Advanced Compression Unlikely to Free Up TV Spectrum for Broadband
June 25, 2010
video compression is unlikely to free up TV spectrum for broadband. At least
any time soon. This was the upshot of one of today’s roundtables at the FCC’s
Broadcast Engineering Forum. Andy Setos of Fox Broadcasting (pictured left) delivered the findings
“As you might know, DTV uses MPEG-2,” he said. “It was completed in 1992 and published
in ’94. Here we are 12 years later... no meaningful efficiency improvements are
expected for MPEG-2.”
The commission is looking to reallocate 120 MHz of broadcast TV spectrum for
its Omnibus Broadband Initiative. Channel-sharing was proposed as one option
for TV stations, but MPEG-2 doesn’t support such a scheme, especially in an
increasingly high-definition environment, Setos’ group said. He was joined by
Richard Chernock of Triveni Digital, David Converse of ABC, Greg Coppa of CBS, Sterling
Davis of Cox, Matthew Goldman of Tandberg Television/Ericsson, Brett Jenkins of
ION, Glenn Reitmeier of NBC Universal, Mark Richer of the ATSC and Peter Sockett,
of Capitol Broadcasting.
“The OBI document proposed two stations sharing a
single channel,” Setos said. “This would require severe constraints on
programming options.... One or both of the HD services would be degraded under
There are examples of channel-sharing through the use of multiplexing, but it’s
carried out under extensive planning and typically by stations under single
ownership, Setos said. Also, while MPEG-2 encoders have improved over the
years, it’s usually at the expense of resolution.
“You’re forced to create a winner and a loser in the HD streams, when one
instantaneously exceeds the bandwidth stream of another,” he said. Both HD signals
likely would be degraded.
“The other option suggested was using statistical multiplexing, or statmuxing,”
Setos said. “Statmux efficiency is dependent on the mix of content.”
Bandwidth requirements are determined on a frame-by-frame basis, and so are not
under a broadcaster’s control. Bandwidth requirements of commercials especially
are not under a broadcaster’s control. Statmuxing two HD services requires,
again, picking a winner and loser, the group concluded. The loser would be put
at a disadvantage. Statmux efficiencies are lost with two separate licensees.
There’s potential with MPEG-4 Part 10, which is more efficient than MPEG-2, and
available. However, the current television infrastructure is built around
MPEG-2. Networks and TV stations encode in MPEG-2, and TV stations in homes
across the country decode it.
“Conservatively, there are 100 million DTV sets in homes, not to mention
converter boxes,” Setos said. At minimum, the group concluded it would take 13
years to realistically implement MPEG-4 Part 10.
Setos noted that the OBI suggested combining HD and SD-only stations, but those
opportunities are dwindling. Licensees are still migrating to HD. Setos said all
of Fox’s owned-and-operated stations have transitioned to HD news, and produce
about five or six hours of it a day. And just about all major league sports is
in HD, and scripted content is around 90 percent in HD. Also, 75 percent of the
commercial advertising on Fox is in HD. The industry trend is toward 100
percent HD, he said.
“All these programs are being produced in HD. SD equipment is no longer being
manufactured,” he said later during the Q&A. “The fact is, if you took, a
snapshot of viewing... the density of HD content is extraordinarily high. We
simply observe that this is a real trend.”
-- Deborah D. McAdams