Analogue playback Wander

[amirm]

Again, think about why close-in phase noise is called flicker noise - it's thought to be the result of some event at the quantum level although this is not confirmed!

No, no, no. We are not driving the output of a DAC with a quartz crystal. We are driving it with an oscillator where the crystal is a component. ALL the components contribute to the final phase noise. Here is one of countless examples of oscillators, this time using a discrete circuit:

All of these components work together and each has its own noise and profile. It is not just the quartz crystal that you keep mentioning. Did you not follow what I explained regarding CLT and PDF???

There are tons of papers online that explains this. Please download and read them before throwing out these terms.

 

[amirm]

Right, I recognise the Audiophileo from your description - right?

Good detective work. You know your competitors well.

If so then the clock phase noise as posted on their website isn't considered low close-in phase noise
View attachment 32476

I have post the right graph and explanation in ASR Forum. I won't repeat it here: http://www.audiosciencereview.com/forum/index.php?threads/close-in-jitter.1621/page-10#post-41230

Net, net, I did indeed test the effect of some 40 db of close in/low-frequency jitter and it was not audible to me in blind tests. Just like the fact that analog jitter similar range is not audible to people.

This is adding random jitter - the most benign as far as audibility is concerned.

Now you got it. Random bandlimited jitter is even less audible due to strong masking. See critical bands.

 

[FrantzM]

You still don't get it. Maybe an example:

Take 10 clock ticks. Each number below is the time the tick occurs relative to the "correct" time. A positive number means the tick is late (clock slow) and vice versa for negative.
0 -1 +1 -2 0 +3 -2 0 +1 0
Phase noise / jitter, correct? And corresponding to relatively phase noise frequencies.
Now the same noise, but with a low frequency component:
0 0 +3 +0 +3 +6 +0 +2 +2 0
It's not obvious from that. or indeed from a waterfall display such as you posted a picture of, but there is half a cycle of LF noise in there:
0 +1 +2 +2 +3 +3 +2 +2 +1 0
In frequency terms, the clock slows down then returns to the correct speed over a 10 sample period.

That is exactly what happens in a jittery clock. And in fact it happens more often than not, because as the phase noise plots show, longer-term variations in clock timing occur more frequently than short term variations. So there is indeed a connection between the mistimings. They can indeed move in one direction.The most extreme example of this is where all the ticks come consistently early or late (the clock is running fast or slow.) It's exactly the same effect as turntable wow. The "pitch control" analogy applies exactly.

I hope I've made it clear enough for "those following along at home", so I'm done here unless anyone else has any questions or comments.

You have but.... see my post somewhere in this thread ....

 

[jkeny]

No. Just no. That scope is sampling at 5 Ghz. That is 5,000,000 Khz. Our audio bandwidth on a good day may be 50 Khz. Bandwidth limit that scope to 50 Khz and then come and show that kind of jitter.

Oh dear, it seems you can't follow what's being discussed?
Don was claiming that clock timing moved as an interconnected group of timings "In frequency terms, the clock slows down then returns to the correct speed over a 10 sample period."
I posted that graph to show clock timings & how clock timings don't move in this manner.
Your objection is totally wrong - what Don & I were talking about it was about the clock tim

Answer me this - do you agree with Don's assertion above?

Audio circuits do not have such slew rates as I explained earlier. The timing of the output of the DAC cannot and will not change instantaneously. If it did, it would be a step function and that would show up as odd harmonics of that step with bessel frequency coefficients.

This is why I keep saying you MUST measure the output of the DAC, not what goes into it. The turntable measurements I post are the final output of what comes out of the turntable. No one is showing you the component distributions to it which may indeed have step functions like you imagine for digital (but again, slew limited due to inertia of the mechanical system).

Let me repeat: the turntable measurements I post where for the final output of the system. You can't keep comparing that to what goes into the DAC. And with the wrong instrumentation with massive bandwidth to boot.

The J-test measurements posted earlier were measurements at the output of the DAC - there are plenty of examples of these.

