are similar to channel 3, but smaller, a bit more subtle, and
> scroll directly up.  And channel 43 is an even more subtle version of
> channel 4.
>
> You can see screen shots from channels 3 & 69 on my flickr site:
> http://www.flickr.com/photos/83544026@N00/
>

What you have are called "beats" in the vernacular. They are caused  
by a carrier mixing or "beating" with the visual carrier causing  
hetrodyne products that fall within the baseband video frequency range.

The fact that they are more or less steady indicates that the  
interfering carrier is more or less CW (ie" not modulated), and more  
or less "on" a multiple/submultiple of the horizontal scan frequency  
or, possibly, the vertical scan rate. The source is almost certainly  
from your computer somehow. If the interfering beats are moving  
slowly it indicates that the beat product is merely "close" to a  
multiple/submultiple of the scan frequency, and as the beat gets  
farther and farther away from such multiple (moving faster and  
faster) it eventually gets unrecognizable as a "beat" and rather just  
results in a general fuzziness or lack of detail in the baseband video.

If they are absolutely stationary then they are right on some  
multiple or sub-multiple of the horizontal scan frequency. You could  
theoretically count the number of bars and thus calculate the precise  
frequency of the interfering beat (actually there would be 2 possible  
solutions, depending on whether the interfering carrier was above or  
below the scan frequency). There's no real point in doing this  
because it would not help you in eliminating the problem.

The beats *could* be third-order products happening in the RF realm,  
and being enhanced by cheap tuner on the PVR card. You could try  
attenuating the signal going into the card. The bottom line here is  
that expecting broadcast-quality performance from a $100 card is  
pretty much the height of optimism.

But what you're interested in is what you can do about it. From the  
behavior you describe it sounds like it's happening in the recording  
itself, you could prove that by playing the mpeg back on another  
machine.

You could try moving cards and/or cables around inside your case.  
Video cards are a notorious source of crap so you could try having  
the backend make a recording with no video card in the machine (set  
it up then remove the card and re-boot and let it run headless). You  
could try some sort of shielding, check that all mobo screws are  
tight, make sure everything's grounded well.

Even power supplies can generate crap that's within the video  
bandwidth ( essentially 0 - 5 Mhz). Cheap switcher supplies  
especially are prone to this.

But in the end there may be nothing you can do short of a better- 
designed PC case and/or motherboard combination. Even well-designed  
professional broadcast gear is prone to this sort of thing if the  
designer is not extremely careful.

But at least you now know *why* you are having problems :-)

All this should be obvious to any decent engineer after a 1 second  
glance at your pictures, but I guess I have to keep reminding myself  
that few here on this list worked in TV broadcasting for 35 years :-)  
Obviously what I've said here is a gross oversimplification but I  
hope it makes the point clear.