ASTC 8VSB not a bad choice for Canada, as 95% of the househoulds get TV via cable -- where 256 QAM is the standard...

"Paul Keinanen" wrote:

The spectral efficiency in both ATSC and DVB-T is about 3 bits/s/Hz.

Yes, and that's adjustable in COFDM. The closest equivalent spectral
efficiency, between basic 8T-VSB and COFDM, is achieved when COFDM is
set to 64-QAM, 3/4 convolutional FEC, and 1/16 guard interval. The next
step is to compare the C/N margin needed to achieve reception, in
different types of RF propagation channel.

Early 8-VSB receivers did quite poorly in Rayleigh channels (no direct
path between transmitter and receiver), so they got a bad reputation.
But ever since 2002, much better receiver designs have become available,
designs which use multiple redundant techniques to achieve symbol sync,
and which manage pre-echo far more cleverly than the inadequate 1st
generation ones. Part of the problem was that people had underestimated,
back in the early 1990s, just how much echo tolerance was required for
reasonably good reception. It was thought back then that just a few usec
of pre-echo tolerance was perfectly adequate, where instead many tens of
usec of pre-echo tolerance may be required in urban or indoor antenna
settings. So now, I think these "physical layer" differences between
DVB-T and ATSC don't matter much anymore.

And aside from the difference of the modulation scheme, ATSC and DVB-T
are much more similar than they are different.

The reception distances claimed by many analog System-M/NTSC viewers
seems to be quite large compared to what is achievable by
System-B/D/G/H/I (PAL) viewers in Europe, even if we consider the
bandwidth difference 6/7/8 MHz.

No doubt, in the US plains the reception distances might be more
common, but in Europe, such plains are rare, perhaps Ukraine would
contain such plains.

Before doing comparisons between digital systems, the analog side
should first be normalised.

I suppose. But, you know, much of the difference there is simply caused
by transmitter power. In analog and digital, US transmitters tend to be
higher power than many European ones.

For example, the Eiffel Tower DVB-T transmitters in Paris only emit 20
KW ERP per 8 MHz multiplex, in the UHF band
(http://www.csa.fr/pdf/ParisTourEiffel.pdf). In the US, UHF transmitters
are almost always more powerful than that. I'd say anywhere from 50 KW
all the way to 1 MW ERP is common, for each 6 MHz UHF multiplex.

Bert

"Bob Miller" wrote:

SFN's and on channel repeaters are very important tools to have
especially where you have "vast land areas" like Australia, Russia and
China.

Bob, as you well know, this is certainly not the case. Large area SFNs
make little sense, require way too many towers, and if anything, make
reception more troublesome and unpredictable.

Bert

Albert Manfredi wrote:
"Bob Miller" wrote:

SFN's and on channel repeaters are very important tools to have
especially where you have "vast land areas" like Australia, Russia and
China.

Bob, as you well know, this is certainly not the case. Large area SFNs
make little sense, require way too many towers, and if anything, make
reception more troublesome and unpredictable.

Bert

It certainly is the case. I did not mention large area SFNs though I
think they are possible.

There are a lot of ways to use an SFN and or on-channel repeaters and
Australia is experimenting with a number.

Bob Miller

Albert Manfredi wrote:
"Bob Miller" wrote:

SFN's and on channel repeaters are very important tools to have
especially where you have "vast land areas" like Australia, Russia and
China.

Bob, as you well know, this is certainly not the case. Large area SFNs
make little sense, require way too many towers, and if anything, make
reception more troublesome and unpredictable.

Bert

It certainly is the case. I did not mention large area SFNs though I
think they are possible.

There are a lot of ways to use an SFN and or on-channel repeaters and
Australia is experimenting with a number.

Bob Miller

On Tue, 6 Mar 2007 15:44:15 GMT, "Albert Manfredi"
wrote:

"Bob Miller" wrote:

SFN's and on channel repeaters are very important tools to have
especially where you have "vast land areas" like Australia, Russia and
China.

I don't see why SFNs would be especially important in those countries
with large plains. SFNs are nice as gap fillers in difficult terrain.

Of course if you do not intend to cover the whole area but only cover
selected villages with small repeaters then this might be a good idea,
so you can use the channel in each village a few kilometers away from
each other.

However, with village distance of 10-50 km, you are still going to
have problem due to insufficient guard time. If the villages are in
different valleys and thus sufficient isolation between the
transmitters, this would not be an issue.

Bob, as you well know, this is certainly not the case. Large area SFNs
make little sense, require way too many towers, and if anything, make
reception more troublesome and unpredictable.

The number of towers is not an issue these days, since the cellular
phone service will require quite a lot small towers. Putting SFN
transmitters in 30-90 m high towers would be quite realistic.

