Questions about equivalents of audio/video and digital/analog.

[Jerry Avins]

[email protected] (Don Pearce) wrote:
...


Bullshit Don, that is abjectly stupid to claim.


Nyquist rate:
The reciprocal of the Nyquist interval, i.e., the
minimum theoretical sampling rate that fully
describes a given signal, i.e., enables its
faithful reconstruction from the samples. Note:
The actual sampling rate required to reconstruct
the original signal will be somewhat higher than
the Nyquist rate, because of quantization errors
introduced by the sampling process.

So, quantization error is the reason one has to sample faster than the
Nyquist rate! If you believe that, there's a bridge I want to sell you.

...

Jerry

 

[Floyd L. Davidson]

Hmmm... says "Nyquist rate" right there.

Hmmm... so that's exactly what I looked at.

Actually I was talking about this:

You don't say? Why didn't you mention before what you
were talkiing about?

Nyquist's theorem: A theorem, developed by H. Nyquist, which states
that an analog signal waveform may be uniquely reconstructed, without
error, from samples taken at equal time intervals. The sampling rate
must be equal to, or greater than, twice the highest frequency
component in the analog signal. Synonym sampling theorem.

Go ahead and defend.

Claude Shannon proved it mathematically in the late
1940's.

You are free to dispute Shannon...

 

[Jerry Avins]

Arny Krueger wrote:

...

Yet another *eggspurt* who has never looked at the output of a good DAC with
a scope.

Scope? SCOPE? isn't that an analog device? True digital work is don with
pencil, paper, and calculator. Theory rules!

Jerry

 

[Floyd L. Davidson]

Bullshit. Unless of course you are claiming that all digital signals
have perfectly flat tops and vertical edges. Is that your claim,
Floyd? You are no longer waving, but drowning. Just admit you have
this all wrong and bow out with as much grace as you can manage.

Phase may not affect your ears much, but it does affect
the data. It has nothing to do with flat tops and
vertical edges, which seem to be something you can't get
around.

All of a sudden he has no answer. Come on Floyd - it is your claim
that an amplifier is a D to A converter. Now put up or shut up. Defend
it or admit you have screwed up.

My point is that there is a lot more to a practical
amplifier than just simple gain. Sorry that you missed
the point, again.

 

[Floyd L. Davidson]

Actually, they all do.

However, if the upper limit is more than 2 or 3 times
higher than the frequency spectrum of the signal, in
effect they don't.

The point is we are *not* talking about wideband
amplifiers and narrow band signals, we are talking about
practical applications where the two are very closely
matched.

Usually, its pretty simple.

Sure it is... That's why there are complete books
published on what happens?

Nope, the output of good DAC is generally even more severely band-limited

Limited by the sampling rate.

than that of a good amplifier. That's why they call the filtering in DACs
"brick wall filters".

That is to prevent aliasing though, and is actually a
higher frequency cutoff than what is required to affect
the resulting output. And they are used on the *input*,
not on the output.

 

[Floyd L. Davidson]

Whatever good DAs put out, it is good enough analog to fool my scope!

Your scope tells you if it is digital or analog information
on the display???? Wow, where do you find one like that?

Hint: you don't, because it can't.

Yet another *eggspurt* who has never looked at the output of a good DAC with
a scope.

Toss your old tube type Heathkit scope in the dumpster.
Get a good one, and sync it to the sampling clock.

Regardless, the idea that you think looking at a signal
with a scope will tell you if it is digital or analog
suggests that you weren't paying attention when you
studied basic signal theory, if you ever did.

 

[Andor]

Floyd L. Davidson wrote: ....

So, quantization error is the reason one has to sample faster than the
Nyquist rate! If you believe that, there's a bridge I want to sell you.

Indeed.

On another note, I just read Shannon's "Communication in the presence
of noise" paper the other day. He does call the critial interval
between two sampling instants the Nyquist interval:

"Nyquist pointed out the fundamental importance of the time interval
1/2 W seconds in connection with telegraphy, and we will call this the
Nyquist interval corresponding to the band W."

This would mean that the Nyquist frequency, being the inverse of the
Nyquist interval, is the sampling frequency. However, I've never seen
the name used that way. Typically, the Nyquist frequency denotes
_half_ the sampling frequency. If anything, nomenclature will be the
end of this newsgroup.

Regards,
Andor

 

[Floyd L. Davidson]

So, quantization error is the reason one has to sample
faster than the Nyquist rate! If you believe that,
there's a bridge I want to sell you.

If you don't understand what they said, you probably do
have a bridge that somebody sold you...

 

[Floyd L. Davidson]

Arny Krueger wrote:

...


Scope? SCOPE? isn't that an analog device? True digital
work is don with pencil, paper, and calculator. Theory
rules!

And you, like Arny, probably have no idea what you'd
see on a decent scope anyway.

I'd like Arny to explain how he can look at a scope and
tell if a single cycle of a sine wave is an analog signal
representing one cylce of a pure tone, or is just a digital
signal that represents 8000 different bytes of data from
a digital image.

 

[Bob Myers]

Analog signals are by *definition* continous.

Nonsense. You don't believe it is possible to sample
an "analog" signal and have it remain analog?

You have misunderstood what that means though. The
analog value of a signal is continuous,

Well, THAT certainly makes it clearer....

Since you seem to be so hung up on definitions, Floyd,
try this one on for size:

Continuous: unchanged or uninterrupted: continuing
without changing, stopping, or being interrupted in space
or time.

Note that this does not say anything at all about the range
of possible values being "continuous" (which is what you
seem to be trying to say in the above).

Bob M.

 

[Arny Krueger]

Your scope tells you if it is digital or analog information
on the display???? Wow, where do you find one like that?

Hint: you don't, because it can't.

Toss your old tube type Heathkit scope in the dumpster.

Never had one.

Get a good one, and sync it to the sampling clock.

Been there, done that.

Regardless, the idea that you think looking at a signal
with a scope will tell you if it is digital or analog
suggests that you weren't paying attention when you
studied basic signal theory, if you ever did.

I aced "Signals And Systems Analysis" and placed high in "Linear Stochastic
Optimal Control" - probably the heaviest signals trips I ever took.

 

[Bob Myers]

There needs to be a way to describe sampled but not quantized
signals. They are not continuous in time, but the function
can take on any value at each sample point.

Such signals are simply referred to as "sampled."
Whether the data they carry is encoded in "digital"
or "analog" form is a separate question.

The value that any symbol has (to use the proper
communications or information theory term) is
ALWAYS limited in accuracy/resolution, regardless
of the encoding system used. There is no such thing
as an information encoding system or communications
channel which provides "infinite" accuracy/resolution,
therefore to say that a signal can take on "any"
value becomes to a large extent meaningless.

Bob M.

 

[Bob Myers]

Sheesh! That *is*, by definition a digital signal.

Really? Suppose I show you an oscilloscope screen which
is displaying a single line of video, which happens to be
carrying an 8-level gray-scale pattern. It clearly shows
a set of discrete levels. Further, since this video happened
to be created by a D/A converter with only three bits at
the input (our video generator was built on the cheap!), those
are the ONLY levels this signal may exhibit. Is this a
"digital" signal?

Bullshit son. Look it up. I've provided you with
quotes from an authoritative reference, twice now. You
don't have to take my word for it, that *is* the agreed
technical definition of the term.

Ah, Floyd - argument from authority again, huh?

Bob M.

 

[Arny Krueger]

However, if the upper limit is more than 2 or 3 times
higher than the frequency spectrum of the signal, in
effect they don't.

Nope, the phase shift is appreciable down to Fc/10. Ever hear of Bode plots?

The point is we are *not* talking about wideband
amplifiers and narrow band signals, we are talking about
practical applications where the two are very closely
matched.

Most digital audio is brick walled at 0.95 * Fs/2 or about 21 KHz. Most
power amps are pretty flat up to about 50 KHz.

Sure it is... That's why there are complete books
published on what happens?

For people who don't already know?

Limited by the sampling rate.

Well, by the brick wall filter which is usually set as stated above.

That is to prevent aliasing though,

DACs can't alias. Only ADCs can alias. Improperly filtered DACs may produce
images.

and is actually a
higher frequency cutoff than what is required to affect
the resulting output. And they are used on the *input*,
not on the output.

The input side of a digital system is called an ADC, not a DAC as you just
said.

 

[Arny Krueger]

And you, like Arny, probably have no idea what you'd
see on a decent scope anyway.

(1) A decent scope gives a pretty close approximation of what theory
predicts.

(2) I was probably working with decent scopes before you were born.

I'd like Arny to explain how he can look at a scope and
tell if a single cycle of a sine wave is an analog signal
representing one cylce of a pure tone, or is just a digital
signal that represents 8000 different bytes of data from
a digital image.

I've seen both kinds of data many times. Imaging data almost never looks
like sine waves.

 

[Bob Myers]

It wasn't analog until you ran it through an analog amplifier.

More nonsense.

Again, the ONLY definitions of "analog" and "digital"
which make any sense treat these as distinctions in
the form of information encoding being used. If I
run EITHER a "digital" or "analog" signal through an
amplifier, what comes out can still be interpreted (the
information recovered from the signal) ONLY if the
encoding intent is understood and the proper decoding
applied.

A serial stream of digital data still makes sense, whether
the amplitude assigned to the "1" or "true" state is
0.1V, 1V, 10V, or 100 kV. But it makes sense
ONLY when interpreted AS a serial stream of bits.

Similarly, an analog representation of, say, video makes
sense only if interpreted AS "analog". No matter how
"digital-ish" it might look, if you try to interpret THIS
signal as a "digital" stream, you'll get gibberish.

Floyd, you would be well advised to stop treating your
"definitions" as though they were somehow handed
down by God, and instead try to employ arguments
that are actually based in something sensible.

The NTIA is an authority, and MilStd specifications are
also authoritative. That is the reason I cited them.
And the *fact* is that you have not and cannot cite any
authoritative standards body that does not agree with
them.

Neither of these - and for that matter, NO standards body
- is an Infallible Source of Absolute Truth, and no standard
should be looked at as a substitute for good ol' basic
theory and experimentation. This is the fundamental flaw
with any argument "from authority": wrong is wrong, no matter
who writes it down on a piece of paper. God knows I've
spent way more than enough time in my career working
with various standards organizations (in fact, I am currently
chairing one fairly well-known such group), including both
"industry" and "government" efforts, and I can tell you from
long and painful experience that simply because something
appears in a standards document does not make it correct.
With the right people paying close attention, these documents
can often turn out pretty darn good - but they should NEVER
EVER be used as a substitute for some actual thought and
understanding of the subject matter at hand.

Bob M.

 

[Arny Krueger]

Really? Suppose I show you an oscilloscope screen which
is displaying a single line of video, which happens to be
carrying an 8-level gray-scale pattern. It clearly shows
a set of discrete levels.

Well it shows what was once a set of discrete levels. Since it is now in the
analog domain, there will be rise time, overshoot, tilt, simple inaccuracy,
and etc. .

Further, since this video happened
to be created by a D/A converter with only three bits at
the input (our video generator was built on the cheap!), those
are the ONLY levels this signal may exhibit. Is this a
"digital" signal?

It's an analog signal that represents something that was once quantized.

The definitions are fine, it is the misapplication of them that sticks.

 

[Bob Myers]

It's best to call that a quasi-analog signal...

Why? What does that mean, EXACTLY, that
isn't already conveyed (and conveyed more accurately)
by other, more appropriate terms? What additional
information does this "quasi-analog" nonsense bring
to the party?

Bob M.

 

[Bob Myers]

It is by definition. You can stand there and deny it all
you like, but that just makes *you* look damned silly.

I've cited impeccable and authoritative sources. You have not
and cannnot cite anything that supports your opinion.

Well, no, he's cited nothing but evidence and reason.
I suppose those are ignorable to anyone who treats
mere "definitions" as Holy Writ, but the rest of us somehow
find them helpful.

Again, I can cite equally acceptable and peer-reviewed
sources which use EXACTLY the "definitions" and
interpretation of these terms that Don and I and others
are using here. Do you really think this sort of think should
be decided by "battling authorities"?

Tell you what - you present an argument that ISN'T
based on an appeal to an authority, one which shows how
your notions make rational sense, and I will in turn show
you said "authorities" on the other side.

Bob M.

 

[Don Bowey]

Oh, perhaps 1950, maybe a few as late as 1960 or so...

Do you know of any telco that actually uses them today?


Of course I suppose it is possible they are still being
used where *you* live. But I don't know of any telco in
all of Alaska that has used an E repeater in the last 30-40
years. In particular, in the last 10-20 years that would
be totally unacceptable.

I didn't leave my telco job until the end of 94. At which times they were
still in use, but there was talk of using gain within the switching
machines. It wouldn't surprise me if that is being done now, being a simple
process. In any case, there are loops that require gain to meet minimum
requirements. Also, we had a tariff that provided additional gain (for a
price) where feasible.

My concerns were not just for where "I lived." I was on the Transmission
Engineering staff, and we had 14 states with which to be concerned.