simple-minded non-OFDM signal design question

Randy Yates wrote:


> 
> And to be even more explicit, Vladimir postulated that such equivalent
> performance would require a MLSE (Viterbi) and that such an algorithm
> requires on the order of M^L calculations per symbol. 

No, I didn't.

First, I am talking about the general case and not about the comparison 
of the particular standards.

Second, MLSE is the straightforward solution which does not require any 
special properties of the signal. MLSE is universal; however you can 
design a special signal which allows for more efficient approach then 
MLSE. The key to that is divide and conquer, and this is exactly what 
OFDM does. However you don't have to divide in the FFT manner; you can 
chose another transform whatever you like.

Vladimir Vassilevsky

DSP and Mixed Signal Design Consultant

http://www.abvolt.com

Vladimir Vassilevsky <antispam_bogus@hotmail.com> writes:

> Randy Yates wrote:
>
>
>> And to be even more explicit, Vladimir postulated that such
>> equivalent
>> performance would require a MLSE (Viterbi) and that such an algorithm
>> requires on the order of M^L calculations per symbol.
>
> No, I didn't.

You didn't??? Then what did you mean by this:

  Consider QAM-16 at the rate of 100Mbit/s, i.e. 25 Mbaud. The delay
  spread will be at the order of 100ns, i.e. ~3 symbols. The number of
  operations will be 16^3 = 4096 per symbol, or ~100G ops per second.
  This is not unfeasible however it is not practical either.

? You weren't intimating that with such an algorithm equivalent
multipath performance could be derived from a single-carrier system?

If you weren't, then what were you saying?
-- 
%  Randy Yates                  % "Ticket to the moon, flight leaves here today 
%% Fuquay-Varina, NC            %  from Satellite 2"
%%% 919-577-9882                % 'Ticket To The Moon' 
%%%% <yates@ieee.org>           % *Time*, Electric Light Orchestra
http://home.earthlink.net/~yatescr

Randy Yates wrote:

> Vladimir Vassilevsky <antispam_bogus@hotmail.com> writes:
> 
> 
>>Randy Yates wrote:
>>
>>
>>
>>>And to be even more explicit, Vladimir postulated that such
>>>equivalent
>>>performance would require a MLSE (Viterbi) and that such an algorithm
>>>requires on the order of M^L calculations per symbol.
>>
>>No, I didn't.
> 
> 
> You didn't??? Then what did you mean by this:
> 
>   Consider QAM-16 at the rate of 100Mbit/s, i.e. 25 Mbaud. The delay
>   spread will be at the order of 100ns, i.e. ~3 symbols. The number of
>   operations will be 16^3 = 4096 per symbol, or ~100G ops per second.
>   This is not unfeasible however it is not practical either.
> 
> ? You weren't intimating that with such an algorithm equivalent
> multipath performance could be derived from a single-carrier system?
> 
> If you weren't, then what were you saying?

Oh, you such a bore.

It is not about multicarrier vs single carrier, it is about a bunch of 
thin channels vs one thick channel. It does not really matter in which 
way the thin channels are combined. The point is the optimal demodulator 
for the thick channel is MLSE with the exponential complexity, whereas 
there are better algorithms for a pack of the thin channels.


Vladimir Vassilevsky

DSP and Mixed Signal Design Consultant

http://www.abvolt.com

Vladimir Vassilevsky <antispam_bogus@hotmail.com> writes:

> Randy Yates wrote:
>
>> Vladimir Vassilevsky <antispam_bogus@hotmail.com> writes:
>>
>>>Randy Yates wrote:
>>>
>>>
>>>
>>>>And to be even more explicit, Vladimir postulated that such
>>>>equivalent
>>>>performance would require a MLSE (Viterbi) and that such an algorithm
>>>>requires on the order of M^L calculations per symbol.
>>>
>>>No, I didn't.
>> You didn't??? Then what did you mean by this:
>>   Consider QAM-16 at the rate of 100Mbit/s, i.e. 25 Mbaud. The delay
>>   spread will be at the order of 100ns, i.e. ~3 symbols. The number of
>>   operations will be 16^3 = 4096 per symbol, or ~100G ops per second.
>>   This is not unfeasible however it is not practical either.
>> ? You weren't intimating that with such an algorithm equivalent
>> multipath performance could be derived from a single-carrier system?
>> If you weren't, then what were you saying?
>
> Oh, you such a bore.

Why? Because I don't insult you like you do me (and many other folks
here)? 