And the point of what is seen on the widened skirt at the base of the signal spike & rising with increasing slope to meet the signal spike is that it is not due to the same underlying causes as analogue wow - Don tried to assert that was the case in so far as he asserted that a sequential grouping of clock ticks slipped in time in an interconnected manner to provide the equivalent of a pitch change & therefore this was the equivalent of analogue wow

I'll ask you again - do you agree with his contention about such behaviour of the clock timings?

 

[jkeny]

No, no, no. We are not driving the output of a DAC with a quartz crystal. We are driving it with an oscillator where the crystal is a component. ALL the components contribute to the final phase noise. Here is one of countless examples of oscillators, this time using a discrete circuit:

All of these components work together and each has its own noise and profile. It is not just the quartz crystal that you keep mentioning. Did you not follow what I explained regarding CLT and PDF???

There are tons of papers online that explains this. Please download and read them before throwing out these terms.

Your post has nothing to do with what I said & does not disagree with anything I've said about close-in phase noise which is predominantly due to the crystal itself, its cut, polish etc.

You are missing basic understanding & throwing anything to obfuscate, hoping something will stick
It won't because what we are talking about is close-in phase noise of clocks & analogue playback wander

 

[amirm]

I'll ask you again - do you agree with his contention about such behaviour of the clock timings?

I am not following Don's argument. I am following yours and they continue to show errors. Bigly.

I will sum up your position and mine:
1. You are relying on what you think you have heard in sighted tests with strong bias ("good clock must sound better"). Even if you changed nothing, you would be hearing such improvements. So there is no basis here in scientific investigation.

2. You are then working backward, convincing yourself that there must be a difference here between digital and analog. Everything you say are your words, not presented in any paper, research, references, etc. Outside of that you are ignoring the fact that analog jitter is tons, tons higher than digital. So even if your argument was right, which it is not, you still have no explanation for why heaps of speed errors in analog at low frequencies is inaudible to people

3. Psychoacoustics completely backs what I explain. None of yours follows the same.

Bottom line is this: you can believe *subjectively* that better clock sounds better to you. And other people for that matter. That is perfectly fine. It is no different than folks claiming improvements for much lesser devices.

What you can't do is say audio science backs any of this. No audio science will accept your sighted, biased tests as the basis of anything. So you need to dispense with all of this technical talk. You can't be a part-time vegetarian.

If it is too uncomfortable for you to distance yourself from audio science, then you need to take audio science seriously. Start with proper testing protocols. Present that data. Let us repeat it. And then we can go on.

Alternatively adopt subjectivity fully and be done with it. Talk about free radicals this, and molecule alignment that, and you will be golden. Same number of folks will buy your gear that do with the kind of talk you have been presenting.

 

[jkeny]

Good detective work. You know your competitors well.

I know the area well & the technologies used

I have post the right graph and explanation in ASR Forum. I won't repeat it here: http://www.audiosciencereview.com/forum/index.php?threads/close-in-jitter.1621/page-10#post-41230

Net, net, I did indeed test the effect of some 40 db of close in/low-frequency jitter and it was not audible to me in blind tests.

If you want to make statements then bring the evidence. With the Audiophileo you didn't test close in phase noise as I already showed you from the Audiophileo site graph - it's not a clock which exhibits low close-in phase noise - no amount of repeating this will change that graph

Just like the fact that analog jitter similar range is not audible to people.

Now you got it. Random bandlimited jitter is even less audible due to strong masking. See critical bands.

You are chasing your tail around here, Amir - yes your test on the Audiophileo was of no worth & not a test that has anything to do with low PN clock. So doing a test which has no relevance to the subject being discussed & declaring it showed no difference is the same as doing a measurement which has no relevance & declaring no difference - both endeavours are useless & a waste of everybody's time

 

[amirm]

Your post has nothing to do with what I said & does not disagree with anything I've said about close-in phase noise which is predominantly due to the crystal itself, its cut, polish etc.

