[opendtv] Ed Williams 1999 Sinclair Baltimore Report
- From: "John Shutt" <shuttj@xxxxxxxxx>
- To: "OpenDTV" <opendtv@xxxxxxxxxxxxx>
- Date: Fri, 25 Mar 2005 16:34:08 -0500
Here is Ed's report of his observations in Baltimore in 1999. I found it on
Google, but the Web-Star site link no longer worked, so this is from
Google's cached version.
John Shutt
I visited the Sinclair COFDM/8-VSB tests in Baltimore yesterday (July 14,
1999).
I was among 11 others that visited that morning. Another demonstration was
scheduled for the afternoon. Sinclair is accommodating up to 25 visitors a
day and hopes to keep the facility on the air for the rest of July. With any
luck they expect to have some HD material for the COFDM signal as they do
for 8-VSB.
The transmission facility was reasonably well setup in a spacious and well
equipped transmitter building. There is not much documentation but the high
quality test equipment indicated both signals were in good form. About 28 dB
SNR and 3.1% EVM for the 8-VSB signal - well within normal operating
specifications. The COFDM system operated in 2K, 64QAM, 3/4 FEC, 1/8
spacing, 24.88 Mbps gross data rate mode. Useful data rate for 8-VSB data
rate was 19.39 and COFDM was 18.66 Mbps. Close enough for useful comparison.
Bandwidth occupancy was 5.7 MHz for COFDM and 5.38 for 8-VSB. The wider
bandwidth for COFDM forced a compromise on the bandpass filter so it did not
meet FCC specs - not needed for this test.
The DTV transmission was on channel 40 (the DTV assignment for Sinclair's
channel 54 station) with an ERP of about 50 kW average for both signals
using a 30 gain antenna at about 1000 ft above ground. Peak-to-average power
was 8.5 dB for COFDM and 6.5 dB for 8-VSB. One of four 60 kW IOT power
amplifiers normally used for Sinclair's channel 45 station was used for the
DTV tests at an average output power of about 5.5 kW.
The COFDM signal consisted of a 7 second clip of standard definition PAL
material with an encoded rate of about 3.5 Mbps followed by bit stuffing in
the Rohde & Schwarz modulator for the 18.66 Mbps total. For 8-VSB some HD
material was provided from a server. There was no real complaint about using
SD on one and HD on the other but it did generate some discussion about what
effect the difference could make in reception.
The receiving setup seemed a bit complicated but there were good reasons for
each element of the system all well explained by Mark Aitken, Ray Kiesel and
Harvey Arnold, the Sinclair engineers. Two UHF antennas, a single stand
alone bow-tie (comes with most TV sets) and double bow-tie with reflector
(Radio Shack), were used for the demonstrations. A diagnostic display on one
of the COFDM receivers provided indications of the transmit mode of the
COFDM system.
As for the two sites visited my experience was along the lines of that
reported by Mark Shubin in his earlier report to this group. COFDM is easy
to receive and 8-VSB is more difficult. Both can be made to fail, of course.
However, I found (Sinclair allowed us to conduct our own tests at each site)
that at locations where both 8-VSB and COFDM could be received the margin to
threshold was nearly identical although there were significant variations
among the two COFDM receivers and the two 8-VSB receivers.
By the way, there was an NTSC receiver available (connected to an MATV
system) at the first site (11th floor apartment) but no one in the group
expressed an interest in using it for comparison on an antenna although we
were told that NTSC pictures were not very good at this location because of
the substantial amount of multipath.
Several of the observers (including a receiver mfg) commented that the
signal level variations displayed on the spectrum analyzer under both
receivable (+/- 5 dB) and non-receivable (+/- 10 dB) conditions for 8-VSB
did not appear to be unreasonably distorted. That was my view as well having
seen many DTV signals on the DTV Express and having decoded most of them but
using different make receivers. So it may be that the difference in
reception capability is in the design of the receivers.
My immediate reaction to the demonstration is there clearly is a difference
in receivability. We would not have seen that difference without the efforts
of Sinclair to mount this impressive demonstration.
The next step is to determine the theoretical limits of 8-VSB reception
followed by implementation of circuitry to meet the criteria as closely as
possible. Then I think another test might be in order. It is no longer a
comparison between NTSC and DTV, it is now between two conceptually
different DTV systems.
Ed Williams, PBS
email: ewilliams@xxxxxxx
----------------------------------------------------------------------
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:
