[mythtv] Ticket #7736: DVB captions for Danish broadcasting channels (H.264 over DVB-T)

[Morten Grouleff]

MythTV wrote:
> #7736: DVB captions for Danish broadcasting channels (H.264 over DVB-T)
> -----------------------------------------------+----------------------------
>  Reporter:  Morten Grouleff <mg@…>             |       Owner:  janne     
>      Type:  patch                              |      Status:  new       
>  Priority:  minor                              |   Milestone:  unknown   
> Component:  MythTV - DVB                       |     Version:  0.22-fixes
>  Severity:  medium                             |     Mlocked:  0         
> -----------------------------------------------+----------------------------
>
> Comment(by stuarta):
>
>  The patch is not correct because it is making the assumption
>  that anything in a DVB private stream is a subtitle. This is
>  not correct in wild.
>   
The patch does not assume that every private stream is a subtitle. Only 
those tagged with "DVB_SUBT_DESCID" (0x59) has their type set.

Could you take another look at the function "mpegts_parse_desc"? That 
function already did inspect the "desc_tag" and switch on the the value 
to decide the kind of contents within the private stream. For streams 
tagged as audio, the kind detection of this patch is proposing, is 
already used.  The patch does not change this.

What the patch removes is the requirement on (dvbci->comp_page | 
dvbci->anc_page) to be non-zero. The streams in question have a zero 
value here, but are never the less subtitle streams tagged as such.

Are you saying that in order to be a proper subtitle stream, it has to 
have both been tagged with "DVB_SUBT_DESCID" and have a 
comp_page/anc_page value that is non-zero? I have tried to find that 
requirement in dvb-specifications with no luck, as I would like to be 
able to tell the broadcaster to change the way they encode the stream, 
but with no luck.

Is it possible to have streams with audio tagged as "DVB_SUBT_DESCID"?

(I believe we're not suppose to use trac for discussion when possible, 
thus I have replied here on the mailing-list and not in trac.)

Thanks for taking the time to comment.

Regards,
Morten Grouleff