No, no, no. Did I not say that before? You continue to confuse the Crystal itself, with the performance of the entire subsystem that produces the clock for the DAC.

Again this stuff is in any online paper/app note from chip companies. Here is an example:

Notice how the curve tracks the phase noise plots posted. And notice the last line and such ultra-low levels we are talking about.

Anyway, I am bored correcting you. As I said, your choices are simple. Don't try to straddle objectivity without any solid foundation/references in your arguments.

 

[jkeny]

I am not following Don's argument. I am following yours and they continue to show errors. Bigly.

Are you channeling Trump?
I see so you won't admit Don is wrong & won't deal with the core of the issues being discussed preferring instead gross deflects into non-sequitors yet again

I will sum up your position and mine:
1. You are relying on what you think you have heard in sighted tests with strong bias ("good clock must sound better"). Even if you changed nothing, you would be hearing such improvements. So there is no basis here in scientific investigation.

2. You are then working backward, convincing yourself that there must be a difference here between digital and analog. Everything you say are your words, not presented in any paper, research, references, etc. Outside of that you are ignoring the fact that analog jitter is tons, tons higher than digital. So even if your argument was right, which it is not, you still have no explanation for why heaps of speed errors in analog at low frequencies is inaudible to people

3. Psychoacoustics completely backs what I explain. None of yours follows the same.

Bottom line is this: you can believe *subjectively* that better clock sounds better to you. And other people for that matter. That is perfectly fine. It is no different than folks claiming improvements for much lesser devices.

What you can't do is say audio science backs any of this. No audio science will accept your sighted, biased tests as the basis of anything. So you need to dispense with all of this technical talk. You can't be a part-time vegetarian.

If it is too uncomfortable for you to distance yourself from audio science, then you need to take audio science seriously. Start with proper testing protocols. Present that data. Let us repeat it. And then we can go on.

Alternatively adopt subjectivity fully and be done with it. Talk about free radicals this, and molecule alignment that, and you will be golden. Same number of folks will buy your gear that do with the kind of talk you have been presenting.

Aha, the old fallback snub at commercial interests - what's the problem, Amir, run out of argument - I would normally say logical argument but your is far from logical as I said it's a series of non-sequiturs posts hoping something will stick
The fact that you won't back Don's argument is very telling as he completely disagree with you & you with him

Let's summarise. There are 3 posters here (I don't count Frantz) who simply contradict one another (at least two of them do)

When I post that analogue wow is fundamentally different to close in clockl jitter - one is pitch wander whose origins are mechanical & the other is much finer grained & based on atomic fluctuations (what I termed macro & micro timing differences)
- Amir states "You are painting an imaginary scenario with an arbitrary difference built into it. There is no justification for either one. The analog one for example could also have "sudden" jumps due to bearing of a platter. Or snaps of the rubber belt. The digital one also has slew limiting and won't all of a sudden jump from one value to another value that is too large. On and on." He Contradicts Don Hills who says that analogue wow is a pitch error, ramping up & ramping down from the correct speed.
- Don Hills who maintains that clock jitter does the same ramping up & ramping down (in timing errors) & this is why it is exactly the same as analogue wow
- Robh3606 who presented the phase noise of a mysterious clock & somehow disagrees with J-test plots which show a steep slope to the rising skirt at the base of the main signal spike but I'm not sure what his disagreement is?


And on & on it goes

 

[amirm]

If you want to make statements then bring the evidence. With the Audiophileo you didn't test close in phase noise as I already showed you from the Audiophileo site graph - it's not a clock which exhibits low close-in phase noise - no amount of repeating this will change that graph

What are you talking about John? Here is the graph which has the same blue component you post:

What do you call all that jitter below 100 Hz? At 0.1 Hz the "bad" clock had massive low frequency jitter to the tune of 0 db!!! The good clock was 40 db better. Yet the difference was not audible to me. Just like psychoacoustics explains.

