[opendtv] Signal levels, probability and contours
- From: "John Willkie" <johnwillkie@xxxxxxxxxx>
- To: <opendtv@xxxxxxxxxxxxx>
- Date: Sun, 3 Oct 2004 08:38:11 -0700
(resent because I screwed up and sent it to the old Topica list.) Sorry if
anybody "resents" it.
While more than a few members of this list (largely many of the lurkers I
have encountered in the wild) know much more about this subject than I, many
of list members are still susceptible to spin on the subject of signal
levels, signal contours and the broader subject of probability.
First, there is NO WAY TO PREDICT SIGNAL LEVELS. Never has been, never will
be, if you ask me.
There are time-honored ways to measure the signal level at a particular
point on the planet (or above or below it) at a particular point in time.
One can log or chart that signal level.
But, there's no way -- even with the most modern of systems -- to predict
with any certainty, what the signal level will be the next second, the next
week, or the next moment.
What the FCC and other regulators have done is to state the desired
transmitted signal level for a particular frequency band, and service areas
of a station. The strongest signal contour is the city-grade signal
contour. Under the rules, stations need to provide a city-grade signal over
all portions of their city of license.
The second signal contour of note is the Grade A contour, which is a few dB
below the city grade signal, and then the grade B contour.
Unfortunately, the idiots that don't like the current satellite white area
situation want to change all the engineering for TV stations because they
are not happy with the fact that FCC uses the grade B signal level to
determine where people cannot receive over the air signals.
Largely, their arguments -- proffered by EchoStar -- are totally full of
unmitigated bullshit. The FCC permits people to do real-world signal
measurements (a friend of mine has done this for EchoStar, but his results
are thrown out by EchoStar if he finds a usable signal where EchoStar wants
none to be found.) However, my friend's signal strength meter is not
industrial grade, and has never been externally calibrated. This is the
quality of Echo Star's engineering.
The FCC also permits viewers to provide alternative signal measurement
analysis. Unfortunately for EchoStar, these are expensive, so expensive
that only if the installation is a multiple dwelling unit one, is it likely
to be done.
OF COURSE PEOPLE HAVE TROUBLE WITH the R-6602 Longley-Rice model. The model
is just that, and since the model was adopted fifty or more years ago, we
have achieved much advances in electronics that make reception easier. The
model does not reflect that. It's a reference model.
Also, the model does not predict absolute or relative signal level. Indeed,
the measurements (for service area) are called F(50,50). They predict the
median signal level at a fixed point, given no interfering terrain. The
criteria are quite clear: the probability of a certain signal level at 50
percent of the receiving locations at that point, 50 percent of the time.
There are even other ways to cut the numbers, to find the highest signal
level one can expect at that point f(90,10) or even the interference
contours f(10,90).
The methodology used to measure -- for proof of performance purposes -- a
station's signal level at fixed points should also be pointed out. The
measurement truck extends a mast 30 feet about ground level. Then, the
truck travels very slowly a fixed distance (I seem to remember it being 200
feet on the level road, taking 50 measurements evenly spaced by time and
distance. The result for that "point" is then averaged. The differences
between the minima and maxima at that point can be quite wide: only the
average is reported.
Since it's probability, of course one can expect at some points in time the
signal will be much higher, at other times, there might be no signal. Lack
of signal is the issue, of course.
Only idiots expect the signal to be as unwaveringly usable at the fringes of
a station's coverage area as they are at the transmitter site. Until early
ATSC receivers came on the market, nobody had ever complained about too high
of signal levels.
So, when I hear people saying that the R-6602 model needs to be thrown out
because they found a higher reading than predicted at their point, I ask
them how many measurements they took, over how long a time, and what the
mimima and maxima were. When they (universally) tell me that they only took
one measurement, at one time, I tell them that's irrelevant and they offer
nothing more than noise. Of course, those are the VERY SAME anecdotes that
EchoStar tries to use (unsuccessfully in court, where their experts are no
longer qualified as expert witnesses) to get access by operation of law to
content that they haven't even tried to negotiate with rightful licensees
for access to.
That's just the kind of junk science that have been eliminated (under the
Merrill-Dow decision by the Supreme Court) in courts. That doesn't mean
that one cant try in Congress to get the anecdotal tail to wag the
engineering dog.
John Willkie
----------------------------------------------------------------------
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:
