[linux-dvb] Re: Reconstruct PAT/PMT tables from TS?

[Josef Wolf <jw@raven.inka.de>]

On Sun, Dec 05, 2004 at 02:05:15AM +0100, w-thiel@gmx.net wrote:

> Well, what's your problem exactly?

My problem seem to be that I am missing some puzzle pieces to see the
whole picture :)

> I have written a program to process D2Mac streams a while ago, and I
> have adopted it to DVB-S without much problems, thanks to EN 300 468.

Oh, I think you misunderstood me.  I don't want to challenge or criticise
it.  I want to _understand_ it.

> I think you are missing something: a table_id is a table_id a table_id.

So the table_id defines how the section looks like.  But it is not enough
to identify the PMT for a specific program.  To get the PMT for 3sat on
astra 11954MHz horizontal, I need sections with table_id==2 from PID==0xc8.
Sections with table_id==2 from different PID's won't help to find the PID's
that belong to 3sat, do they?

> > Hmm, but table_id is not enough to identify a specific section.  Different
> > PMT's (in different PID's) all have table_id==2, don't they?  So
> No. Have a look at 30west to get an idea, how PMTs can be transmitted.
> Or, even worse, have a look at 42east.

Oh, I don't have a rotating dish.  Can you please describe in more detail
what is wrong with 30w/42e?  Do they use different table_id's for the PMT's?
How do receivers find the PMT's then?

> > - to identify the sections containing PAT, the PID _or_ table_id is
> >   sufficient.
> You have to scan the PIDs given in the PMT.

Ough?  I can not assume that the PAT is in PID==0 and has the table_id==0?
How do receivers find the PAT if they can't make this assumption?

> You cannot be sure what they will look like: you have to accept the
> things written in EN 300468,

In order to accept EN 300468, I need to _understand_ it.  This is why I am
asking questions :)

> and you have to accept the private things.

You mean the NIT (which is placed in private sections) or do you mean that
providers ignore the standards?

> > Is it possible that PMT's for different programs are transported withhin a
> > single PID?  If it's possible, can they be placed withhin the same section
> Yes. See 30west.

OK.

> > So (in contrast to EN-300468, section 5.11.1), to uniquely identify a
                                          ^^^^^^
This was a typo.  Should have been 5.1.1.  Sorry!

> > section following information is needed:
> > 
> > - PID
> > - table_id (evtl table_id_extension)
> > - section_number
> > - version_number
> > - current_next_indicator
> >
> Hmh. IMHO, EN-300468 is quite clear.

Hmm, it's not clear to me.  5.1.1 says

- table_id
- table_id_extension
- section_number
- version_number
- current_next_indicator

is enough to uniquely identify a section.  OTOH, different sections
with table_id==2 can be transported on different PID's.  Therefore,
to uniquely identify a section, the PID is needed in addition to the
attributes mentioned in 5.1.1.  I just can't see how this fits together.

> Perhaps, you should tell what your intention is, and what for.

Given a TS, I want to find out which programs are available from this
TS and which PID's are needed to completely forward a specific program.

> You cannot expect all transponders to provide all streams required
> by the standards: the astras are quite ok, usually, 1west/5east
> is ok (with exceptions), 30west usually is ok, 42east usually is a
> big mess.

But how do set-top-receivers find out which programs are available if
required streams (like PAT/PMT's) are missing?

-- 
Please visit and sign and http://www.ffii.org
-- Josef Wolf -- jw@raven.inka.de --