 

[jkeny]

What are you talking about John? Here is the graph which has the same blue component you post:

What do you call all that jitter below 100 Hz? At 0.1 Hz the "bad" clock had massive low frequency jitter to the tune of 0 db!!! The good clock was 40 db better. Yet the difference was not audible to me. Just like psychoacoustics explains.

And I already told you that the "good" clock you are showing in your graph is not a low PN clock when it comes to close-in PN (I already posted the clock I use - some 10 -20dB better for 10H to 0.1Hz) - so your claim that you have tested a clock with close-in PN is false .

Yes you have tested a clock with OK PN & one with bad PN - you heard nothing - so?

Another non-sequitur

 

[FrantzM]

I take offense on not being counted

 

[amirm]

And I already told you that the "good" clock you are showing in your graph is not a low PN clock when it comes to close-in PN (I already posted the clock I use - some 10 -20dB better for 10H to 0.1Hz) - so your claim is not that you have tested a clcok with close-in PN is false .Yes you have tested a clock with OK PN & one with bad PN - you heard nothing - so?

So this: the difference in that massive increase of 40 to 60 db in low frequency jitter was not audible. If that massive increase is not audible, you want us to believe that a 10 to 20 decrease is audible?

And furthermore, even higher jitter in analog systems is not audible to folks as the "roughness" or whatever you say. I listen to the audiophilleo every day on my system and it has none of the problems you mention.

Have someone run a blind test on you and report the results. And then send the same fixture and I will measure and do the same blind test if your results are positive. That is how you follow science.

My request remains: adopt subjectively fully.

 

[jkeny]

I take offense on not being counted

Then stand up & be counted - I can't see you when you are hiding in the grass

 

[jkeny]

So this: the difference in that massive increase of 40 to 60 db in low frequency jitter was not audible. If that massive increase is not audible, you want us to believe that a 10 to 20 decrease is audible?

1st order thinking, Amir, yet again. So apart from this not being a low close in PN clock but a standard clock that you are listening to - along with the increase in the close-in PN there's also a 20dB increase in the phase noise beyond 100Hz. Any guesses what might be the result of this?

Do you see anything wrong with changing a number of parameters other than just close-in PN & then claiming no difference as a result of close-in PN?

Hmmmm what was it you were saying about objectivity & science?

And furthermore, even higher jitter in analog systems is not audible to folks as the "roughness" or whatever you say.

Not the same mechanism as I said

I listen to the audiophilleo every day on my system and it has none of the problems you mention.

It's like a lot of what you don't hear, Amir - you try to say that there are "problems" which you don't hear so what is this 'solving'. It was the same with other devices as with this - you fail to understand that the sound can get noticeably more realistic, not that there are audible problems. When you can't hear something which is audible to many others you re-frame it into a 'problem' that is being solved & claim if you can't hear the problem, there's nothing to 'solve'.

I don't know if it's your training that has locked you into identifying 'problems' but whatever the reason something can sound better even though you never identified an audible problem in the first place with what you were hearing prior to the change.

Have you ever experienced this?

 

[amirm]

Do you see anything wrong with changing a number of parameters other than just close-in PN & then claiming no difference as a result of close-in PN?

No. As I explained, the bulk of the added jitter was close-in phase noise. What is added at higher frequencies is at far lower levels to be of audible concern.

Not the same mechanism as I said It's like a lot of what you don't hear, Amir - you try to say that there are "problems" which you don't hear so what is this 'solving'. It was the same with other devices as with this - you fail to understand that the sound can get noticeably more realistic, not that there are audible problems. When you can't hear something which is audible to many others you re-frame it into a 'problem' that is being solved & claim if you can't hear the problem, there's nothing to 'solve'.

I don't know if it's your training that has locked you into identifying 'problems' but whatever the reason something can sound better even though you never identified an audible problem in the first place with what you were hearing prior to the change.

Have you ever experienced this?

You haven't shown anything that is "audible to others" based on how audio science works. So the discussion ends there.

