--- In HDTV-in-SFbay@yahoogroups.com
, "epicurusradium" <vinced@...> wrote:
> What I do know is that with 18 ft of coax and an amp of 20db gain The maximum SS reading on the Sony is typically (summer weather) 58 to 65 reading here on KNTV ch 12. While KGO digital, when back on ch 24 was between 78 and 88 year round. Channel 20.x and 65.x also run in the same signal strength range regardless of temp. KRCB in Cotati typically is from 70 to 84! There have been a few days when it is not thereat all.
> Ok, let me see if I understand how the digital signal looks.... since it is composed of bits consisting of 1's & 0's, the 1's representing a uniform signal value of X and the 0's representing essentially no signal or a very much lesser amount, the point where all the bits representing 1's are no longer seen (decoded) is at the 'cliff' signal level in the receiver's tuner.
It's a lot more complicated than that.
8vsb is actually an amplitude modulated signal with 8 different legal "amplitude" steps. At each sampling point, the signal is supposed to be at one of those steps. If you feed an 8vsb signal into an oscilloscope with the sampling point being the trigger, you get the classic "eye" diagram. You'll see just a slew of noise, but every so often (at the sampling points), you'll see clear "eyes" open up where the signal settles down to one of those 8 amplitude steps.
Since there are 8 steps, that's 3 bits (2^3=8). So each sampling point represents a "symbol" of 3 bits. The baud rate (symbols per second) is about 10 megabaud. This is because the Nyquist filtering - to cram the whole thing inside of a bandwidth of a little more than 5 MHz, the maximum sampling rate you can use for a digital signal is a little more than 10 megabaud. Since there's 3 bits per symbol that's ~30 megabits per second. But not every progression of one symbol to the next is legal. Trellis coding is done to dictate the symbol progression. Think of trellis coding as a dance: if you just moved your right foot, you can't move it again until you move your left foot.
Anyway, a visual representation of the digital cliff effect would be how "open" those eyes are. That is, how easy is it to tell which of the 8 amplitude levels the signal is at when a sampling period happens. A good, solid signal will pigeonhole the signal perfectly in one of the 8 spots, making it easy for the trellis decoder to follow along. When the signal smears out, the trellis decoder has to follow a few steps of the dance to see which levels were the actual ones being used. Do that too much and the trellis decoder can be lead down the wrong path and lose synchronization, forcing it to start over.
> It would then seem that as far as the TV tuner is concerned a peak level (1 bits) must be high enough in amplitude to be above the 'cliff'. Therefore the receiver does not care that the digital signal has an average power that may or may not be higher than that of an analog signal. PG&E however probably does.
> Multipath; If the ratio of reflected signal to direct signal is determined by physical objects in the local environment, how would that change if the transmitted ERP is increased?? Logically it would not.
The ratio would not increase, but when any signal (main or reflection) gets low enough, it merges into the noise floor, whereupon it no longer has any impact on reception.
The other issue with high power is that the broadcasters around here operate fairly close to urban centers. Even Mt. Allison and Monument Peak are not too far away from Fremont and Milpetas. If they were to just blast out huge amounts of power, they'd wind up desensitizing local receivers, making it hard for those folks to receive other programming.
Not only that, but Sutro Tower is located in a residential neighborhood. There are RF exposure safety rules that all generators of RF have to be cognizant of (even us Amateurs have to worry about field strength if we use large amplifiers and high ERP antennas), and the folks who live around Sutro are a fairly vocal bunch of malcontents, from what I see on the web.