Disney has to drool when they look at Time-Warner's DVD revenues.
The problem with DTV audio sync.............................
tvbroadcast.com
DTV Signals Out Of Sync
By Robin Berger
(May 10, 1999) If you want to become instantly unpopular in this industry, just ask some of the various consumer electronics manufacturer whether they've experienced any synchronization problems between the audio and video signals in their new digital TV receivers. By most accounts, it's a "growing pain" of the digital television (DTV) infrastructure yet to be worked out.
But behind closed doors, the Advanced Television Systems Committee (ATSC) subgroup that tracks these issues concluded, "clearly there is work to be done."
Latency generated by the new DTV gear has been estimated at anywhere from "a matter of milliseconds" to "several seconds," with the audio preceding as well as trailing the video. Home viewers will notice the problem if audio lags video by as little as 30 to 68 milliseconds (a frame is 34 milliseconds), depending upon who you talk to. And one industry guru, who preferred not to be named, said the audience will also detect a bug if the audio is as little as 17 milliseconds too early.
What's more, this A/V latency varies from receiver to receiver and can get worse as a feed progresses.
So what needs to get tweaked? Broadcasters pointed the finger at receiver manufacturers and vice versa. More than one source indicated that certain aspects of the MPEG standard were not well defined, like the timing reference point for audio/video streams (which, they concluded, could cause decoders to make untrue assumptions about incoming streams).
According to Craig Todd, senior member of the technical staff at Dolby Laboratories, what's missing is an analysis tool to verify that encoders are correctly applying time stamps.
"The tool would be a computer program that could analyze the bit stream produced by the transmission equipment," he elaborated. Todd noted that he'd actually seen a prototype of this kind of software, and hopes it becomes commercially available soon.
To date, broadcasters have used homemade remedies to determine whether their encoder switchers are set correctly.
Frank Geraty, director of broadcast operations and engineering at KTLA, in Los Angeles, is working with Dolby to find a better solution--one that "would generate a video flash and an audio tone with visible time code."
The station relied on its own test tape to analyze the digital broadcast it did of the Tournament of Roses Parade last January (see 2/99 issue, page 66). The on-air test was done using digital receivers manufactured for the general public.
But, Dolby's Todd pointed out, it's better to test the transmission in isolation (i.e., have software analyze a bit stream).
"You can use a receiver, but then you're dependent on the accuracy of the receiver," Todd elaborated.
Washington DC's Model HDTV Station Project is also consulting with Dolby, and it recently decided to replace its own homemade test tapes with a "lip sync baseband tape" ordered from the Sarnoff Corporation. (The Sarnoff Corp. declined to comment on the technology, preferring to make any announcements about encoder testing at NAB.)
However, said Model Station president Bill Miller, getting better test equipment for encoders is only part of the solution. Unlike KTLA's Geraty--whose Zenith, Panasonic and Harris sets "perform identically"--Miller found marked differences among the models he had. In fact, his group included receivers, he concluded, that failed to check time stamp references throughout the transmission.
"Apparently they check only when a program first airs and not thereafter," Miller surmised. His theory came after he ran tape loops through a computer server and noted that even those in sync at the beginning could fall out of sync by the end of the tape.
Lew Zager, vice president of technology at Washington's WETA station, also noted that decoders handled "problem bit streams"--those with faulty time stamps--differently.
"Some decoders said, 'yeah I can handle it,' others just spit up," Zager explained.
Dean Malmstrom, head of Sarnoff Corp.'s digital television professional products division backed Zager's assessment, adding "most of today's encoders use only a part of the syntax defined by MPEG."
As a result, he warned, "no amount of testing will guarantee that a decoder will work under any condition with bit streams from any encoder;" though a well-executed, well-crafted test plan "goes a long way towards that goal."
Despite the diversity of opinion on the magnitude of the A/V sync problem and its cause, one conclusion was unanimous: this too shall pass. That, and the need to air this message.
"I think that there is a tremendous fear that the story will go out
that the system doesn't work," said one observer, "that the press will interpret these problems as mortal problems when they are just growing pains."
|
