[vdr] Re: PATCHES: vdr-1.2.6 (getSTC and mp2)

[Oliver Endriss <o.endriss@gmx.de>]

On Friday 07 November 2003 15:56, Dr. Werner Fink wrote:
> On Fri, Nov 07, 2003 at 03:46:31PM +0100, Oliver Endriss wrote:
> > On Friday 07 November 2003 15:08, Sven Goethel wrote:
> > > 
> > > hmm .. so ain't there a way to get the seen PTS (video/audio)
> > > of the played (replay/transfer/live) stream ?
> > 
> > Imho the *PTS* (presentation timestamp) is not directly available,
> > but the *STC* (System Time Clock) can be used instead.
> > 
> > According to my tests, the STC is valid during replay/transfer/live
> > mode. In trick mode (pause, slow motion, fast forward/rewind) it is
> > meaningless anyway.
> > 
> > > afai understand STC is the seen PTS as long as you just play ..
> > 
> > Apparently the STC is set to the value of the PTS in the recorded
> > material.
> > 
> > It might be interesting for you to read section '11. Synchronization' of
> > the av7110 datasheet. There is an explanation how the av7110 performs
> > A/V sync. Maybe you can use some ideas.
> 
> The question is: how to get the offset to STC for audio as described
> in paragraph 11.6?  Especially in live mode.  AFAIS there is no method
> to get it from the firmware/driver.  Nevertheless this would help
> a lot to get AC3 streams in sync with the live picture regardless of
> the used channel/transponder.

If I understand 11.5/11.6 correctly, there are basically 3 STCs:
STCsys, STCvid and STCaud, where
o STC = PCR
o STCvid = STCsys+vidoffset
o STCaud = STCsys+audoffset
o vidoffset = audoffset (0 unless a buffer over/underrun occurred)

If the DMX_GET_STC ioctl delivers STCvid, there is no problem.
But I don't know which STC value is returned.

IMHO it is more critical that the DMX_GET_STC ioctl introduces a
(variable) delay which depends on system load, processor speed etc. 

Oliver


-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe vdr" as subject.