When the GSM cellular phone service started in Finland in the early
1990's, each operator started to build their own towers at each
available hill along each main road. Due to environmental concerns,
the phone companies were required by law to lease space for the
antennas to the competitor in each tower, thus reducing the actual
number of towers built. In practice, the phone companies created a
company to manage this tower and equipment room management.

I don't see a problem, why phone companies and broadcasters could not
share the same towers for cellular phone as well as SFN broadcasting.
Due to the low power required by such SFN broadcasting systems, there
would not be much problems to the cellular phone receivers due to
power levels and phase noise generated by the DVB-T transmitters.

Paul

On Tue, 6 Mar 2007 15:41:59 GMT, "Albert Manfredi"
wrote:

"Paul Keinanen" wrote:

The reception distances claimed by many analog System-M/NTSC viewers
seems to be quite large compared to what is achievable by
System-B/D/G/H/I (PAL) viewers in Europe, even if we consider the
bandwidth difference 6/7/8 MHz.

No doubt, in the US plains the reception distances might be more
common, but in Europe, such plains are rare, perhaps Ukraine would
contain such plains.

Before doing comparisons between digital systems, the analog side
should first be normalised.

I suppose. But, you know, much of the difference there is simply caused
by transmitter power. In analog and digital, US transmitters tend to be
higher power than many European ones.

For example, the Eiffel Tower DVB-T transmitters in Paris only emit 20
KW ERP per 8 MHz multiplex, in the UHF band
(http://www.csa.fr/pdf/ParisTourEiffel.pdf). In the US, UHF transmitters
are almost always more powerful than that. I'd say anywhere from 50 KW
all the way to 1 MW ERP is common, for each 6 MHz UHF multiplex.

This would explain quite a lot.

In Europe, anything over 1 MW ERP is very rare on the analog side on
UHF and the DVB-T ERPs are usually 10-15 dB below the analog ERPs for
comparable viewing areas, so the Paris case with 20 kW digital and 700
kW analog would be typical.

While this power ratio seems to be quite sufficient most of the time,
however, during slow fades caused by abnormal atmospheric refraction
conditions, the signal may vary with more than +/-10 dB, which in
analog reception only may cause some extra noise into the picture and
sound, but in a digital system (DVB-T and the old NICAM sound) going
below the sharp threshold level will make the reception impossible.

Since the analog and digital systems suffer differently during a slow
fade, IMHO, the fade margin for the digital system should be larger
from the beginning, thus running the digital DVB-T service more than
about 10 dB below the analog ERP may reduce the usability of the
system during abnormal atmospheric conditions. It appears that in UK,
they originally started with a much larger ERP difference, but had to
upgrade the digital power to maintain the same service area and
reliability.

Paul

"Bob Miller" wrote:

SFN's and on channel repeaters are very important tools to have
especially where you have "vast land areas" like Australia, Russia
and China.

It certainly is the case. I did not mention large area SFNs though I
think they are possible.

Both of the above are your words. Did I misinterpret what you wrote?

If you expect to cover, uniformly, "vast land areas" such as Australia,
Russia, or China with nationwide SFNs, I call that (a) untenable,
meaning possible but absurd, and (b) large area.

If you don't intend ubiquitous coverage, then of course SFNs are not
hard to install. You can certainly have pockets of coverage, where each
isolated pocket uses the same set of RF frequencies. A no brainer.

On channel repeaters are also a doable do. The CRC demonstrated those
already, also using 8-VSB, in Ottawa. The trick with OCRs is to ensure
that receivers have adequate pre-echo tolerance and to ensure that a
stronger main signal overpowers the OCR signal before echo tolerance
limits of receivers are exceeded. Directional antennas for the OCRs can
help, by reducing areas in which the main signal may appear as a strong
pre-echo with large separation from the repeated signal.

Bert

On Tue, 06 Mar 2007 18:12:38 GMT, Bob Miller wrote:

Albert Manfredi wrote:
"Bob Miller" wrote:

SFN's and on channel repeaters are very important tools to have
especially where you have "vast land areas" like Australia, Russia and
China.

Bob, as you well know, this is certainly not the case. Large area SFNs
make little sense, require way too many towers, and if anything, make
reception more troublesome and unpredictable.

Bert

It certainly is the case. I did not mention large area SFNs though I
think they are possible.

Certainly they are, however, the design philosophy is closer to a
cellular phone network than a traditional broadcast network.

Due to the guard time issue, you really have to keep the towers short
enough, in order to avoid the situation of having significant amount
of signals (at a specific receiver site) with more propagation delay
that the guard interval will allow.

There are a lot of ways to use an SFN and or on-channel repeaters and
Australia is experimenting with a number.

