Re: Strength or no errors?
- From: Jim Lesurf <jcgl@xxxxxxxxxxxxxxxxxx>
- Date: Mon, 03 Oct 2005 16:58:16 +0100
In article <1128350949.605028.46950@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
spiney
<sparkistuf@xxxxxxxxxxxxx> wrote:
> A further reply, digital tv system perfromance, to Jim Lesurf etc ......
> On this topic, my original reply was to Jasmine, ("ber not strength"!),
> then to Kim on the "digital cliff". There most certainly is one,
> whatever Jim thinks.
....thus showing that you haven't actually understood what I (and others)
have written on that specific topic. :-)
> If you consult the usual textbooks, etc, on various digital modulations
> in presence of Gaussian noise, they will all show a steadily increasing
> b.e.r. as S/N decreases, until the point where the (Shannon) channel
> capacity is exceeded, where b.e.r. suddenly starts to rise vertically (1
> in 10,000 being the normally accepted "cutoff point"!).
As pointed out in previous postings (by others as well as myself) the above
sidesteps various factors which mean reality for DTTV isn't simply behaving
as the above. Thus one of the reasons why your descriptions are misleading.
FWIW if you look in one of the undergrad textbooks I wrote you will find a
chapter on the matter of error rates for gaussian noise, and the effects of
error detection and correction. (There is also a chapter on the work of
Shannon.) However although the same sort of thing appears in many texts,
this does not deal with the specific question at issue - which I have asked
about repeatedly [cf other postings to avoid more repetition].
The problem seems to be that what I and others have been saying seems
to go right past you without gaining your actual attention. Instead,
you just repeat the same misleading conclusions derived from over-
simplifying one factor and using it in an unrealistic context.
> Shannon didn't say anything about channel coding, and what actually
> happens depends on the error correction used. Usually, it will be some
> combination of convolution and cyclic coding, coping well with Gaussian
> noise, and to some degree with impulsive noise.
Apart from the vagueness of the "coping well" assertion, all fine *and
irrelevant* to the points we keep making to you... :-)
> Gaussian noise is useful, often allowing simple calculations, and can be
> "skewed" to approximate other distributions. It's the "general case",
> reasonably true in most circumstances. The Ricean noise sometimes quoted
> in cofdm articles comes from a particular reflection mode, which is not
> typical and may not be generally accurate.
Comment as above. :-)
> When discussing marginal signal performance we have to define what we
> mean, ie, is it signal level or ber, and under what conditions etc.
Since you are telling people they will in practice get a "perfect picture"
or "nothing", that is what your explanations would have to deal with. In
the context of a reality which includes varying levels of ignition
interference, fading, multipath, etc. The problem is that what you say is
misleading as it does not deal with the real situation.
> However, just putting a UHF variable attenuator in the aerial lead
> should convince most people there's a well defined signal level at which
> reception suddenly fails.
"suddenly fails" != "statistical rate of visible/audible problems"
> If Jim wishes to refer to technical documents, afraid I'm not
> telepathic, please give the exact link and point being made.
Nice try. :-) Alas, I've already just posted (again) the request that you
support your own assertions with a specific reference, and giving (again)
the context. So far, any telepathy here would have to be mine, trying to
read your mind and discover what one(s) you might quote.
Unlike your own postings, mine quote in the coventional manner. Hence
no telepathy is required. Just read my actual postings (and those of
others) and then make a replay that is based upon their actual
content. :-)
The problem is that I can't find any references that back up your claim in
terms of real-world behaviour. However I can find ones (e.g. BBC rd WHP056)
which indicate otherwise. Still waiting for you to give some *specific*
references I/we can check that support your claims...
> Unfortunately, the UK has a pisspoor record on digital tv (the OnDigital
> disaster etc). The BBC has a research dept for "status" purposes, but
> I'd like to know a single useful thing it's ever produced!
Another nice try. :-) I'm afraid, though, that a generalised attack on
the BBC research dept isn't likely to be regarded as a useful technical
reference for your assertions.
I recall Roy Hattersley saying "when asked a question you can't answer -
abuse the interviewer"... You seem to be his pupil... :-)
I must confess. though, that my personal bias would be to tend to hold the
BBC rd as being more reliable than yourself when it comes to the relevant
issues. Sorry if that seems unreasonable to you. The good news, is that
there may be other (reliable) sources of material. So...
> The (former) IBA also gave us the useless MAC (itself almost a direct
> copy of MUSE!), both together needlessly wasted much money on Eureka,
> etc. I'm not too impressed with these setups, no matter how many
> "papers" they might produce.
I suspect the feeling would be mutual. :-) However you still have not
given any specific references whose reliability and content we could use to
judge the basis for your claims. Perhaps you would like to also dismiss
some other bodies like the IEE and IEEE?...
> Note that cofdm is no better than ATSC, in fact rather worse! ATSC uses
> adaptive filter receiver techniques, a much better (and previously
> proved and tried) method than "guard bands". And cofdm isn't exactly
> "orthogonal" either - as the name might imply - since truncating the
> carriers obviously spreads out their spectra, 64qam doesn't help, and
> what you end up with is a much worse performing system (than the
> original claims).
Thus showing that you probably don't really understand orthogonal
modulation either, I'm afraid!
Maybe I'm wrong, spiney, but I have the feeling that the more you say
things like the above, the more likely it is that you are digging yourself
into a wider and deeper 'hole' so far as many experienced people here are
concerned. Too many hostages to fortune.
Based on your response, I still have the view that your claims are based
on over-simplifying one factors, and disregarding others. The result
being that what you say becomes misleading and inaccurate. My impression
is that I am not the only person here who regards your 'explanations'
as being like this. However if you can quote some specific sources which
people here would regard as reliable, then that might change. Alas, I
doubt that being dismissive about the BBC, etc, (or anyone who questions
what you claim) will be taken very seriously. Afraid I am not telepathic,
either, so can only judge you by what you say. Alas, when you pass
comments like some of those above, my reaction is to regard what you
say as being less and less reliable.
Slainte,
Jim
--
Electronics http://www.st-and.ac.uk/~www_pa/Scots_Guide/intro/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html
.
- References:
- Re: Strength or no errors?
- From: spiney
- Re: Strength or no errors?
- From: spiney
- Re: Strength or no errors?
- Prev by Date: Re: Strength or no errors?
- Next by Date: Re: frequencies of DTT muxes after switchover - do aerial riggers know?
- Previous by thread: Re: Strength or no errors?
- Next by thread: Re: Strength or no errors?
- Index(es):
Relevant Pages
|
