[linux-dvb] DVB file formats

[Andrew de Quincey]

On Friday 17 June 2005 08:34, Marcus Metzler wrote:
> >>>>> "Andreas" == Andreas Oberritter <obi at linuxtv.org> writes:
>
>     Andreas> Hi, On Fri, 2005-06-17 at 00:10 +0100, Andrew de Quincey
>
>     Andreas> wrote:
>     >> The first thing that springs to my mind is something like the
>     >> following:
>     >>
>     >> DVBS: ((frequency / (symbolrate/1000)) << 1) | polarisation
>     >> DVBT: (frequency / bandwidth) DVBC: (frequency / symbolrate)
>
>     Andreas> AFAIK there can not be more than one multiplex one
>     Andreas> frequency except if the polarization differs. That makes
>     Andreas> me think that the divisions are unnecessary. Instead I'd
>     Andreas> like to see the orbital position in there for DVB-S
>     Andreas> somehow like that:
>
>     Andreas> (orbital_position << 22) | (polarization << 20) |
>     Andreas> frequency_mhz
>
> I looked for some examples in my current dvbrc which is for 5 orbital
> positions. With 207 transponders I found 32 cases of the same TSID
> involving 117 of the transponders, i.e. between 3 aand 4 transponders
> per TSID. If you exclude the cases where you have different orbital
> positions to distinguish between the transponders you still have 28
> cases with 108 transponders. If you use the polarisation to
> distinguish the transponders you get 29 cases (15 for V with 49
> transponders and 14 for H with 41 transponders)
> and still 100 transponders.
> If you take the first 4 digits of the frequency, all these remaining
> cases can be resolved.
> So a combination of transport ID, orbital position and frequency
> should be enough. Polarisation does not really help much.

Cool, thanks for confiriming that.