[mythtv-users] Component video adapters

Cory Papenfuss papenfuss at juneau.me.vt.edu
Thu Dec 9 20:01:07 UTC 2004

>> in the A960.  Remember that MPEG *is* YUV... the color transformation to
>> RGB and back to YUV shouldn't even have to be done.
> Makes sense.  I hadn't thought of that.  So aside from multiplexing the
> H&V sync, the rest of the problem is just a matter of transforming the
> frame buffer contents.  Doh!

 	Actually, no frame buffer contents really need to be transformed. 
I posted a few weeks ago (and received no response) about maybe making a 
dummy Xv format like in 'xvinfo'
  Number of image formats: 4
       id: 0x32595559 (YUY2)
         guid: 59555932-0000-0010-8000-00aa00389b71
         bits per pixel: 16
         number of planes: 1
         type: YUV (packed)
       id: 0x59565955 (UYVY)
         guid: 55595659-0000-0010-8000-00aa00389b71
         bits per pixel: 16
         number of planes: 1
         type: YUV (packed)

 	... something that tells the software that the YUV data from the 
MPEG decode should be put into the VRAM unaltered (non-rotated into RGB). 
Do that, and presto... "accelerated" YUV color transformation and cheap 
"transcoder" (VGA to RCA plug is all) for free.  I hadn't thought too much 
about the voltage levels, but IIRC RGB is 0.7V which is the same as Y. 
The Pb and Pr are 350mVp, which should still be doable with 16-bit color. 
DVD MPEG Y is 8 bits and U/V are only 4 bits anyway so 16/24 bpp on the 
DACs is more than enough resolution.

>>  	Why not?  IIRC the monitor ID info is sent as a serial bitstream
>> on one of the pins.  Why can't the driver be "told" it's a HDTV with the
>> adapter?  Seems reasonable.
> The component video cable has no allowance for the clock & data signals
> needed for EDID which is what I was refering too.  Thus the monitor itself
> can't tell the driver anything.  But I see now that perhaps this adapter
> might possibly have an 8051 in it with an on-chip I2C transceiver.  Then
> the dip switches are just a few discrete inputs to some GPIO pins on the
> 8051 and the rest of the solution amounts to just a software issue in the
> 8051...

 	True nothing on the component cables, but that's why there are DIP 
switches.  It's certainly possible.


* Cory Papenfuss							*
* Electrical Engineering candidate Ph.D. graduate student               *
* Virginia Polytechnic Institute and State University 			*

More information about the mythtv-users mailing list