[linux-dvb] [RFC] associating pcr with vcxo

Johannes Stezenbach js at linuxtv.org
Tue Jul 26 18:02:06 CEST 2005


Andreas Oberritter wrote:
> On Tue, 2005-07-26 at 11:23 +0200, Johannes Stezenbach wrote:
> > Andreas Oberritter wrote:
> > > if you have multiple displays and/or Picture-In-Picture decoding, then
> > > you have to define one PCR per display as master for vcxo control.
> > > 
> > > Therefore I propose the following simple interface:
> > > 
> > > struct dmx_stc_mapping {
> > > 	unsigned int clock_num;
> > > 	unsigned int pcr_num;
> > > }
> > > 
> > > #define DMX_SET_STC_MAPPING _IOWR('o', 51, struct dmx_stc_mapping)
> > 
> > I think the naming is a bit misleading, because in your first
> > sentence you say that this is about setting the "master for
> > vcxo control". I.e. (if I get it right) you have one 27MHz
> > crystal, but two PCR/STC. Obviously the vcxo can be synced to
> > only one of them, so you have to choose (for PiP usually the
> > one with the bigger picture).
> > 
> > How about:
> > 
> > struct dmx_vcxo_control {
> > 	unsigned int vcxo_num;	// in case you have more than one vcxo
> >  	unsigned int pcr_num;	// which PCR to sync the vcxo to
> > };
> > 
> > #define DMX_VCXO_CONTROL _IOWR('o', 51, struct dmx_vcxo_control)
> > 
> > or maybe DMX_SET_VCXO_SYNC_SOURCE?
> 
> You are right, "stc" was quite confusing.

STC was actually not confusing, but "mapping" was.

> I've been told that using a VCXO is not the only way to synchronize the
> transmitter and receiver clock, so how about your proposal with "vcxo"
> replaced by the more generic "clock"?

Maybe we should use "STC" then. What hardware is there when
no VCXO is used?

Johannes




More information about the linux-dvb mailing list