[mythtv-users] DVB-T Hell in Sydney

[Doug Scoular (dscoular)]

Hi Igor et al,

Igor wrote:
"Not sure if this helps but I'm in Sydney and I did update mythtv (on
mythbuntu 14.04 though) from 0.27 to 0.28 about a month ago and I can
confirm that all channels work properly for me as they did on 0.27.
Tuning&recording all good here."

Yes, that was almost my situation but with 16.04 all worked properly until
May 30th/June 1st. There was a software update around that time but I'm
not ready to blame it for my woes.

The big news is that 0.27 still works perfectly... so there's nothing
wrong with my hardware, antenna or whichever transmitter my antenna is
pointing to.

Anyway, I rebooted into 0.27 and checked that I could tune in to every
network and every network's channels. All worked perfectly so I exported
the 0.27 tables which I think relate to DVB-T as:

dscoular-M-51PJ:~ dscoular$ ls -1 mythtv-0.27/
capturecard.csv
channel.csv
channelscan.csv
channelscan_channel.csv
channelscan_dtv_multiplex.csv
dtv_multiplex.csv
dtv_privatetypes.csv
inputgroup.csv


I rebooted back into 0.28 and I created a new export of some of the 0.28
tables and started making comparisons (again no channels can get lock
under 0.28 and tzap can only get a subset of channels to lock).

The only tables that I think are really significant are: capturecard,
channel and dtv_multiplex.

To my dismay the dtv_multiplex tables are identical! I had expected these
to be the root of my problems, but no!

The capturecard table has evolved to have a new parentid column and a few
other extras. 

I created a little sqlite3 database so that I could drill down on some of
the differences of the capturecard table's specific rows:

sqlite> .mode csv capturecard_27
sqlite> .import mythtv-0.27/capturecard.csv capturecard_27
sqlite> .mode csv capturecard_28
sqlite> .import mythtv-0.28/capturecard.csv capturecard_28

sqlite> .mode line


	The above now lets me list a single row in "line" mode:

sqlite> select * from capturecard_27 where cardid = 1;
               cardid = 1
          videodevice = /dev/dvb/adapter_playtv-0/frontend0
          audiodevice = NULL
            vbidevice = NULL
             cardtype = DVB
         defaultinput = Television
       audioratelimit = NULL
             hostname = spug
         dvb_swfilter = 0
         dvb_sat_type = 0
dvb_wait_for_seqstart = 1
          skipbtaudio = 0
        dvb_on_demand = 0
      dvb_diseqc_type = NULL
       firewire_speed = 0
       firewire_model = NULL
  firewire_connection = 0
       signal_timeout = 1000
      channel_timeout = 3000
     dvb_tuning_delay = 0
             contrast = 0
           brightness = 0
               colour = 0
                  hue = 0
             diseqcid = 0
          dvb_eitscan = 0

	I sorted the fields and then did a unified diff on cardid 1 between 0.27
and 0.28:

-sqlite> select * from capturecard_27 where cardid = 1;
+sqlite> select * from capturecard_28 where cardid = 1;
 audiodevice = NULL
 audioratelimit = NULL
 brightness = 0
 cardid = 1
 cardtype = DVB
+changer_device = NULL
+changer_model = NULL
colour = 0
 contrast = 0
 defaultinput = Television
 diseqcid = 0
+dishnet_eit = 0
+displayname = PlayTV-0
 dvb_diseqc_type = NULL
 dvb_eitscan = 0
 dvb_on_demand = 0
@@ -16,12 +20,22 @@
 dvb_swfilter = 0
 dvb_tuning_delay = 0
 dvb_wait_for_seqstart = 1
+externalcommand = NULL
 firewire_connection = 0
 firewire_model = NULL
 firewire_speed = 0
 hostname = spug
 hue = 0
+inputname = DVBInput
+livetvorder = 1
+parentid = 0
+quicktune = 0
+recpriority = 0
+schedorder = 1
 signal_timeout = 1000
 skipbtaudio = 0
+sourceid = 1
+startchan = 24
+tunechan = NULL
 vbidevice = NULL
 videodevice = /dev/dvb/adapter_playtv-0/frontend0


	From what I can see there are no significant differences in the
capturecard tables other than the new fields in the 0.28 schema (prefixed
with a plus).


	If I do the same for the channels table, drilling down on channel 1003
(SBS ONE):

bash-3.2$ diff -u mythtv-0.27/channel_1003 mythtv-0.28/channel_1003
--- mythtv-0.27/channel_1003    2016-06-03 17:55:22.000000000 +1000
+++ mythtv-0.28/channel_1003    2016-06-03 17:55:21.000000000 +1000
@@ -1,4 +1,4 @@
-sqlite> select * from channel_27 where chanid = 1003;
+sqlite> select * from channel_28 where chanid = 1003;
 atsc_major_chan = 0
 atsc_minor_chan = 0
 brightness = 32768
@@ -14,7 +14,7 @@
 hue = 32768
 icon = sbs861c4b17d8fc1856b814.png
 iptvid = NULL
-last_record = 2016-04-21 10:30:00
+last_record = 2016-06-01 17:50:00
 mplexid = 4
 name = SBS ONE
 outputfilters =


	Only the last recording date differs, as I would expect. The date was
approximately when I upgraded to 0.28 (21st April) and when 0.28 started
refusing to tune (1st June).

	Anyway, I'm now going to backup my mythtv 0.28, delete all cards and
channels and see how far a fresh scan gets me. I'm dubious because doing a
DVB scan outwith myth either on my old 14.10 or my 16.04 never gets
Southern Cross 10 or Seven... even though tzap will happily tune to them
on 14.10 (actually it might be 14.04... I must check).

	So frustrating to have 0.28 work so well from April 23rd till May 30. At
least there's a rainy weekend forecast for me to stay home and sort this
out.

	Wish me luck,

	Doug