Close in jitter?

[Blumlein 88]

Isn't this, that Blumlein 88 posted earlier in this thread, a "jitter free" (and noise free) J-Test signal (for comparison)?

Yes I just took the Jtest that JA uses and put it in Wave Spectra. As you can see when Benchmark says their DAC is immune to jitter it comes very close to theoretical perfection on that point. The Chord Hugo is also very good though not quite as sharp a tone result around 11,025 hz. The base of that tone spreads just a bit more than the Benchmark.

 

[amirm]

Isn't this, that Blumlein 88 posted earlier in this thread, a "jitter free" (and noise free) J-Test signal (for comparison)?

Some of them, yes.

 

[Blumlein 88]

Here are couple views of the 44.1 khz 24 bit file.

 

[amirm]

That's correct. In 24-bit measurements those spikes are usually invisible.

 

[Jakob1863]

Similar assertions are made in both papers/articles. I have quoted them separately. In either case, it is just said that such and such is audible but with no further explanation, references or anything else.

Not in his AES-convention paper, therefore my suprise. But i had in my archive an older pdf from the grimm website about special test instruments they had to develop (according to their pdf). In it there is brief paragraph about the audibility of low frequency jitter (wander):

"All active analogue circuits suffer some form of flicker (1/f) noise. The effect of this
noise on clock oscillators is to produce lowfrequency jitter (wander), the effect of which
on sound has been greatly underestimated. Published audible jitter limits have so far
been derived from monaural listening tests on tones. A simple listening test establishes that
the minimum audible jitter on stereophonic material is orders of magnitude lower."

and

"The VLF nanovoltmeter addresses one cause of jitter. To measure jitter directly we have
long relied on a Wavecrest test set. Again, the fact that this instrument is designed to
measure jitter at any thinkable clock frequency ultimately limits the noise fl oor to around
3ps. It appears that discrete jitter components well below this fi gure can still lead to audible
degradation. To say that clock stability is a decisive factor in digital audio is nearly an
understatement."

They did not mention a controlled listening test, so it remains unclear, although afair Putzeys does usually use ABX-tests to confirm things like that........

 

[Jakob1863]

Julian Dunn calculates his original threshold curve by pure theoretical considerations, based on masking theory and some premises:
1.) single sinusoidal frequency audio signal
2.) single sinusoidal frequency jitter component
3.) calculated for replay at (for example) 120 dB SPL, so that the provoked distortion is just at/above the threshold

therefore the jitter amplitude should be below 20 ps (wrt 20 kHz) to be safe.
The 250ps fas42 mentioned above is presumably a peak to peak value and based on another method, that computes the maximum amount of jitter without degrading the overall quality of the CD system by more than 1/2 LSB.

Generally spoken (if not appropriate in this thread i´ll start another one) the usual process of discussions like these are always surprising me, because imo it is always the same since roughly 40-50 years ago.

I think we all can agree that Mivera did not provide evidence by a sufficienct number of controlled listening tests, that the close in noise of oscillators make a really audible difference (at the levels shown).
But does that mean, it is indeed inaudible,? No.

Grimm audio´s description might be pure anecdotical, but maybe they did more than sighted listening.

Even, if not, does it mean we can dismiss their results? No.

Does any evaluation based on the sideband tone theory help to decide whether any audibility may exist? No, because the exchange of oscillators is a real world experiment and the reason of an audible difference (if really existent) might be a side effect (means it might be dependent on certain circuits) or a situation simply not covered by our model right now.

We have to remember that our understanding is based on modelling the reality and that our models are always restricted.
The fact/feeling we have to evaluate is at the beginning that the exchange of oscillators gives the impression of an audible difference.
First step would be to do controlled listening tests to gain further evidence, based on the anecdotical impressions.

Second step would be (under the premise that he first step was successfull) to develop a measurement solution to see whats happening at the DAC output.

 

[oivavoi]

Julian Dunn calculates his original threshold curve by pure theoretical considerations, based on masking theory and some premises:
1.) single sinusoidal frequency audio signal
2.) single sinusoidal frequency jitter component
3.) calculated for replay at (for example) 120 dB SPL, so that the provoked distortion is just at/above the threshold

therefore the jitter amplitude should be below 20 ps (wrt 20 kHz) to be safe.
The 250ps fas42 mentioned above is presumably a peak to peak value and based on another method, that computes the maximum amount of jitter without degrading the overall quality of the CD system by more than 1/2 LSB.

