[time-nuts] Using digital broadcast TV for timing?

Magnus Danielson magnus at rubidium.dyndns.org
Thu Feb 9 01:14:36 UTC 2012


On 02/08/2012 05:41 AM, David McGaw wrote:
> The DTV signal coding has a major problem in that it is not
> deterministic - they can't even synchronize the audio and video. There
> have been many workshops on this at the Audio Engineering Society
> conventions. I doubt there is any useable timing in it.

You are mixing up the issue of the relative timing between the picture 
and audio (which is a mess at times) with the issue of having good 
timing in the transmitted signal.

The signal is a very layered thing, where PCR has been mentioned.

The audio and video is compressed and encoded into a program stream. 
Several program steams is then muxed together into a transport stream, 
all according to the ITU-T H.222.0 spec. The PCR ticks in the universal 
27 MHz rate (older MPEG-1 systems used 90 kHz).

The transport stream is then padded into fixed bit-rate, wrapped in 
various ways and then transmitted using some modulation. For instance, 
DVB-T uses a COFDM modulation. Using a combination of the COFDM 
structure, pilot-tones, etc. fairly good frequency and fairly stable 
phase should be retrievable. The absolute phase of the received phase 
could be troublesome to find out.

Chances are that a GPS receiver is feeding you through the transmitted.

Many time-nuts have matching GPS receivers (such as a Thunderbolt) 
already running for them.

Cheers,
Magnus




More information about the Time-nuts_lists.febo.com mailing list