Have you considered the possibility that you were not communicating
clearly? No, I'm sure you haven't. 
-- 
%  Randy Yates                  % "Rollin' and riding and slippin' and
%% Fuquay-Varina, NC            %  sliding, it's magic."
%%% 919-577-9882                %  
%%%% <yates@ieee.org>           % 'Living' Thing', *A New World Record*, ELO
http://home.earthlink.net/~yatescr

On Mon, 02 Oct 2006 16:50:46 GMT, Randy Yates <yates@ieee.org> wrote:

>Eric Jacobsen <eric.jacobsen@ieee.org> writes:
>
>> On Tue, 26 Sep 2006 03:31:14 GMT, Randy Yates <yates@ieee.org> wrote:
>>
>>>Randy Yates <yates@ieee.org> writes:
>>>> [...]
>>>
>>>OK, let me tip my hand. 
>>>
>>>There are some folks/circles that are ragingly debating
>>>whether the COFDM used in DVB/T is better than the US's
>>>8VSB. I'm wondering if there is any real difference IF
>>>the end-to-end systems are designed "properly."
>>
>> Oh, gag a maggot.
>
>Wow, I hadn't heard that phrase since high-school, which dates
>you and I both, Eric!
>
>> I can't believe anybody still thinks that 8VSB is a good idea for
>> anything, anywhere, under any circumstances.
>
>Eric, 
>
>First of all, I'm not trying to drum up support for 8VSB. As far as
>I can tell, the situation is that we're stuck with it, so my questions
>are more like, "What can we do with it now that we've got it?", or 
>"Why can't we make it work as well as (C)OFDM?"
>
>Secondly, this is really about some interesting comm theory (IMHO) and
>not about heated standards battles. I have absolutely no joy or
>interest in going (even if anyone would listen to me, which they
>wouldn't because I'm nobody) to the ATSC or FCC about this. There's
>about as much wisdom in that as there is in aspiring to be a lawyer...

Yes, I got that from the beginning but I didn't know how much detail
you wanted.  The background of the selection tells the foundation of
the story, which is important to understand how 8VSB got selected. The
main point is just that the rationale behind the selection was not
because of a technical superiority to anything else available by
today's standards.

>> I've got some opinion on this, which I'll offer FWIW, but it's really
>> just my opinion.   I've worked with both 8VSB and OFDM and am pretty
>> familiar with both the ATSC and DVB-T standards.
>>
>> As far as I can tell the Grand Alliance, 
>
>I guess you mean the ATSC? The Grand Alliance was some political party
>back in 1600's England from what Google told me. (Yeah, I know - you're
>being cute. OK.)

There was a study group called the Grand Alliance that chose the
technology (8VSB) for the ATSC spec.

>> from which we've been afflicted with 8VSB, was populated mostly by
>> old farts who hadn't been exposed to much radio technology other
>> than NTSC and other television standards.  The only two modulation
>> techniques that were seriously considered, by my understanding, were
>> 16-QAM and 8VSB.  The conclusion was that 8VSB was "superior" at
>> least partly because it only needed a real-valued equalizer in the
>> receiver and it was, therefore, simpler to implement.  And since
>> 8VSB looks pretty familiar to somebody who's been looking at NTSC
>> signals through their career, it probably was very comfortable for
>> them to accept it.
>
>That adds up.
>
>> There's a reason you don't see n-VSB showing up anywhere else.   It
>> pretty much just sucks, IMHO.  The Grand Alliance really didn't do us
>> any favors by selecting that.
>
>Well, you really didn't answer my question. You just basically said
>"8VSB sucks." I'm not really contending against that assertion: maybe
>it does. All I'm asking is "Why?"
>
>It sounds like, from subsequent posts with Vladimir in this thread,
>that the answer, basically, is that it takes a lot more computational
>complexity to obtain the same performance with 8VSB as it does with
>OFDM. Do you agree?

No, I don't think it's possible to get the same performance with 8VSB
as is possible with OFDM, regardless of how much complexity you throw
at it.   There are a number of reasons for this.

Start from the basic requirement:  HD TV needs about 20Mbps (actually
19Mbps and change) and due to the legacy NTSC channelization there is
about 6MHz bandwidth in which to transport the 20Mbps.   For
single-carrier technology the obvious candidate is then 16QAM, which
gives a native 4bps/Hz or about 24Mbps.   Put coding and overhead on
top of that and you're back to the required payload rate.