Generally spoken (if not appropriate in this thread i´ll start another one) the usual process of discussions like these are always surprising me, because imo it is always the same since roughly 40-50 years ago.

I think we all can agree that Mivera did not provide evidence by a sufficienct number of controlled listening tests, that the close in noise of oscillators make a really audible difference (at the levels shown).
But does that mean, it is indeed inaudible,? No.

Grimm audio´s description might be pure anecdotical, but maybe they did more than sighted listening.

Even, if not, does it mean we can dismiss their results? No.

Does any evaluation based on the sideband tone theory help to decide whether any audibility may exist? No, because the exchange of oscillators is a real world experiment and the reason of an audible difference (if really existent) might be a side effect (means it might be dependent on certain circuits) or a situation simply not covered by our model right now.

We have to remember that our understanding is based on modelling the reality and that our models are always restricted.
The fact/feeling we have to evaluate is at the beginning that the exchange of oscillators gives the impression of an audible difference.
First step would be to do controlled listening tests to gain further evidence, based on the anecdotical impressions.

Second step would be (under the premise that he first step was successfull) to develop a measurement solution to see whats happening at the DAC output.

You much sense are making, Jakob1863.

 

[Cosmik]

Is it accepted that the error due to jitter can be calculated? Not too controversial I think.

The next step is to ask: is it accepted that the worst case error for some particular setup when playing music could be simulated and amplified by a large factor (100, say) and played in isolation over a very good DAC at normal listening levels?

If such an experiment resulted in an almost imperceptible hiss or, indeed, silence, could we say that the problem is purely theoretical?

If anyone wanted to argue that an error at 1/100 of the level of imperceptible - even when there is no masking from the main signal - is a problem, they might have to work quite hard at it.

 

[DonH56]

Error in a data converter from RJ is readily calculated and correlates well to measured data. The impact of other forms of jitter, well, depending on the signal, may or may not have closed-form solutions. It should be pretty easy to set up an experiment using a JBERT as a clock source and good DAC to output various signals with varying amounts and types of jitter applied. Of course you'll need the DAC and the JBERT, and the latter is not cheap. The JBERT has programmable jitter sources (that's the "J") including RJ, SJ, high- and low-frequency jitter sources of BUJ, and so forth, in addition to common-mode interference, differential-mode interference, and other impairments you could add to the signal.

Any time you do such an experiment one of the problems is determining if you have limited the variables to only the one of interest. That can be a lot harder than it sounds, especially since they are not always known.

Note there are several source of low-frequency noise and flicker is only one of them. LF modulation has long been a concern for ages (from tape/vinyl sources on) but is not something I have particularly followed. It is also not something I claim to know particularly well since my career has focused more on HF circuits and systems. It does impact a lot of systems, however, such as wideband high-resolution radar and lidar systems that had 120 - 140 dB dynamic range decades ago and required essentially DC response in IQ (in-phase/quadrature) systems. But since that is not technically audio...

I am going to bow out of this; in the previous thread my opinions and experience were not wanted and since it got fairly personal I took the step of placing several members on ignore. It has helped my blood pressure, and hopefully helped prevent me from responding in kind, but made that and this thread impossible to follow.

Enjoy - Don

 

[Thomas savage]

Error in a data converter from RJ is readily calculated and correlates well to measured data. The impact of other forms of jitter, well, depending on the signal, may or may not have closed-form solutions. It should be pretty easy to set up an experiment using a JBERT as a clock source and good DAC to output various signals with varying amounts and types of jitter applied. Of course you'll need the DAC and the JBERT, and the latter is not cheap. The JBERT has programmable jitter sources (that's the "J") including RJ, SJ, high- and low-frequency jitter sources of BUJ, and so forth, in addition to common-mode interference, differential-mode interference, and other impairments you could add to the signal.

Any time you do such an experiment one of the problems is determining if you have limited the variables to only the one of interest. That can be a lot harder than it sounds, especially since they are not always known.

