[linux-dvb] [RFC]

Manu Abraham manu at kromtek.com
Wed Feb 16 16:32:29 CET 2005


Hari_Mohan wrote:
> 
> -----Original Message-----
> From: Manu Abraham [mailto:manu at kromtek.com] 
> Sent: Wednesday, February 16, 2005 4:37 PM
> To: Hari_Mohan
> Cc: linux-dvb
> Subject: Re: [linux-dvb] [RFC]
> 
> Hari_Mohan wrote:
> 
>>-----Original Message-----
>>From: linux-dvb-bounces at linuxtv.org [mailto:linux-dvb-bounces at linuxtv.org]
>>On Behalf Of Manu Abraham
>>Sent: Wednesday, February 16, 2005 3:04 PM
>>To: Hari_Mohan
>>Cc: linux-dvb
>>Subject: Re: [linux-dvb] [RFC]
>>
>>Hari_Mohan wrote:
>>
>>
>>>-----Original Message-----
>>>From: linux-dvb-bounces at linuxtv.org [mailto:linux-dvb-bounces at linuxtv.org]
>>>On Behalf Of Manu Abraham
>>>Sent: Wednesday, February 16, 2005 2:11 PM
>>>To: Hari_Mohan
>>>Cc: linux-dvb
>>>Subject: Re: [linux-dvb] [RFC]
>>>
>>>Hari_Mohan wrote:
>>>
>>>
>>>
>>>>I don't know much about this but I have seen this in a dvb demodulator
>>>>output and ASI convertors. These lines were termed DVALID and these used
>>>
>>>to
>>>
>>>
>>>
>>>>stay high whenever there is valid data in the transport stream
>>>>bus-8bit/serial. This line stays low when there are some special
>>>
>>>characters
>>>
>>>
>>>
>>>>coming in the bus. I think this flag is related to that. Even error line
>>>>also is similar - ie whenever there were some uncorrectable errors this
> 
> is
> 
>>>>supposed to stay low - depends on your configuration
>>>>
>>>
>>>What i was wondering was that, whether this information could help in 
>>>the upper space DVB API/driver in any manner, in terms of error 
>>>correction and or other aspects.
>>>------------------------
>>>According to my knowledge mpeg decoder chips will clock in the data based
>>
>>on
>>
>>
>>>the state of these lines. I don't know whether any hardware mechanism is
>>
>>
>>So is this information more reliable than the 13818-1 Discontinuity flag..
> 
> ?
> 
>>such that this information can be used to identify cases where we have 
>>problems in the TS.. ?
>>since under the Linux DVB API, we don't have an exact solution for 
>>discontinuity, and TS errors to a 100% effectiveness, would these flags 
>>be useful for solving that problem ?
>>
>>For a particular DVB card, i can get these signals out from the frontend 
>>to the card and the driver.. So, considering that aspect if i get these 
>>flags out, i was wondering whether it would be possible to have a 
>>solution for the latter errors..
>>
>>That was the query that i had in my mind..
>>********************
>>
>>The error bit which I am speaking about is the signal which is given by
>>demodulator or a receiver device once it finds that input ts packet is
> 
> Yes, that's the one what i was referring to..
> 
> 
>>errored. But the discontinuity indicator which you refered is inside the
>>transport stream - so these information which are inside stream are
> 
> inserted
> 
>>from the stream generator (encoder/mux) and I don't think that tuner will
> 
> be
> 
>>reading this data to give the error signal/flag.
> 
> 
> What i was thinking was that if the demodulator detects a broken stream 
> and or inconsistencies, this would also result in broken TS's and CRC 
> failures.
> 
> 
>>For discontinuity in transport stream you can tweak software or add some
>>patches but for errored input signal indication I don't think anything can
> 
> 
> 
> Some patches have already gone in regarding broken streams, but that 
> still doesn't seem to solve all the issues.
> That only made me think in this aspect, as well as the availability of 
> these flags. These cards do not feature a MPEG decoder.
> 
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~
> I don't have much idea about v4l drivers but still if you can tell me the
> issues I can try some suggestions. If you are trying to record videos then

Broken TS's and CRC failures.. I see them once a while.. I still got to 
find out where that happens though.


> you can avoid the bad vid pes packets. But if you are playing a video
> realtime you may face problems.

I am not trying to record a video here.. but rather playing it..

> 
> Hari
> ~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> 
>>be done other than making satellite/cable reception proper.
> 
> 
> So do you mean to say that these falgs can be helpful only for MPEG 
> decoders and so on ?
> When we are doing Software MPEG decoding, rather than Hardware doing it, 
> would it be possible to make use of these flags ?
> 
> 
> @@@@@@@@@@@@@@@@@@@@@@
> Do you have APIs which give access to these flags. If you have it then the
> possibility can be explored. Which is the video codec s/w used in this?

It could be any codec.. Currently for the MPEG2 TS i am using mplayer.

> 
> Hari
> @@@@@@@@@@@@@@@@@@@@@@
> 
>>Hari
>>**************************
>>
>>
>>>there to store these values and use for upper layer applications. Because
>>>stv299 may have separate APIs to calculate BER and signal quality.
>>>
>>>Hari
>>>---------------------------------
>>>
>>>
>>>
>>>
>>>
>>>
>>>>Hari
>>>>
>>>>-----Original Message-----
>>>>From: linux-dvb-bounces at linuxtv.org
> 
> [mailto:linux-dvb-bounces at linuxtv.org]
> 
>>>>On Behalf Of Manu Abraham
>>>>Sent: Wednesday, February 16, 2005 12:27 AM
>>>>To: linux-dvb
>>>>Subject: [linux-dvb] [RFC]
>>>>
>>>>Hi,
>>>>
>>>>Is there any advantage in getting the following flags out from the tuner 
>>>>(STV0299 based).. such that it might be of some help.. ?
>>>>
>>>>(1) data valid
>>>>(2) error
>>>>
>>>>
>>>
>>>
> 

Manu




More information about the linux-dvb mailing list