Originally Posted by Trip in VA
Not a clue. I've been assured that it is somehow possible. Maybe the idea is that the signal from the booster will be too weak by the time the echo exceeds 30 ms? Who knows.
The Axcera paper cited below talks about WPSX in State College PA and says this: "The time offset between the transmitters was adjusted to put the concurrent arrival time point approximately over the Rattlesnake site." So apparently they do what you said - put a delay on one of the transmitters so that the two signals arrive at very close to the same time at a particular location. In the case of WFMZ the Allentown transmitter is quite a bit stronger than the proposed Roxborough transmitter plus there are hills in the way between the two sites (which is why they want to do a DTS in the first place) so maybe between these two facts they can get it to work, i.e. put a delay on Roxborough so that south of the two transmitters the two signals arrive within 30 ms of each other and further north depend on the strength of the Allentown transmitter and the intervening hills to prevent the echo from Roxborough from causing problems.
On the contrary, I'm not aware of any watermarking. My understanding is that the two signals must be identical or else they will be treated as interference rather than multipath.
From Axcera paper on DTS:http://www.axcera.com/downloads/tech...hite_paper.pdf
To aid in system setup and diagnostics, the distributed transmission system includes a unique “watermark” signal for each transmitter in a network. The watermark signal is a low-level code, buried underneath the ATSC symbols. It appears as noise to receivers. The noise level is small enough so that it will have little or no practical effect on threshold (typically only about 0.1 dB).
The RF watermark sampling instants are the same as the ATSC symbols. This creates eight minor eye openings in the demodulated I channel, with the normal seven major eye openings interspersed as shown in Figure 5. This figure was produced by a software simulation of the RF watermark injected at a bury ratio of 27 dB.
The RF watermark signal will typically be operated about 30 dB below the transmitter’s average output. However, the coding benefits from a 54 dB coding gain when integrated over one ATSC field. So, it is theoretically possible to identify a buried code from a transmitter that is 24 dB weaker than an interfering transmitter in just one field.
Additional averaging, over multiple fields, would increase the decoding gain.
A constellation display derived from the same simulation is shown in Figure 6. This display shows that each I channel line splits into two. The distance between the two constellation lines is proportional to the RF watermark injection level. The objective in creating the RF watermark system is to be able to identify individual components of a received aggregate signal, and to associate the received components with individual transmitters, as shown in Figure 4 above. Even if one transmitter’s signal is buried underneath a stronger signal, it is still possible to identify the components from the weaker transmitter because of the coding gain of the watermark signal.