Note there are several source of low-frequency noise and flicker is only one of them. LF modulation has long been a concern for ages (from tape/vinyl sources on) but is not something I have particularly followed. It is also not something I claim to know particularly well since my career has focused more on HF circuits and systems. It does impact a lot of systems, however, such as wideband high-resolution radar and lidar systems that had 120 - 140 dB dynamic range decades ago and required essentially DC response in IQ (in-phase/quadrature) systems. But since that is not technically audio...

I am going to bow out of this; in the previous thread my opinions and experience were not wanted and since it got fairly personal I took the step of placing several members on ignore. It has helped my blood pressure, and hopefully helped prevent me from responding in kind, but made that and this thread impossible to follow.

Enjoy - Don

Well, wanted maybe a false prism to view the worth of your contributions don.., very much needed and entirely essential would fit more adeptly.

I apologise for the unpleasant feeling you've expirenced, its entirely my responsibility to prevent Such from happening.

 

[amirm]

Not in his AES-convention paper, therefore my suprise. But i had in my archive an older pdf from the grimm website about special test instruments they had to develop (according to their pdf). In it there is brief paragraph about the audibility of low frequency jitter (wander):

"All active analogue circuits suffer some form of flicker (1/f) noise. The effect of this
noise on clock oscillators is to produce lowfrequency jitter (wander), the effect of which
on sound has been greatly underestimated. Published audible jitter limits have so far
been derived from monaural listening tests on tones. A simple listening test establishes that
the minimum audible jitter on stereophonic material is orders of magnitude lower."

and

"The VLF nanovoltmeter addresses one cause of jitter. To measure jitter directly we have
long relied on a Wavecrest test set. Again, the fact that this instrument is designed to
measure jitter at any thinkable clock frequency ultimately limits the noise fl oor to around
3ps. It appears that discrete jitter components well below this fi gure can still lead to audible
degradation. To say that clock stability is a decisive factor in digital audio is nearly an
understatement."

They did not mention a controlled listening test, so it remains unclear, although afair Putzeys does usually use ABX-tests to confirm things like that........

Well, the first one is definitely wrong. Tests have been performed with music and I quoted them in the other threads: https://secure.aes.org/forum/pubs/conventions/?elib=8354
"Theoretical and Audible Effects of Jitter on Digital Audio Quality

Many digital audio systems now use some form of self-clocked digital interface for audio delivery. With the advent of new digital audio systems that use IEC 61937 to convey non-linearly coded audio, the total number of devices using the IEC 60958 interface has substantially increased. The digital interface may contain jitter that translates to distortion in the audio at the point of conversion back to the analog domain. Sources of digital audio, the digital interface, the mechanisms by which errors are introduced, and the effect on DACs are examined

Authors: Benjamin, Eric; Gannon, Benjamin
Affiliation: Dolby Laboratories Inc., San Francisco, CA
AES Convention:105 (September 1998) Paper Number:4826
Publication Date:September 1, 1998 "

As you see above, music selections were most definitely used (after testing with pure sine waves).

His presumption also that sinewaves show higher thresholds is just incorrect. Real program music has complex spectrum that can mask jitter better than pure tones. Here is Dolby paper on that:

And the tests above were done at much higher Jitter frequencies. At lower jitter frequencies we are talking about, masking becomes so much more powerful. Here is Dolby paper on that:

And:

And the punchline:

Even after training and selecting material that showed jitter artifacts better than others, the threshold remained in tens of *nanoseconds* not picoseconds which Bruno talks about. And again, these are with jitter frequencies in 1000 to 2000 Hz.

I have taken real jitter tests with frequency of 30 Hz and I tell you once that level got to -60 db SPL, it was excruciatingly hard for me to find the degraded file in ABX testing. As far as I know, I am the only one that could pass that test at that level. Nobody else could when the test was presented on AVS Forum.

So if he has evidence that runs counter to everything we know in psychoacoustics and threshold of hearing, he needs to publish that as that will be *major news*. Just declaring it doesn't amount to anything.

Let me say that I am an advocate of systems with low jitter. I think good hygiene in design gets us there and products that fail that, just show incompetence on behalf of the designer. The case from audibility gets very hard and is not a domain people should enter unless they have taken such controlled tests themselves and studied psychoacoustics.

 

[amirm]

I found the names of the files in the 30 Hz jitter test (the actual dropbox link no longer works). The files were revised and my results are for those but can't find that list just now.

