[mythtv] QAM tuning problem with HD-3000 & kernel DVB drivers

Bruce Pennypacker bruce at pennypacker.org
Mon May 30 23:48:51 UTC 2005


Chuck,

I'm at exactly the same point you're at.  I've managed to manually pull 
together a good number of the QAM PID's for Comcast channels in Boston 
(I live just north of the city) and can manually capture the streams. 
It looks like myth relies on a transport ID instead of the individual 
PID's and I've also managed to collect that using dvbscan - at least I 
think I have.  Since the PSIP data isn't included dvbscan's output is 
pretty generic, and there's no documentation other than the source 
code...  All attempts I've made to manually insert this information into 
the channel & dtv_multiplex tables has failed.  myth just keeps 
informing me that the channels are off the air.

If you want a copy of the channels.conf I've managed to hack together 
let me know.  It's pretty ugly (don't have all the channel/station names 
set up properly) but it works.

-Bruce


Chuck Lloyd wrote:
> Bruce,
> 
> I've had a similar problem. I think the cause is that your cable provider
> is not providing the standard PSIP data. I have a similar problem
> with Comcast here in Boston.
> 
> I have been able to find a few stations, but I have not tried putting them
> into the Myth database yet.
> 
> I noted which channels myth seemed to find a frequency lock on
> and then did a little probing.
> 
> For example:
> 
> Myth found something on channel 84 so I tuned to that channel.
> azap -r C84
> 
> using '/dev/dvb/adapter0/frontend0' and '/dev/dvb/adapter0/demux0'
> tuning to 585000000 Hz
> video pid 0x0000, audio pid 0x0000
> status 1f | signal 0977 | snr fde5 | ber 00000000 | unc 00000000 | 
> FE_HAS_LOCK
> status 1f | signal 0977 | snr fdfd | ber 00000000 | unc 00000000 | 
> FE_HAS_LOCK
> status 1f | signal 0394 | snr fdfd | ber 00000000 | unc 00000000 | 
> FE_HAS_LOCK
> status 1f | signal 0394 | snr fdb3 | ber 00000000 | unc 00000000 | 
> FE_HAS_LOCK
> status 1f | signal 0977 | snr fdb5 | ber 00000000 | unc 00000000 | 
> FE_HAS_LOCK
> ......
> 
> Then I ran dvbtraffic to look at the pids:
> 
> dvbtraffic
> -PID--FREQ-----BANDWIDTH-BANDWIDTH-
> 0000     3 p/s     0 kb/s     5 kbit
> 0030    19 p/s     3 kb/s    29 kbit
> 0031    19 p/s     3 kb/s    29 kbit
> 0800  9362 p/s  1718 kb/s 14081 kbit
> 0801   261 p/s    47 kb/s   393 kbit
> 0802    88 p/s    16 kb/s   133 kbit
> 0840 11093 p/s  2036 kb/s 16685 kbit
> 0841   260 p/s    47 kb/s   391 kbit
> 1fff  4718 p/s   866 kb/s  7096 kbit
> 2000 25830 p/s  4742 kb/s 38848 kbit
> 
>  >From above it looks like there is a hi-bitrate signal at PID 800 
> (2048)and a low
> one at 801 & 802. ( 2049 & 2050 )
> 
> I leave dvbtraffic running and then I run mplayer:
> 
> mplayer -vid 2048 -aid 2050  -vc ffmpeg12mc -vo xvmc /dev/dvb/adapter0/dvr0
> 
> Viola, WGBH in HD.
> 
> Just need to figure how it all works into the database. (and maybe 
> automate it.)
> 
> 
> On 5/28/05, *Bruce Pennypacker* <bruce at pennypacker.org 
> <mailto:bruce at pennypacker.org>> wrote:
> 
>     Hi all,
> 
>     I've been unable to tune QAM channels on my HD-3000 with mythtv despite
>     being able to extract it using the azap, dtvscan, etc. tools.  In
>     searching around I've come across many similar reports of problems with
>     this combination.   I've tried both the 0.18.1 release and the CVS tree
>     from yesterday with the same results - doing a channel scan simply
>     results in all of the channels resulting in "Timeout Scanning Channel"
>     when it's able to lock onto a channel.
> 
>     I was wondering if anybody knows about this problem and/or may be
>     working on it.  It's been years since I've done any serious c/c++
>     development/debugging and I know next to nothing about DTV, QAM, etc.
>     but I was thinking of firing up gdb to see if I could figure out why
>     this may be happening.  Of course I don't want to expend the effort if
>     somebody else is already looking into it, hence my query here before I
>     take the plunge.
> 
>     -Bruce


More information about the mythtv-dev mailing list