[mythtv-users] Hardware for Best TV-Out?

[Cory Papenfuss]

> | 	I thought you said that the Harmonic Research does exactly
> |that.
>
> Sorry, I was unclear.  I meant a complete solution to provide NTSC out for
> Myth.  Without seeing for myself, I don't yet accept your premise that
> running the RGB output of an nVidia card through (even a good) NTSC encoder
> will give me what I want. :)  I'd have to think about whether it would be
> worth spending the money just to try.
>
 	"Complete solution."  Boy, that's marketspeak if I ever heard it! 
:)  Seriously, though, the biggest problem I've had with mine is 
convincing a video card modern enough to do Xv to go low enough and 
produced bugless interlaced video.  I can assure you that the quality of 
what I've cooked up is much better than any TVOUT card I've ever seen.  I 
was fairly careful on the PCB layout and termination of the video lines, 
etc.  Also, the AD724 is a SMC chip, so it keeps the inductance low, too. 
I'm sure that a crappy layout will produce crappy results (part of why I 
haven't built a RGB->Component transcoder yet... too lazy to lay out a 
PCB)


> This is interesting.  I have a CV233.  The RGB version is the CV223.  The
> CV121/CV131 series used to be the ones that did not provide a locked sub-
> carrier and cost less than half as much.  Looking at the description of
> the CV121A/CV131A they now list locked sub-carrier as a feature!  I wonder
> if they really changed the design or this is more market speak.  If they
> really changed the design and kept the cost low, I can't see why anyone
> would bother with the CV2* series anymore (surely not just for the BNC
> connectors).

 	Here's something you might try to see if it's worth investing more 
into the experiment.  I did this the other day to simulate B&W NTSC from 
the VGA card.  If you've got a VGA->BNC or equivalent breakout, try using 
a vid card with sync-on-green into the Y-channel of your component->NTSC 
box.  You'll only get B&W (of green), but it's good enough to get 
luminance resolution, look for ringing, and see if the RF-modulator like 
the line-locked signal it produces.  Actually, aside from the colors being 
good-n-groovy, go ahead and put R&B on there as well... then you *will* 
get (horribly wrong) color to test.

 	That's a pretty cheap test, even if you don't have the $20 
RGB->BNC breakout cable.  I've had *really* good luck with Matrox cards 
and funky vid options (sync-on-green, csync, interlace, low clocks, etc). 
The Millenium doesn't do Xv, but for a static test burn the CPU for 
psychodelic-colored video... :)

>
> |You could always buy/build an RGB-Component transcoder (an
> |analog generation loss, though) to run through the one you've got.
>
> Yes, I've considered that.
>
> 				Dan Lanciani
> 				ddl at danlan.*com

 	Might have renewed vigor if the RGB one works.  I was going to 
build rev2 of my circuit here when I get some time, so I'll look into 
replacing the crystal with a PLL.  If I get something that's user-friendly 
for others, I'll post a page describing the construction.

-Cory