But if what you have in your head for a radio is an NTSC tuner and
you're skeptical of the digital baseband "black box", then you
immediately start to worry about phase noise, group delay in the
filters, and, most of all, a mysterious equalizer that has to work on
complex-valued symbols.   The EQ was deemed a complexity problem and
there were some very simplistic channel studies done that "proved" to
the Grand Alliance that 16QAM would not survive the multipath channel.
Today we know this to be a limitation of their testing or creativity,
because it's very feasible to use 16QAM for this task.   The EQ is not
trivial, but it's definitely workable.

So given that they didn't think 16QAM would work, a clever way to get
4bps/Hz is to use 8-PAM and whack off half the spectrum (i.e., make it
Vestigial SideBand).   Doing so presents some synchronization
problems, so if you put a pilot tone in there (just like for NTSC,
even at the same location) then you can lock to the pilot tone and
recover the symbols.   An added benefit is now that the equalizer is
real-valued (since the base modulation was 8-PAM) and, compared to the
16QAM equalizer you've reduced the complexity by much more than half.

The problem is that the VSB process causes some distortion in a
practical system, the addition of the pilot tone reduces the power
efficiency, and 8-PAM is not a power efficient modulation to begin
with.

An example of applying VSB could be taking BPSK and filtering off half
the spectrum.   The spectral efficiency is then the same as QPSK and
the signal is still real-valued is properly synchronized.  Nobody does
this because it's less painful (and more efficient) to just use QPSK.

So, yes, I think 8VSB is pretty much crap compared to even 16QAM, and
we've not even started talking about OFDM yet, which I do think works
in broadcast channels better than 16QAM due to the issues previously
discussed.

Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

On Mon, 02 Oct 2006 16:50:46 GMT, Randy Yates <yates@ieee.org> wrote:

>Eric Jacobsen <eric.jacobsen@ieee.org> writes:
>
>> As far as I can tell the Grand Alliance, 
>
>I guess you mean the ATSC? The Grand Alliance was some political party
>back in 1600's England from what Google told me. (Yeah, I know - you're
>being cute. OK.)

FWIW, here's a reference explaining the connection between the Grand
Alliance and the ATSC:

http://www.atsc.org/granda.html

Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Hey Eric,

Sorry I let this slide so long, but I did want to finally get back to
you on it. See comments interjected below.

Eric Jacobsen <eric.jacobsen@ieee.org> writes:
> [...]
> Yes, I got that from the beginning but I didn't know how much detail
> you wanted.  The background of the selection tells the foundation of
> the story, which is important to understand how 8VSB got selected. The
> main point is just that the rationale behind the selection was not
> because of a technical superiority to anything else available by
> today's standards.

Are you talking about me or the Grand Alliance? 

I never have thought of 8VSB as being particularly superior, but
neither do I see why it is particularly inferior. Specifically in the
area of multipath susceptability, I don't understand why it couldn't
be made just as robust as, e.g., OFDM.

> There was a study group called the Grand Alliance that chose the
> technology (8VSB) for the ATSC spec.

That makes the picture clearer <ahem>. Thanks for the link, too.

>>> There's a reason you don't see n-VSB showing up anywhere else.   It
>>> pretty much just sucks, IMHO.  The Grand Alliance really didn't do us
>>> any favors by selecting that.

Assertions. I know the assertion - what I'm asking for is the precise
reasons why the assertion is true.

> No, I don't think it's possible to get the same performance with 8VSB
> as is possible with OFDM, regardless of how much complexity you throw
> at it.   There are a number of reasons for this.
>
> Start from the basic requirement:  HD TV needs about 20Mbps (actually
> 19Mbps and change) and due to the legacy NTSC channelization there is
> about 6MHz bandwidth in which to transport the 20Mbps.   For
> single-carrier technology the obvious candidate is then 16QAM, which
> gives a native 4bps/Hz or about 24Mbps.   Put coding and overhead on
> top of that and you're back to the required payload rate.
>
> But if what you have in your head for a radio is an NTSC tuner and
> you're skeptical of the digital baseband "black box", then you
> immediately start to worry about phase noise, group delay in the
> filters, and, most of all, a mysterious equalizer that has to work on
> complex-valued symbols.   The EQ was deemed a complexity problem and
> there were some very simplistic channel studies done that "proved" to
> the Grand Alliance that 16QAM would not survive the multipath channel.
> Today we know this to be a limitation of their testing or creativity,
> because it's very feasible to use 16QAM for this task.   The EQ is not
> trivial, but it's definitely workable.
>
> So given that they didn't think 16QAM would work, a clever way to get
> 4bps/Hz is to use 8-PAM and whack off half the spectrum (i.e., make it
> Vestigial SideBand).   

