[opendtv] Re: ATSC and Lip Sync



Cliff Benham wrote:
Almost ten years ago I wrote an email to this list [see attachment] asking for discussion about the lip sync problems in ATSC. 

Today it is much worse, not better than it was then.
Just watch any local or national news cast and you'll see/hear lip sync disparities sometimes as much as two seconds [!!] worth in various segments.
Somehow all audio must be embedded at 'the' source so it remains synced with the video all the way into my eyes and ears. 
It should just work.
There are three completely different issues here, and none of them is the fault of the ATSC.
1. Programming distributors are doing more separate processing of video and audio without compensating. As you point out, news is a major culprit, and we started noticing the problem with the near-concurrent arrival of digital video effects devices and international news connections roughly 30 years ago (before there was an ATSC looking into ghost cancellation, let alone DTT). DVEs, frame syncs, and the like delay video; few delay audio to match. Fortunately, there are many devices already on the market or coming that can help this part of the problem, IF they are used (and, for most of them, only if they are used in many stages of the production process). Even then, things are tricky. If you're doing a live interview with a wireless camera, and you compensate for the encode-decode delay of the wireless system, the frame syncs that might be used on the wireless camera, the outbound path to the distant site, and the inbound path from the distant site, the encode-decode delays on those paths, and any DVEs that might be used, then, even without satellites, you may be introducing a substantial conversational delay. A solution is using things like the Boxx Meridian wireless system (<1 ms delay) and the NTT ultra-low-latency transmission system (likewise), both demonstrated at NAB 2009 and both requiring lots of bandwidth. And even they don't solve the problem completely. Run fiber between Europe and the U.S., and you're looking at delays (ignoring all of the electronics) in the range of maybe three to ten 29.97-fps frames, depending on locations and routing -- not tragic for a conversation but enough to inhibit intercontinental musician jams and to require interesting echo cancellation even with "mix-minus" production.
2. Non-CRT display technologies introduce video delays. At an HPA Technology Retreat some years back, Bill Hogan showed a video sequence shot of multiple monitors each getting the same video signal with burned-in time code. Three different frame numbers were visible at any time the sequence was frozen. Throw in multiviewers, and the delay increases. The audio to the control room can be delayed to match, but then there's a conflict with intercom and IFB. So it can be hard even to determine what IS correct lip sync at the source.
3. There is no MPEG police force. Although the ATSC standard (like most other audio-and-video bit-rate-reduction standards) includes a perfect way to deal with any encode-decode slip (presentation time stamps), receiver manufacturers rarely (if at all) check them after acquiring a signal. When your lip sync goes out, try changing channels briefly back and forth; it might lock in and then start drifting again. Maybe this could come within FCC purview, but it is not in any current requirements.
In summary, the bottom line is the bottom line. If you're willing to spend $300 for your set-top box instead of $50, and you're willing to pay $20 a month for your favorite news, these problems are resolvable. 

If not, it's probably going to get a lot worse before it gets better.

TTFN,
Mark



----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org 
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: