Ethan Winer Builds a Wire Null Tester

[restorer-john]

What about DSD stream / PWM signal ?
It is a digital signal but when fed directly into a speaker you hear analog sound.

DSD is not PWM, it's PDM and yes, it is a digitally stored but ultimately an analog system where absolute values are not calculated, just the density or lack thereof, of fixed amplitude pulses.

 

[solderdude]

You persist in confusing the data with the signal. The analog signal may just as well consist of information that represents the Klingon alphabet; it's still an analog signal. If we are to agree that it's a digital signal simply because it consists of an analog signal representing digital data, then Morse Code transmitted over the air via radio in 1930 is a digital signal via a digital transmission, simply because Morse Code can be described as a binary language (1).

We don't even have to agree with Petzold on that point either, as radio can be used to transmit digital data today, and was just as capable to have done so 100 years ago had [some digital data format] existed at that time. Now we would have to find ourselves proposing that shortwave is no longer an analog medium. And then we'd have to agree that Smoke Signals are no longer analog either because that is a binary language being transmitted.

It doesn't make sense to make such a claim.

More importantly, grasping the concept will help you to understand much better what is and what is not important in a digital transmission, and that includes what is desireable and what can fail in a digital cable, radio signal, or internal circuit layout and configuration. If you fail to grasp it, then you won't see where, in digital audio (or video) transmissions, the "bits are bits" crowd is wrong and why, and where they are correct and why

We are speaking of audio in this case. A digital audio signal can be 'PWM/PDM/DSD' which can be converted back to analog after conditioning and LPF or it can be PCM for instance when we talk about an AUDIO signal... you know ... music.

It should be quite clear that there is a difference between a digital signal (which transfers '1' and '0') and an 'analog' AUDIO/MUSIC 'signal'.
It does not matter if the audio contains a Klingon aria, morse code, a single tone or any other form of spoken word or music.

Your remark about the 'bits is bits crowd is wrong' says something as to why you are debating this.
In the subjective crowd there is a long standing belief that the actual waveform of the 'digital' signal has an influence on the actual analog (audio) signal.
Let's name a few: jitter, receivers having to 'work harder' to clean up the signal which influences ground planes and power supply lines, 'EMC' garbage riding along over ground and power lines running along datalines, ground loops (related), PCB designs (routing, ground planes), chips used, digital source quality, receiver configurations, galvanic separation, PLL, clock(s) used to name but a few.

And yes, these things do matter in some 'severity' or not but it does not change the fact that a changed waveform in the 'signal' has the same 'audible' effect which is often read in subjective reviews. Bass quality, treble extension and detailing being different between 'digital cables' etc is bollocks and can not be shown to exist in actual wave forms.

Sure... we can see ground loops, common mode signals entering gear, which is not affecting sound unless it somehow changes to differential mode and other 'nasties' everywhere.

Sure, in Amir's measurements we can see differences in measured performance between various inputs in certain DAC's but this is all due to implementations being far from optimal and most of the measured differences may not be audible, some may or will be.

It does not change the fact that when nulling an analog audio signal yields no differences between various cables (outside of hum being induced depending on the cable construction) yet the same 'hearing' crowd that can clearly hear differences between interlinks in 'analog' audio signals (which is not there as shown) also hear similar 'differences' when using 'digital' signals through different cables.

Let's just assume that a LOT of reported differences that are not clearly shown in measurements are well... maybe, just maybe, caused by something else than being of electrical origin. I base this on lack of 'hard evidence' on the contrary and personal experiences.

As said... we are talking of audio signals and their 'form' which is either 'digital' or 'analog'.

 

[solderdude]

Did you know Clive Sinclair married a lap-dancer (haven't we all) 36 years his junior, ditched her after 7 years and hooked up with his ex-secretary?

https://www.dailymail.co.uk/news/article-4371950/Clive-Sinclair-76-ditches-lap-dancer-wife.html

That may well explain why it does not work any more.
Is it known if he did he take her on bike rides in his 'blimps' ?

 

[solderdude]

DSD is not PWM, it's PDM and yes, it is a digitally stored but ultimately an analog system where absolute values are not calculated, just the density or lack thereof, of fixed amplitude pulses.

yep, that's why I mentioned both as DSD and PWM which both are digital 'streams' that can be converted to analog again using conditioning and LPF but are not the same but do something similar.

 

[Theo]

A digital signal will have some analog characteristics but remains a digital signal.

Referring to the definition (Oxford dictionary), I'd say this is not properly worded:
Signal : A gesture, action, or sound that is used to convey information or instructions, typically by prearrangement between the parties concerned.
So, I would propose that we say that the signal is the real (like physical) mean of transmitting information or codes (like virtual), including digits. Actually if there is no decoding of the signal, including by our brain, the signal is just physical or analog "noise". Its value is in the information it carries and which needs to be interpreted. Music, for example, is not a signal, it is information.
So, shouldn't we say that the digital we are talking about is the information, not the signal which remains analog, ie physical. "Digital", meaning based on mathematics (numbers), can't be real, so can't be physical, so not a signal. Numbers are information and need decoding on the receiver side, same as radio or language.

So, please, when speaking of digital, I would suggest that we use the word "information" or "code", but not signal...

 

[solderdude]

You persist in confusing the data with the signal. The analog signal may just as well consist of information that represents the Klingon alphabet; it's still an analog signal. If we are to agree that it's a digital signal simply because it consists of an analog signal representing digital data, then Morse Code transmitted over the air via radio in 1930 is a digital signal via a digital transmission, simply because Morse Code can be described as a binary language (1).

We don't even have to agree with Petzold on that point either, as radio can be used to transmit digital data today, and was just as capable to have done so 100 years ago had [some digital data format] existed at that time. Now we would have to find ourselves proposing that shortwave is no longer an analog medium. And then we'd have to agree that Smoke Signals are no longer analog either because that is a binary language being transmitted.

It doesn't make sense to make such a claim.

More importantly, grasping the concept will help you to understand much better what is and what is not important in a digital transmission, and that includes what is desireable and what can fail in a digital cable, radio signal, or internal circuit layout and configuration. If you fail to grasp it, then you won't see where, in digital audio (or video) transmissions, the "bits are bits" crowd is wrong and why, and where they are correct and why

... semantics, agreements. I'll rephrase ..

We are speaking of audio information (recorded music) in this case. A digital audio signal code can be 'PWM/PDM/DSD' which can be converted back to analog information after conditioning and LPF or it can be PCM for instance when we talk about an AUDIO signal information... you know ... music.

It should be quite clear that there is a difference between a digital signal code (which transfers '1' and '0') and an 'analog' AUDIO/MUSIC 'signal' information.
It does not matter if the audio information contains a Klingon aria, morse code, a single tone or any other form of spoken word or music.

Your remark about the 'bits is bits crowd is wrong' says something as to why you are debating this.
In the subjective crowd there is a long standing belief that the actual waveform of the 'digital' signal code has an influence on the actual analog (audio) signal information.
Let's name a few: jitter, receivers having to 'work harder' to clean up the signal code which influences ground planes and power supply lines, 'EMC' garbage riding along over ground and power lines running along datalines, ground loops (related), PCB designs (routing, ground planes), chips used, digital source quality, receiver configurations, galvanic separation, PLL, clock(s) used to name but a few.

And yes, these things do matter in some 'severity' or not but it does not change the fact that a changed waveform in the 'signal' code has the same 'audible' effect which is often read in subjective reviews. Bass quality, treble extension and detailing being different between 'digital cables' etc is bollocks and can not be shown to exist in actual analog information wave forms.

Sure... we can see ground loops, common mode signals entering gear, which is not affecting sound unless it somehow changes to differential mode and other 'nasties' everywhere.

Sure... in Amir's measurements we can see differences in measured performance between various digital (code) inputs in certain DAC's but this is all due to implementations being far from optimal and most of the measured differences may not be audible, some may or will be.

It does not change the fact that when nulling an analog audio signal information yields no differences between various cables (outside of hum being induced depending on the cable construction) yet the same 'hearing' crowd that can clearly hear differences between interlinks in 'analog audio signals' information (which is not there as shown) also hear similar 'differences' when using 'digital signals' code through different cables.

Let's just assume that a LOT of reported differences that are not clearly shown in measurements are well... maybe, just maybe, caused by something else than being of electrical origin. I base this on lack of 'hard evidence' on the contrary and personal experiences.

As said... we are talking of audio/music signals information and their 'form' which is either 'digital / code' or 'analog / information'.

 

[svart-hvitt]

We are speaking of audio in this case. A digital audio signal can be 'PWM/PDM/DSD' which can be converted back to analog after conditioning and LPF or it can be PCM for instance when we talk about an AUDIO signal... you know ... music.

It should be quite clear that there is a difference between a digital signal (which transfers '1' and '0') and an 'analog' AUDIO/MUSIC 'signal'.
It does not matter if the audio contains a Klingon aria, morse code, a single tone or any other form of spoken word or music.

Your remark about the 'bits is bits crowd is wrong' says something as to why you are debating this.
In the subjective crowd there is a long standing belief that the actual waveform of the 'digital' signal has an influence on the actual analog (audio) signal.
Let's name a few: jitter, receivers having to 'work harder' to clean up the signal which influences ground planes and power supply lines, 'EMC' garbage riding along over ground and power lines running along datalines, ground loops (related), PCB designs (routing, ground planes), chips used, digital source quality, receiver configurations, galvanic separation, PLL, clock(s) used to name but a few.

And yes, these things do matter in some 'severity' or not but it does not change the fact that a changed waveform in the 'signal' has the same 'audible' effect which is often read in subjective reviews. Bass quality, treble extension and detailing being different between 'digital cables' etc is bollocks and can not be shown to exist in actual wave forms.

Sure... we can see ground loops, common mode signals entering gear, which is not affecting sound unless it somehow changes to differential mode and other 'nasties' everywhere.

Sure, in Amir's measurements we can see differences in measured performance between various inputs in certain DAC's but this is all due to implementations being far from optimal and most of the measured differences may not be audible, some may or will be.

It does not change the fact that when nulling an analog audio signal yields no differences between various cables (outside of hum being induced depending on the cable construction) yet the same 'hearing' crowd that can clearly hear differences between interlinks in 'analog' audio signals (which is not there as shown) also hear similar 'differences' when using 'digital' signals through different cables.

Let's just assume that a LOT of reported differences that are not clearly shown in measurements are well... maybe, just maybe, caused by something else than being of electrical origin. I base this on lack of 'hard evidence' on the contrary and personal experiences.

As said... we are talking of audio signals and their 'form' which is either 'digital' or 'analog'.

Do you disagree with Jim Lesurf’s text (see link above)?

 

[solderdude]

What he states is that by looking at signals we can't be 100% sure a waveform is analog information or digital code.
In certain cases that may be very true.
For instance when looking at an 'analog' squarewave generator signal information (which is an analog signal information) yet looks like a digital signal code.
Also when connecting a line level source to a mic input. The audio signal information looks a lot like a digital signal code.
We can look at a digital signal code that is being fed through a bandfilter and it may look like an analog waveform information when distorted enough.

So yes, he has a point. However, when you listen to those waveforms it will quickly dawn on you what the 'signal' you are dealing with is.
And... I might add, having seen enough clipped, distorted digital and analog waveforms, and looking at parts used on PCB's, I can make a fairly educated guess, leaning towards certainty, whether or not a signal/waveform is analog information or digital code but of course, cannot tell what the information exactly is untill it is made audible. Code (unless PWM/PDM) will not sound anything like audio information.
Both visual and audio information will give a lot of certainty (dare say 99.99%) whether or not a signal is 'digital' code (or modulation) or 'analog' not coded / modulated audio information.

So I agree and disagree with certain points but that is true for a lot of thesis and 'info' out there.
I also believe that most readers/participants of this thread are fully aware of what a 'digital' and 'analog' signal is.

The Wiki definition linked to by @danadam is a much clearer definition when it comes to audio related signals whether they are 'code' or 'information'.

An analog signal is any continuous signal for which the time-varying feature (variable) of the signal is a representation of some other time varying quantity, i.e., analogous to another time varying signal. For example, in an analog audio signal, the instantaneous voltage of the signal varies continuously with the pressure of the sound waves.

 

[mansr]

If we are to agree that it's a digital signal simply because it consists of an analog signal representing digital data, then Morse Code transmitted over the air via radio in 1930 is a digital signal via a digital transmission, simply because Morse Code can be described as a binary language (1).

Morse code is absolutely digital, consisting of pulses classified by the receiver as short or long. The carrier medium is irrelevant.

And then we'd have to agree that Smoke Signals are no longer analog either because that is a binary language being transmitted.

Quite so.

 

[svart-hvitt]

My favorite Jim Lesurf article on analog vs digital:

https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/iandm/part12/page2.html

He summarizes (again) like this:

«You should now understand that the terms ‘analog’ and ‘digital’ are based on idealisations. Real systems and signals will show a mixture of analog (smooth continuous) and digital (quantised) properties. Although it's often convenient to assume a signal/system is one thing or the other, this mixed behaviour is an unavoidable consequence of the way the world works».

Lastly, a background link on Lesurf:

https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/MMWave/Jim.html

 

[solderdude]

He is talking about vinyl.

The above analysis is a simplified one. It leaves out many features of a practical LP system. Despite that, it does serve to show that even a system which appears essentially ‘analog’ will still have underlying properties similar to a digital information processing system.

He claims that LP is basically limited because of atom size and calculates 110dB dynamic range !
Then he goes on about vinyl molecules (which are a LOT bigger in size drastically reducing the dynamic range) but could just as well be electrical noise from amps used or magnetic particles (tape).

For a recorded signal this produces an effect similar to dithering a signal before digital sampling.

I call this 'background noise' and would not go as far as stating this is a 'bit' digital. But ... each to his own. Personally I find the conclusion a bit misleading but do understand the Philosophical point he is trying to make.

All of this has little to do with a nulling device though.

 

[svart-hvitt]

He is talking about vinyl.



He claims that LP is basically limited because of atom size and calculates 110dB dynamic range !
Then he goes on about vinyl molecules (which are a LOT bigger in size) but could just as well be electrical noise from amps used or magnetic particles (tape).


I call this 'background noise' and would not go as far as stating this is a 'bit' digital. But ... each to his own. Personally I find the conclusion a bit misleading but do understand the Philosophical point he is trying to make.

Lesurf is an experienced teacher of physics and electronics, and is a winner of the National Physical Laboratory Prize, twice.

I like the way he writes. To me, it gives insight.

 

[solderdude]

He has quite a few accomplishments !

I like to explain things as simple as possible.
His thesis:... vinyl is analog with a bit of 'digital dither' in the form of vinyl atoms and grains ?

To me vinyl is a (quite flawed) audio medium that sounds surprisingly good despite its shortcomings, mostly due to the 'forgiving' nature of our hearing and pleasant-ness of the added distortions.
I can see people preferring the analog medium, with a bit of digital in it, over digital with a bit of analog in it.

 

[svart-hvitt]

He has quite a few acomplishments !

I like to explain things as simple as possible.
His thesis:... vinyl is analog with a bit of 'digital dither' in the form of vinyl atoms and grains ?

To me vinyl is a (quite flawed) audio medium that sounds surprisingly good despite its shortcomings, mostly due to the 'forgiving' nature of our hearing and pleasant-ness of the added distortions.
I can see people preferring the analog medium, with a bit of digital in it, over digital with a bit of analog in it.

I think his vinyl article should be read as an attempt at explaining to vinyl lovers what could be achieved by that format in über ideal conditions. And then you end up - ridicilously theoretically - with something which is well below 24 bits*.

So the vinyl lovers hate him for pointing that out.

I know, because I’ve done that


*And he uses his mastering of both the digital and the analog domain to build up that point.

 

[DonH56]

Good grief, such a pissing contest over what is digital? I'm thinking "What is digital?" and he washes his hands of it...

Digital signals happen when you take an analog signal and apply a thresholding or windowing (comparison) operation to it. I have found the definition of "digital" depends upon who does the talking. The digital designers at my company consider pretty much every signal in the path digital, including the clocks and Rx/Tx signals to/from the SerDes, digital signals. I do not. Nor do they when the signal is captured wrongly and all of a sudden the analog team is brought in to debug their digital signals. But it has been many years since I dealt with RTL/VHDL/Verilog code so make no claim to their expertise. We need it all to make our products and keep the world going 'round. And our teams work together well.

I had in mind going into more detail, but this is obviously another of those "religious" threads where nobody is going to change anybody's mind so am disinclined.

 

[svart-hvitt]

Good grief, such a pissing contest over what is digital? I'm thinking "What is digital?" and he washes his hands of it...

Digital signals happen when you take an analog signal and apply a thresholding or windowing (comparison) operation to it. I have found the definition of "digital" depends upon who does the talking. The digital designers at my company consider pretty much every signal in the path digital, including the clocks and Rx/Tx signals to/from the SerDes, digital signals. I do not. Nor do they when the signal is captured wrongly and all of a sudden the analog team is brought in to debug their digital signals. But it has been many years since I dealt with RTL/VHDL/Verilog code so make no claim to their expertise. We need it all to make our products and keep the world going 'round. And our teams work together well.

I had in mind going into more detail, but this is obviously another of those "religious" threads where nobody is going to change anybody's mind so am disinclined.

I think my «enthusiasm» on this matter reflects how little I know - and of how even less I knew - on this matter... Let me explain:

When I first read Lesurf it was some sort of reveleation, at least eye-opening to me. It got me thinking.

I think this is what happens when myths and magical thinking are replaced by more well-thought out ideas.

I enjoy very much such episodes when my thinking changes on a matter. You also need to take into account that I have no physics or engineering background.

Just a personal reflection on my «enthusiasm» here.

 

[bennetng]

Adding fuel to the fire on this thread:

 

[solderdude]

I have found the definition of "digital" depends upon who does the talking.

For the noobs (not being Don and a few others )

Yes, semantics. For me the differences are clear..
A to D conversions happens in the AD converter chip(s) and in the DA conversion chip the digital signal is converted back to analog.
Everything in between the A-D and D-A chip is 'digital' in nature and could be called 'data' or 'code' if you will.
For me code means something else.. it means 'a machine code program' to me so data for me.

In the 'digital' signal there is a lot of 'analog' happening such as 'cleaning' up signals and the PLL part.
So yes, there is analog happening in the digital realm to make it work.
The clock generator is analog (it is a neat sine wave) which is basically 'clipped' so it becomes more square-wave alike.
That clock is usually divided a couple of times as well.
Even though the clock is digital in nature and needed (essential) for everything to work I don't regard it as a digital signal. Some will though.
I would call the SPDIF signal 'digital' and consists of data + clock so transmissions are easily decoded and only 1 (signal) wire is needed.
Also I would call the L-R clock, and data digital.
A clock is a constant frequency and is not 'modulated'. It always has a 50% duty-cycle (well... very close to it) so is not a 'digital signal' to me.
L-R clock, SPDIF and data all have a varying signal and is what I consider a 'digital signal' as it contains data about the 'enclosed information'.
The clock has no such information (aside from the speed)

Digital signals can be actual '0's and '1's but can also be just indicating transitions from 0 to 1 or 1 to 0. For the latter a synchronized clock is essential.
That clock tells you how many consecutive 0's (or 1's) are there after each other. Regardless if they are Eight to Fourteen coded or not. Even if the 'digital data' were transmitted via 2 sine-waves of different frequencies which would be an analog transmission the signal itself still remains 'digital' in nature as it contains 'data' which in turn contains 'information.

Curious to hear what others find to be 'analog' or 'digital' and above all... why.

* Perhaps a mod could create another thread with all this non 'wire tester' related blabla or move that to another thread if it exists. *

The eye-pattern in a CD player (as described in the video) he considers 'analog' and the laser photo diodes pre-amp he describes as an ADC which it isn't.
Well.. the data coming of the eye-pattern is 'digital' but the amplitude is used in an 'analog' way.
The laser tracking and focusing looks at the analog waveform (amplitude) and adjusts the tracking to keep it at 'maximum' output.
The disc speed is controlled by adjusting the speed so that the 'smallest' detected 'sine-wave' is 3 clock pulses 'wide'
The eye-pattern is made up of a series of 'transitions' from 0 to 1 or 1 to 0. The minimum width of a 'land' or 'pit' is 3T (3 clock-pulses) and maximum 11T (11 clock-pulses) and can be any width in between (in steps of 1 clock pulse width)
The pits and lands thus are not '1's and '0's as in the actual 'digital audio data'. That data (which contains the analog 'information') is enclosed in other data. This is needed to ensure small 'errors' on the disc or blemishes do not cause audio data to be corrupted.
Eight to Fourteen modulation, interleaving and error correction as well as bit frames and other essential data is enclosed in the eye-pattern.
All this I call digital signals. The focusing and tracking is 'analog' to me.

 

[svart-hvitt]

For the noobs (not being Don and a few others )

Yes, semantics. For me the differences are clear..
A to D conversions happens in the AD converter chip(s) and in the DA conversion chip the digital signal is converted back to analog.
Everything in between the A-D and D-A chip is 'digital' in nature and could be called 'data' or 'code' if you will.
For me code means something else.. it means 'a machine code program' to me so data for me.

In the 'digital' signal there is a lot of 'analog' happening such as 'cleaning' up signals and the PLL part.
So yes, there is analog happening in the digital realm to make it work.
The clock generator is analog (it is a neat sine wave) which is basically 'clipped' so it becomes more square-wave alike.
That clock is usually divided a couple of times as well.
Even though the clock is digital in nature and needed (essential) for everything to work I don't regard it as a digital signal. Some will though.
I would call the SPDIF signal 'digital' and consists of data + clock so transmissions are easily decoded and only 1 (signal) wire is needed.
Also I would call the L-R clock, and data digital.
A clock is a constant frequency and is not 'modulated'. It always has a 50% dutycycle (well... very close to it) so is not a 'digital signal' to me.
L-R clock, SPDIF and data all have a varying signal and is what I consider a 'digital signal' as it contains info about the 'enclosed information'.
The clock has no such information (aside from the speed)

Digital signals can be actual '0's and '1's but can also be just indicating transitions from 0 to 1 or 1 to 0. For the latter a synchronised clock is essential.
That clock tells you how many consecutive 0's (or 1's) are there after each other. Regardless if they are Eight to Forteen coded or not. Even if the 'digital data' were transmitted via 2 sinewaves of different frequencies which would be an analog transmission the signal itself still remains 'digital' in nature as it contains 'data' which in turn contains 'information.

Curious to hear what others find to be 'analog' or 'digital' and above all... why.

* Perhaps a mod could create another thread with all this non 'wire tester' related blabla or move that to another thread if it exists. *

The eye-pattern in a CD player (as described in the video) he considers 'analog' and the laser pre-amp he describes as an ADC.
Well.. the data coming of the eyepattern is 'digital' but the amplitude is used in an 'analog' way.
The laser tracking looks at the analog waveform (amplitude) and adjusts the tracking to keep it 'maximum' in output.
The disc speed is controlled by adjusting the speed so that the 'smallest' detected 'sinewave' is 3 clock pulses 'wide'
The eyepattern is made up of a series of 'transitions' from 0 to 1 or 1 to 0. The minimum width of a 'land' or 'pit' is 3T (3 clockpulses) and maximum 11T (11 clockpulses) and can be any width in between (in steps of 1 closck pulse width)
The pits and lands thus are not '1's and '0's as in the actual 'digital audio data'. That data (which contains the analog 'information') is enclosed in other data. This is needed to ensure small 'errors' on the disc or blemishes do not cause audio data to be corrupted.
Eight to Forteen modulation, interleaving and error correction as well as bit frames and other essential data is enclosed in the eye-pattern.
All this I call digital signals.

Digital vs analog...I always liked this one:

It all depends on the encoder, doesn’t it?

 

[bennetng]

It all depends on the encoder, doesn’t it?

Interesting no one call that chip by the official name (Photo diode signal processor for compact disc players) and call it laser pre-amp or ADC. Clearly people are giving definition of things on their own.

A signal is a signal.

Undeniable fact.