OK, wait - there seems to be a definite error here. SSB 8PAM is 6 bits/Hz.
At least that's what Proakis says - remember this chart?:

  http://www.digitalsignallabs.com/efficiency.jpg

Right? Or is there something about the "VSB" in "8VSB" that makes
this untrue?

> Doing so presents some synchronization
> problems, so if you put a pilot tone in there (just like for NTSC,
> even at the same location) then you can lock to the pilot tone and
> recover the symbols.   An added benefit is now that the equalizer is
> real-valued (since the base modulation was 8-PAM) 

Assuming you've performed carrier recover - I think that's the standard
architecture, right Eric?

> and, compared to the 16QAM equalizer you've reduced the complexity
> by much more than half.

Right - no complex multiplies!

> The problem is that the VSB process causes some distortion in a
> practical system, the addition of the pilot tone reduces the power
> efficiency, and 8-PAM is not a power efficient modulation to begin
> with.
>
> So, yes, I think 8VSB is pretty much crap compared to even 16QAM, 

Well, all I've heard you really say is that 8VSB is less power
efficient than 16-QAM. I agree with that; from the Proakis chart, if
my eyeball is correct, it looks like about 3 dB more efficient (at
10^{-5} SER).

You seemed to start out implying that 8VSB wasn't as bandwidth efficient
as 16QAM, but that isn't really true. I guess I don't know what you're 
point is, other than the power efficiency issue, about 8PAM vs. 16QAM.

So if I could distill it down, you're saying 8VSB is crap because it
is 3 dB less power efficient? That doesn't sound like a really BIG
disadvantage since this is a terrestial system that has access to
large power reserves. It may not be optimal (i.e., a more power
efficient technique could be chosen that also fits the other requires
such as spectral efficiency), but I don't see why it's "crap."

I'm not playing any games here, Eric. My questions are honest. I look
forward to your response. 

PS: One other thing - I don't see why SSB 8PAM would have synchronization 
problems (any more than SSB BPSK has syncrhonization problems) - could 
you shed some light on that?
-- 
%  Randy Yates                  % "Ticket to the moon, flight leaves here today 
%% Fuquay-Varina, NC            %  from Satellite 2"
%%% 919-577-9882                % 'Ticket To The Moon' 
%%%% <yates@ieee.org>           % *Time*, Electric Light Orchestra
http://home.earthlink.net/~yatescr

Randy Yates <yates@ieee.org> writes:

Eric (et al.), I also wanted to throw this in: I read somewhere
in one of the ATSC documents that the system requires on the order
of 15 dB SNR to get acceptable picture performance. Does that
sound like a reasonable SNR requirement for John Q. Public? 
-- 
%  Randy Yates                  % "With time with what you've learned, 
%% Fuquay-Varina, NC            %  they'll kiss the ground you walk 
%%% 919-577-9882                %  upon."
%%%% <yates@ieee.org>           % '21st Century Man', *Time*, ELO
http://home.earthlink.net/~yatescr

On Fri, 13 Oct 2006 00:45:42 GMT, Randy Yates <yates@ieee.org> wrote:

>Hey Eric,
>
>Sorry I let this slide so long, but I did want to finally get back to
>you on it. See comments interjected below.
>
>Eric Jacobsen <eric.jacobsen@ieee.org> writes:
>> [...]
>> Yes, I got that from the beginning but I didn't know how much detail
>> you wanted.  The background of the selection tells the foundation of
>> the story, which is important to understand how 8VSB got selected. The
>> main point is just that the rationale behind the selection was not
>> because of a technical superiority to anything else available by
>> today's standards.
>
>Are you talking about me or the Grand Alliance? 

Yes.   ;)

I had the GA in mind.

>I never have thought of 8VSB as being particularly superior, but
>neither do I see why it is particularly inferior. Specifically in the
>area of multipath susceptability, I don't understand why it couldn't
>be made just as robust as, e.g., OFDM.

My intuition is that the lack of frequency diversity (for the coding)
in 8-VSB is going to create a gap that's insurmountable.   Since
multipath is the dominant impairment, that's an issue.

