Tvheadend

mark ius wrote:

Hi

I'm having some issues with tvheadend on the Astra constellation @ 28.2E. I'm relatively new to tvheadend so forgive me if this is a dumb question...

I configure tvh as follows:

1. Create a new network using the attached dvbv5 scanfile. The scanfile was generated by parsing the data available at kingofsat. I'm considering the kingofsat data to be correct.

Why are you using this scan file? I've always found the NIT from the satellite to be more than adequate.

1. Assign the network to one of the LNBs on my TBS6985 DVB-S2 card
2. Initiate a "Force Scan" of the new network

What I'm seeing after the scan has completed is multiple instances of duplicate TSIDs show on different MUXes. Example:

• TSID 2043 is shown at both 10714.25H and 10743.75H (10743.75H is correct)
• TSID 2094 is shown at both 11538V and 11567.5v (11538V is correct)

If I rescan the MUXes that have the wrong TSID, they are changed to the correct value. EG 10714.25H changes to TSID 2041; 11567.5V changes to TSID 2096.

I suspect the problem is caused by the transmitted TS NIT on some MUXes being incorrect.

I've never seen the NIT for 28.2E to be wrong, not that it isn't possible.

I thought I'd be able to fix this by disabling Network Discovery for the network, but tvh still seems to populate the TSID information on "other" MUXes during a scan. This is evidenced when doing the first scan after setting up the network - as soon as the first MUX is scanned, multiple other muxes have their TSID changed from 65535.

I've also noticed that the frequency of some MUXes is being altered. For example, my scanfile has a MUX at 10744000 H. During the scan this is getting changed to 10743750 H which according to KingOfSat is incorrect.

Worse still, if I switch on "Idle Scanning", tvh eventually gets into a loop, repeatedly scanning two MUXes (10714.25H and 10743.75H) and changing the TSID back and forth over and over.

So my questions:

• Am I doing something wrong?
• What does "Network Discovery = Disabled" actually mean? It seems like tvh is still doing something Network Discovery-ish.

I'm running 4.1-2194~g54715ff~trusty

TBH it looks like Tvheadend is getting muddled up as you've used both a Scan file and NIT, most use one or the other, but still this is rather strange, you may want to open a bug report. Might be a good idea to include nit debugging info too (force a scan), e.g.

sudo usermod -aG video hts
sudo -u hts /usr/bin/tvheadend -c --debug nit --trace nit --logfile /tmp/nit.log

Mark Clarkstone wrote:

mark ius wrote:

1. Create a new network using the attached dvbv5 scanfile. The scanfile was generated by parsing the data available at kingofsat. I'm considering the kingofsat data to be correct.

Why are you using this scan file? I've always found the NIT from the satellite to be more than adequate.

I'm the scanfile because it's more up-to-date than the ones included with tvheadend (pre-defined muxes) or the ones included with Debian Jessie.

TBH it looks like Tvheadend is getting muddled up as you've used both a Scan file and NIT, most use one or the other, but still this is rather strange, you may want to open a bug report. Might be a good idea to include nit debugging info too (force a scan), e.g.

I thought using pre-defined MUXes was a fairly standard setup procedure?
I'll debug and trace the NIT stuff and see if I can pinpoint what's happening.

mark ius wrote:

Mark Clarkstone wrote:

mark ius wrote:

1. Create a new network using the attached dvbv5 scanfile. The scanfile was generated by parsing the data available at kingofsat. I'm considering the kingofsat data to be correct.

Why are you using this scan file? I've always found the NIT from the satellite to be more than adequate.

I'm the scanfile because it's more up-to-date than the ones included with tvheadend (pre-defined muxes) or the ones included with Debian Jessie.

Yes, they are very out of date, this is why I only add the two major home transponders for Sky and Freesat (11778V 11425(8)H 27500) and let Tvheadend use the NIT to find the rest.

TBH it looks like Tvheadend is getting muddled up as you've used both a Scan file and NIT, most use one or the other, but still this is rather strange, you may want to open a bug report. Might be a good idea to include nit debugging info too (force a scan), e.g.

I thought using pre-defined MUXes was a fairly standard setup procedure?

It is (by most I meant the people I've spoken to)!

I've found that using both can cause issues like this, I believe if there are certain variations in frequency info Tvheadend can sometimes treat it like a totally different mux.

I'll debug and trace the NIT stuff and see if I can pinpoint what's happening.

So I've done some testing and I think I can rule out tvh as the cause of the issue.

It appears that TSID information is "leaking" between the tuners on my TBS6985 quad-tuner card.

If I scan using a single tuner no duplicates are recorded.
If I use tuner A and B to scan, duplicates TSIDs start to appear.
Using A and C to scan yields no duplicates.

This is probably hardware or driver related so probably a good excuse to replace the card with a SAT>IP tuner.