And here are my ABX test results from bottom to top (easy to hard):

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/06/18 16:44:39

File A: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\no jitter.wav
File B: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\30 Hz jitter strong level .025.flac

16:44:39 : Test started.
16:45:05 : 01/01 50.0%
16:45:15 : 02/02 25.0%
16:45:28 : 03/03 12.5%
16:45:36 : 04/04 6.3%
16:45:54 : 05/05 3.1%
16:46:17 : 06/06 1.6%
16:46:29 : 07/07 0.8%
16:46:45 : 08/08 0.4%
16:46:55 : 09/09 0.2%
16:47:05 : 10/10 0.1%
16:47:19 : 11/11 0.0%
16:47:33 : 12/12 0.0%
16:47:36 : Test finished.

----------
Total: 12/12 (0.0%)

===================
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/06/18 16:39:16

File A: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\no jitter.wav
File B: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\30 Hz Severe Jitter 0.05.flac

16:39:16 : Test started.
16:39:52 : 01/01 50.0%
16:40:26 : 02/02 25.0%
16:40:38 : 03/03 12.5%
16:40:48 : 04/04 6.3%
16:40:58 : 05/05 3.1%
16:41:07 : 06/06 1.6%
16:41:24 : 07/07 0.8%
16:41:32 : 08/08 0.4%
16:41:51 : 09/09 0.2%
16:42:04 : 10/10 0.1%
16:42:12 : 11/11 0.0%
16:42:21 : 12/12 0.0%
16:42:43 : Test finished.

----------
Total: 12/12 (0.0%)

===================

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/06/18 19:04:40

File A: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\no jitter.wav
File B: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\30 Hz noticable jitter 0.0125.flac

19:04:40 : Test started.
19:05:27 : 01/01 50.0%
19:05:54 : 02/02 25.0%
19:06:19 : 03/03 12.5%
19:06:35 : 04/04 6.3%
19:06:57 : 05/05 3.1%
19:07:16 : 06/06 1.6%
19:07:43 : 07/07 0.8%
19:08:15 : 08/08 0.4%
19:08:37 : 09/09 0.2%
19:09:05 : 10/10 0.1%
19:09:30 : 11/11 0.0%
19:10:05 : 12/12 0.0%
19:10:09 : Test finished.

----------
Total: 12/12 (0.0%)

=================

Then it starts to get hard. See my multiple trials to get this level right:
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/06/21 14:50:24

File A: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\no jitter.wav
File B: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\30 Hz jitter marginal level .00625.flac

14:50:24 : Test started.
14:51:40 : 01/01 50.0%
14:52:05 : 02/02 25.0%
14:52:33 : 03/03 12.5%
14:52:57 : 03/04 31.3%
14:53:22 : 03/05 50.0%
14:54:11 : 03/06 65.6%
14:54:54 : 04/07 50.0%
14:55:21 : 05/08 36.3%
14:56:02 : 06/09 25.4%
14:56:55 : 07/10 17.2%
14:57:24 : 08/11 11.3%
14:58:17 : 08/12 19.4%
14:59:37 : 09/13 13.3%
15:00:44 : 10/14 9.0%
15:03:01 : 10/15 15.1%
15:04:11 : 11/16 10.5%
15:04:32 : Test finished.

----------
Total: 11/16 (10.5%)

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/09 09:31:59

File A: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\no jitter.wav
File B: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\30 Hz jitter marginal level .00625.wav

09:31:59 : Test started.
09:32:47 : 01/01 50.0%
09:32:58 : 02/02 25.0%
09:33:09 : 03/03 12.5%
09:33:22 : 04/04 6.3%
09:33:33 : 04/05 18.8%
09:34:03 : 05/06 10.9%
09:34:18 : 06/07 6.3%
09:34:30 : 07/08 3.5%
09:34:46 : 08/09 2.0%
09:34:56 : 08/10 5.5%
09:35:19 : 09/11 3.3%
09:35:34 : 10/12 1.9%
09:35:49 : 11/13 1.1%
09:36:38 : 12/14 0.6%
09:37:21 : 12/15 1.8%
09:37:41 : 13/16 1.1%
09:37:52 : 14/17 0.6%
09:38:13 : 15/18 0.4%
09:38:26 : 16/19 0.2%
09:38:39 : 17/20 0.1%
09:38:45 : Test finished.