In most cases I tell people that in general single-carrier systems and
OFDM systems can both be made to work in the same environments, but in
many cases the complexity required for the SC system makes the OFDM
system more attractive.  

>>>> There's a reason you don't see n-VSB showing up anywhere else.   It
>>>> pretty much just sucks, IMHO.  The Grand Alliance really didn't do us
>>>> any favors by selecting that.
>
>Assertions. I know the assertion - what I'm asking for is the precise
>reasons why the assertion is true.
>
>> No, I don't think it's possible to get the same performance with 8VSB
>> as is possible with OFDM, regardless of how much complexity you throw
>> at it.   There are a number of reasons for this.
>>
>> Start from the basic requirement:  HD TV needs about 20Mbps (actually
>> 19Mbps and change) and due to the legacy NTSC channelization there is
>> about 6MHz bandwidth in which to transport the 20Mbps.   For
>> single-carrier technology the obvious candidate is then 16QAM, which
>> gives a native 4bps/Hz or about 24Mbps.   Put coding and overhead on
>> top of that and you're back to the required payload rate.
>>
>> But if what you have in your head for a radio is an NTSC tuner and
>> you're skeptical of the digital baseband "black box", then you
>> immediately start to worry about phase noise, group delay in the
>> filters, and, most of all, a mysterious equalizer that has to work on
>> complex-valued symbols.   The EQ was deemed a complexity problem and
>> there were some very simplistic channel studies done that "proved" to
>> the Grand Alliance that 16QAM would not survive the multipath channel.
>> Today we know this to be a limitation of their testing or creativity,
>> because it's very feasible to use 16QAM for this task.   The EQ is not
>> trivial, but it's definitely workable.
>>
>> So given that they didn't think 16QAM would work, a clever way to get
>> 4bps/Hz is to use 8-PAM and whack off half the spectrum (i.e., make it
>> Vestigial SideBand).   
>
>OK, wait - there seems to be a definite error here. SSB 8PAM is 6 bits/Hz.
>At least that's what Proakis says - remember this chart?:
>
>  http://www.digitalsignallabs.com/efficiency.jpg
>
>Right? Or is there something about the "VSB" in "8VSB" that makes
>this untrue?

No, I think that's right.   And I had 16-QAM in my head and that may
not be right as to what was used in the comparison.   I found an
abstract here:

http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?tp=&arnumber=720253

that says it was 32-QAM, but, as you point out, that's not quite
jiving with the spectral efficiency comparison.   Now I'm wondering
whether they compared with 64-QAM...ah, crap.

>> Doing so presents some synchronization
>> problems, so if you put a pilot tone in there (just like for NTSC,
>> even at the same location) then you can lock to the pilot tone and
>> recover the symbols.   An added benefit is now that the equalizer is
>> real-valued (since the base modulation was 8-PAM) 
>
>Assuming you've performed carrier recover - I think that's the standard
>architecture, right Eric?

That's the textbook vanilla how-to-demodulate VSB architecture.

An interesting tidbit is that what a lot of people wound up doing in
the receiver is mixing the VSB signal to baseband (i.e., the center of
the transmitted spectrum mixed to DC) and demodulating it as a 64-QAM
constellation (which works, oddly enough).   So the idea of using a
real-valued equalizer was of no benefit to many implementers, who went
the full-complex-demod route, anyway.   My undertanding is that this
is how most 8VSB receivers actually do the processing.

>> and, compared to the 16QAM equalizer you've reduced the complexity
>> by much more than half.
>
>Right - no complex multiplies!
>
>> The problem is that the VSB process causes some distortion in a
>> practical system, the addition of the pilot tone reduces the power
>> efficiency, and 8-PAM is not a power efficient modulation to begin
>> with.
>>
>> So, yes, I think 8VSB is pretty much crap compared to even 16QAM, 
>
>Well, all I've heard you really say is that 8VSB is less power
>efficient than 16-QAM. I agree with that; from the Proakis chart, if
>my eyeball is correct, it looks like about 3 dB more efficient (at
>10^{-5} SER).

For 16-QAM, but if the comparison was really with 64-QAM then there's
no difference in basic power efficiency.    There will be some loss
due to the VSB filtering and the addition of the pilot tone, though.

>You seemed to start out implying that 8VSB wasn't as bandwidth efficient
>as 16QAM, but that isn't really true. I guess I don't know what you're 
>point is, other than the power efficiency issue, about 8PAM vs. 16QAM.