In Finland, experiments are done with an antenna at the roof of an
apartment building, receiving horizontally polarised DVB-T signals,
which are then retransmitted downwards to other floors using vertical
polarisation (to help in keeping the repeater input-output isolation
at a sufficient level). The top floors may receive the original signal
directly from the main transmitter, while the lower floor only use the
repeated signals. Middle floors receive signals which are a
combination of the direct horizontally polarised signal and the
repeated vertically polarised signal.

Paul

"Paul Keinanen" wrote:

While this power ratio seems to be quite sufficient most of the time,
however, during slow fades caused by abnormal atmospheric refraction
conditions, the signal may vary with more than +/-10 dB, which in
analog reception only may cause some extra noise into the picture and
sound, but in a digital system (DVB-T and the old NICAM sound) going
below the sharp threshold level will make the reception impossible.

Since the analog and digital systems suffer differently during a slow
fade, IMHO, the fade margin for the digital system should be larger
from the beginning, thus running the digital DVB-T service more than
about 10 dB below the analog ERP may reduce the usability of the
system during abnormal atmospheric conditions. It appears that in UK,
they originally started with a much larger ERP difference, but had to
upgrade the digital power to maintain the same service area and
reliability.

I agree completely. It is those fade margins that have to be considered
with digital.

The number of towers is not an issue these days, since the cellular
phone service will require quite a lot small towers. Putting SFN
transmitters in 30-90 m high towers would be quite realistic.

The problems are two:

1. In the US, terrestrial broadcasters don't have tax revenues or
subscription fees to help pay for installing and maintaining a large
number of towers. And also, the majority of households subscribe to
cable or satellite, so broadcasters rely heavily on these other
distribution media. As a result, terrestrial broadcasters tend to be
very parsimonious when it comes to their OTA plants (alas!).

2. A cellular telephone system is not an SFN, as I'm sure you know.
Multiple cell towers are really nothing more than a network of
translators. If you try to obtain ubiquitous coverage with a wide area
SFN, mutliple towers carefully synchronized, it takes many antennas, and
areas outside the antenna cluster will get very unreliable service. As
weather conditions change, they will cause the signal from different
antennas to interfere in unpredictable ways (exceeding the GI in
unpredictable ways). So in large area SFNs, coverage becomes reliable
only in close proximity to the antennas.

I think the safe way to implement SFNs is to make them small area,
meaning two or three towers, all within the GI, primarily intended to
make urban reception easier. Distant reception is not really helped by
this, compared with a single big stick. If the urban SFN uses only small
towers, distant reception will actually be hurt compared to a big stick.

By the way, The French CSA has a very good document on these topics (en
Francais):

http://www.csa.fr/pdf/Rapport-GT2-As..._de_la_TNT.pdf

Bert

Paul Keinanen wrote:
On Tue, 6 Mar 2007 15:41:59 GMT, "Albert Manfredi"
wrote:

"Paul Keinanen" wrote:

The reception distances claimed by many analog System-M/NTSC viewers
seems to be quite large compared to what is achievable by
System-B/D/G/H/I (PAL) viewers in Europe, even if we consider the
bandwidth difference 6/7/8 MHz.

No doubt, in the US plains the reception distances might be more
common, but in Europe, such plains are rare, perhaps Ukraine would
contain such plains.

Before doing comparisons between digital systems, the analog side
should first be normalised.
I suppose. But, you know, much of the difference there is simply caused
by transmitter power. In analog and digital, US transmitters tend to be
higher power than many European ones.

For example, the Eiffel Tower DVB-T transmitters in Paris only emit 20
KW ERP per 8 MHz multiplex, in the UHF band
(http://www.csa.fr/pdf/ParisTourEiffel.pdf). In the US, UHF transmitters
are almost always more powerful than that. I'd say anywhere from 50 KW
all the way to 1 MW ERP is common, for each 6 MHz UHF multiplex.

This would explain quite a lot.

In Europe, anything over 1 MW ERP is very rare on the analog side on
UHF and the DVB-T ERPs are usually 10-15 dB below the analog ERPs for
comparable viewing areas, so the Paris case with 20 kW digital and 700
kW analog would be typical.

While this power ratio seems to be quite sufficient most of the time,
however, during slow fades caused by abnormal atmospheric refraction
conditions, the signal may vary with more than +/-10 dB, which in
analog reception only may cause some extra noise into the picture and
sound, but in a digital system (DVB-T and the old NICAM sound) going
below the sharp threshold level will make the reception impossible.

Since the analog and digital systems suffer differently during a slow
fade, IMHO, the fade margin for the digital system should be larger
from the beginning, thus running the digital DVB-T service more than
about 10 dB below the analog ERP may reduce the usability of the
system during abnormal atmospheric conditions. It appears that in UK,
they originally started with a much larger ERP difference, but had to
upgrade the digital power to maintain the same service area and
reliability.

Paul

They did increase power levels in the UK but they are still at very low
levels. There will be another increase after analog shutdown.

Bob Miller