----------
Total: 17/20 (0.1%)

==================

And the final test:

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/09 17:39:55

File A: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\no jitter.wav
File B: C:\Users\Amir\Music\Arny's 30 Hz Jitter File\Arny's new files\30 Hz jitter barely noticable level .015.flac

17:39:55 : Test started.
17:40:40 : 00/01 100.0%
17:41:30 : 01/02 75.0%
17:41:41 : 02/03 50.0%
17:41:52 : 03/04 31.3%
17:42:04 : 04/05 18.8%
17:42:19 : 05/06 10.9%
17:42:32 : 06/07 6.3%
17:42:46 : 07/08 3.5%
17:42:58 : 07/09 9.0%
17:43:12 : 07/10 17.2%
17:43:27 : 07/11 27.4%
17:43:42 : 08/12 19.4%
17:43:53 : 08/13 29.1%
17:44:15 : 08/14 39.5%
17:44:46 : 09/15 30.4%
17:45:00 : 10/16 22.7%
17:45:12 : 11/17 16.6%
17:45:30 : 12/18 11.9%
17:45:52 : 12/19 18.0%
17:46:23 : 13/20 13.2%
17:46:28 : Test finished.

----------
Total: 13/20 (13.2%)

============

As you see I could no longer achieve 95% confidence.

I took the initiative to run these tests. It is not hard to create these tests. Just a modicum of effort needs to go into such strong claims that run counter to full science of psychoacoustics.

 

[amirm]

The music track was from Joni Mitchell by the way. So not pure tones.

 

[Thomas savage]

From my point of view I want the best engineered replay system , there's all sorts of areas in audio ( home audio reproduction) where engineering out imperfections takes precedence over ' what we might or might not hear' .

As long as we can show a objectively measured improvement in performance I don't care if it's audible.. that's a never ending argument that unless the engineering undertaking is exceedingly expensive is imo irrelevant to the end user. ( it's just too complicated for me or most audiophiles to understand and takes a huge amount of money and effort to prove)

It's for manufacturers to decide if increasing performance ideals is market appropriate or not..

if only the manufacturers would admit it's a objective improvement that might well only be academic in nature, rather than claiming all sorts of unproven audibly benefits.. most of the guys into high end audio buy gear for excellence, engineering excellence is a enjoyable ideal regardless of audibly.

Before you all lynch me, save your indignation for more pertinent areas of your life's, after all this is only audio and ment to be fun

 

[Nightlord]

Right. Pushing the limits further into the undetectable I rank in the category of 'feelgood'.

 

[amirm]

if only the manufacturers would admit it's a objective improvement that might well only be academic in nature, rather than claiming all sorts of unproven audibly benefits.. most of the guys into high end audio buy gear for excellence, engineering excellence is a enjoyable ideal regardless of audibly.

Exactly. This was my message to Mike (Mivera) which he never listened. That it is OK in this forum to advocate engineering excellence demonstrated through proper measurements. But he didn't listen and advocated all of these audibility claims that just have no foundation.

Jitter has been a topic that I have debated for more than a decade. Indeed for vast majority of that time, I have taken the above position. But not to the point of ridiculousness just 0.5 ps of jitter. It just doesn't pass the "red-face" test when our DACs don't have that kind of resolution.

 

[Thomas savage]

Right. Pushing the limits further into the undetectable I rank in the category of 'feelgood'.

Great, feeling good never hurt anyone... well unless you enjoy the suffering of others like amir

 

[RayDunzl]

Here at Neverland East, we strive to "doo da best we can".

The low end performance metric for gear and recordings is whether or not it is "better than nothing".

The high end metric is that it (whatever) has to be (but not certainly not limited to) "good enough".

 

[Thomas savage]

Here at Neverland East, we strive to "doo da best we can".

The low end performance metric for gear and recordings is whether or not it is "better than nothing".

The high end metric is that it (whatever) has to be (but not certainly not limited to) "good enough".

I'd want a beer ⛺️....

 

[RayDunzl]

Down in Joe's Garage
We didn't have no dope or LSD
But a couple of quarts of beer
Would fix it so the intonation
Would not offend your ear
And the same old chords goin' over and over
Became a symphony
We could play it again and again and again
Cause it sounded good to me
(One more time)