In an attempt to discover the root cause for the audio dropouts occuring on the TU-DST50/51's SPDIF output, my colleagues and I captured and analyzed the SPDIF output using a PC sound card.
Avoiding the gorry details, we believe that the audio dropouts are occuring due to bad time stamps in the ATSC transport stream as played back from a D-VHS deck. The time stamps embedded in a transport stream indicate the sense of time at the transmit end of the signal chain; these time stamps are used to keep audio and video synchronized and properly buffered at the receive end. Whenever a transport stream is re-transmitted with a new clock, the data must be restamped to the new clock's timing.
What is likely happening in the Panasonic combo (and in the JVC DH30000 deck as well) is that the time stamps from the original broadcast are being used upon play back, but the D-VHS deck, with its own transmitting clock, is not restamping the data to its own clock or is not adjusting its transmit clock to sync up with the time stamps embedded in the recorded data.
The net result is the MPEG decoder experiences a buffer overrun or underrun due to the flawed time stamps, and hence audio gets dropped, repeated, or buffered with zeros (bad).
I'll follow up later with an explanation of what we found in the SPDIF output, and how different AC-3 decoders react to anomalies in the AC-3 data.
Displays: Sony VPL-HW40ES | 110" SI Motorized Slate screen | Pioneer PRO-150FD
Sources: Sony BDP-S5000ES | Dune HD Smart D1 | Roku HD | LG LST-3410A HD-DVR
Speakers: Nearfield Pipedreams 18 (L,R,Sub1,Sub2), Pipedreams 15 (C)
| Magnepan MC-1 (WL, WR, SL, SR, SBL, SBR)
Audio: Denon AVR-4311CI | Bryston 6B-SST | Bryston 14B-SST
Control: Control4 home automation