As I recall the spectral efficiency comparisons of what was tested
were fair, but I have to admit I'm confused now about what was
actually used for the QAM tests.   32-QAM is mentioned in the cited
abstract, but 64-QAM would be the equivalent in a spectral efficiency
sense (and was already understood for digital cable).   64-QAM
received as a single-carrier would be hard to make work in heavy
multipath, but, as I mentioned, that's how at least some people are
treating the 8VSB signal, anyway.

>So if I could distill it down, you're saying 8VSB is crap because it
>is 3 dB less power efficient? That doesn't sound like a really BIG
>disadvantage since this is a terrestial system that has access to
>large power reserves. It may not be optimal (i.e., a more power
>efficient technique could be chosen that also fits the other requires
>such as spectral efficiency), but I don't see why it's "crap."

To me if there's no advantage to technique A compared to other
available techniques, but there are disadvantages to it, then that
qualifies it as "crap" because there's no reason to use it.   I think
that's the case with 8VSB.  There doesn't seem to be any real
advantage to it as a modulation technique, but it has some downsides,
including channel survivability.   The frequency diversity provided by
OFDM generally improves the performance of the coding in frequency
selective channels, which is the main reason for using it.   Broadcast
channels are actually a quite good example of the exact sort of
channel that you'd cite for showing off OFDM over SC.

But, as I've mentioned, you can always do other trick in an SC system,
like really macho time interleaving, and really mondo complex
equalizers, that will make it work as well as, or at least nearly as
well as, OFDM.   It's just generally not worth the effort.

But since the GA didn't consider OFDM for their study there were no
good comparisons made.  There were a bunch by other folks after the
fact, but even those weren't very good as far as real apples-apples
tests were concerned.   I didn't think the GAs test of VSB-vs-QAM were
apples to apples at the time, either, but clearly my memory of that
isn't too good any more.   I can't seem to find some of the materials
that I had from then to refresh it, either.  :(

FWIW, in post benefit-of-hindsight analysis, I've seen a fair amount
of opinion that the difference between OFDM (DVB-T) and ATSC that
makes ATSC desirable is the difference in PAPR.   PAPR for 8VSB as
defined by the ATSC is about 6.3dB at the 99.9% percentile.   FWIW, I
did a study comparing some OFDM systems with a 20% RRC 64QAM system
and in that case I was getting about 6dB PAPR for 64QAM and about
9.5dB for an OFDM system with 240 subcarriers.  At that number of
subcarriers PAPR doesn't go up much as more are added, so DVB-T would
be comparable.  It's generally accepted that there's a difference of
about 3dB.    When you're a broadcaster and you're shelling out gobs
of money for extremely high-powered transmitters, that last 3dB is
pretty expensive.   Since European broadcast footprints are much
smaller, and they use single-frequency networks to make them bigger
rather than building bigger transmitters, the PAPR difference is
inconsequential since they're not using as much power.

Oh, and you can't do single-frequency networks with VSB, that's
another advantage of OFDM, FWIW.

>I'm not playing any games here, Eric. My questions are honest. I look
>forward to your response. 

No problem, it's a good discussion.   You're making me think hard
about this stuff again.  ;)

>PS: One other thing - I don't see why SSB 8PAM would have synchronization 
>problems (any more than SSB BPSK has syncrhonization problems) - could 
>you shed some light on that?

Just that it's trickier, since half the power is gone, there's some
distortion due to the VSB filter, and it's harder to find the
suppressed carrier frequency.   SSB BPSK has the same problem, so
you're right that it's no different than that case.   My point was
just that SSB BPSK isn't as easy to synch as BPSK.  For ATSC the pilot
is added to help make up the difference.


Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

On Fri, 13 Oct 2006 00:53:10 GMT, Randy Yates <yates@ieee.org> wrote:

>Randy Yates <yates@ieee.org> writes:
>
>Eric (et al.), I also wanted to throw this in: I read somewhere
>in one of the ATSC documents that the system requires on the order
>of 15 dB SNR to get acceptable picture performance. Does that
>sound like a reasonable SNR requirement for John Q. Public? 

Yes, it's 14.9dB in some of the early docs that I have.   I bet that's
in AWGN, too, which isn't the real channel model.

For the broadcaster, getting the coverage that they want is then
dictated by the power output of the transmitter.   I don't know what
the detection threshold is for the typical NTSC analog receiver, but I
wouldn't be surprised if it's in the same general range.

Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org