What you have shared is that in sighted, fully subjective listening, you and some unknown people think they hear an improvement. Which again prompts me to tell you to adopt full subjectivism and not constantly try to say there is some science to back what you are doing. There isn't. Period, end of discussion.

 

[jkeny]

No. As I explained, the bulk of the added jitter was close-in phase noise. What is added at higher frequencies is at far lower levels to be of audible concern.

Really? And we should believe you that 20dB added noise is of no consequence? We should believe you without any details about your listening configuration. We should believe you that the PLL in the SPDIF receiver in your DAC doesn't add it's own jitter? We should believe everything you say because of your exemplary measurement/testing history?

You haven't shown anything that is "audible to others" based on how audio science works. So the discussion ends there.

What you have shared is that in sighted, fully subjective listening, you and some unknown people think they hear an improvement. Which again prompts me to tell you to adopt full subjectivism and not constantly try to say there is some science to back what you are doing. There isn't. Period, end of discussion.

I'm happy you end it there as you have spectacularly failed at every attempt to discredit me or my analysis so end it there as I'm really not interested in your continual non-sequitor posts hoping something will stick or your constant deflections & obfuscations.

What you & Don have shown in this thread & many others is that you have little credibility regarding audio science

Don still thinks that clock close-in timing errors somehow work in an interconnected way - timing errors track up & down, just like a TT platter speed errors ramp up & down. Hence his arguing that it is just the same as analogue wow
In this thread you support Don, knowing that he is wrong but you wish to win in this thread as in every other thread

On this forum your past history is well known for your mistatements & misundrestandings which are bigly (just to speak your lingo)

You still haven't answered why you think USB isochronous protocol retransmits packets in error & claim John Swenson told you this
You still haven't corrected your mistaken belief that FFTs involve oversampling
You publish measurements of Schiit DAC which Stereophile disagree with & refuse to admit you might be wrong

 

[Don Hills]

You have but.... see my post somewhere in this thread ....

I did. I took your advice - eventually.
I only persisted because it wasn't a derailing of an existing thread - anyone reading the thread title and the first post would know instantly whether they wanted to wade into the mud pit or ignore it.

 

[jkeny]

I see Donh56 having to correct Amir's fundamental lack of understanding of basic maths & FFTs - yet again!! That's at least two completely basic wrongs now about FFTs - oversampling & now phase.

Did someone mention objective science?

Hmmm... While admitting I have not followed too closely lately (work and Life issues), I rarely use just real or imaginary data alone unless looking for a reactive component. The magnitude reported in an FFT is normally the RSS of real and imaginary data. The magnitude of a complex number is not accurately given by just the real (or imaginary) component; the magnitude is affected by each. I must have missed something crucial...

Tell us again, Amir - who educated you so we can avoid their teaching?

 

[jkeny]

Let's see what Zwicker & Fastl have to say about roughness (Psychoacoustics - Facts * Models) - your bible AFA psychoacoustics is concerned:

Bands of noise often sound rough, although there is no additional amplitude modulation. This is because the envelope of the noise changes randomly. These changes become audible especially for bandwidths in the neighbourhood of 100 Hz, where the average rate of envelope change is 64 Hz (see Sect. 1.1). Therefore, roughness is particularly large at such bandwidths.

And later gets more specific about FM modulation:

Frequency modulation can produce much larger roughness than amplitude modulation. A strong frequency modulation over almost the whole frequency range of hearing produces a roughness close to 6 asper. Only amplitude modulation of broad-band noises is able to produce such a large roughness.

And

Our hearing system is most sensitive for sinusoidal frequency modulations at frequencies of modulation in the neighbourhood of 4 Hz.

Note the words "sinusoidal frequencies", not dynamically changing music about which they say:

Musical tones, however, are rarely sinusoidal tones; they comprise many harmonic components and the frequency changes of these high-frequency harmonics can be more easily detected than frequency changes of the fundamental.

at frequencies below 500 Hz, we are able to di?erentiate between two tone bursts with a frequency di?erence of only about 1 Hz;