RE: RE : [Moderator Action] Bugs and TT (was TT and subtitling)

 
No, PCR is indeed often discontinuous in real life
broadcast usage. See ISO 13818-1 for the details.
All real world MPEG-2 decoders accommodate this reality.
 
G.

	-----Original Message-----
	From: Johnb@screen.subtitling.com [mailto:Johnb@screen.subtitling.com] 
	Sent: Thursday, February 06, 2003 9:26 AM
	To: Glenn A. Adams
	Cc: public-tt@w3.org
	Subject: RE: RE : [Moderator Action] Bugs and TT (was TT and subtitling)
	
	

	In response to: 
	>> By subtitling I am referring to my admittedly narrow 
	>> perspective of subtitling and captioning of broadcast video. 
	>> I have looked at SMIL and Quicktime and cannot see how to 
	>> reconcile the timing aspects of these standards with the 
	>> timecode in an external broadcast signal (where the timecode 
	>> may be discontinuous due to advert insertion). Comments please? 

	Glenn A. Adams wrote: 

	>It is necessary to distinguish between media play time, such as 
	>represented by the NPT mechanism, and a broadcast transport 
	>stream's program clock reference (PCR). The latter (PCR) is indeed 
	>discontinuous at insertion in/out points; however, the former (NPT) 
	>is not. 

	Actually isn't it the other way round? 

	In broadcast transport stream (by which I assume you mean a DVB stream) the PCR is always continuous (but wraps at just over 24 hours). Discontinuities in PCR would have a disastrous effect on the decoder. It is the apparent media play time prior to the encoder that is discontinuous.

	The timecode I refer to here is the VITC or LTC timecode seen by any equipment in the broadcast chain downstream from the switch that effects the changeover between program and advert.

	It is common practice for a broadcast program to be timecoded starting at 10 hours. Adverts tend to be timecoded from 1 hour (this helps engineers tell the material type from the timecode display on the front of the tape machine!). So when playing out a program the timecode will start at 10 hours and go forward 1s per s. When an advert occurs the timecode will jump backwards to 1 hour then go forward. Several backward jumps will probably occur! When the interrupted program resumes the timecode continues from the program timecode value prior to the ad-break.

	Subtitle inserters have to match the apparent timecode in the broadcast signal with an absolute timecode value in the subtitle file and output the appropriate subtitle.

	Hope this helps! 

	regards 
	John Birch 

	The views and opinions expressed are the author's own and do not necessarily 
	reflect the views and opinions of Screen Subtitling Systems Limited.