[tfarney]

Quote:

Wes Phillips found that the Oppo player sounded somewhat soft and overripe in the midbass, which is one consequence of high amounts of random jitter, in my experience.

Wow. People pay big money for that effect in audiophile equipment every day. Who knew they could get it from a $200 dvd player?

Tim

 

[tfarney]

Quote:

Originally Posted by Pars /img/forum/go_quote.gif I thought you guys had summarily dispensed of jitter as even being a factor worth investigating? Also I don't see any jitter measurements of transport/DAC combos in those posted. Those could be interesting.

Evidently the dismissal will not be summary.

Tim

 

[nick_charles]

Quote:

Originally Posted by Pars /img/forum/go_quote.gif I thought you guys had summarily dispensed of jitter as even being a factor worth investigating? Also I don't see any jitter measurements of transport/DAC combos in those posted. Those could be interesting.

You need to dredge through the Stereophile measurements sections more closely , they do sometimes discuss the effect of different transports, specifically they do talk about jitter of the RAdio Shack player as a transport.

Probably they do the same for more esoteric transports.

 

[nick_charles]

Quote:

Originally Posted by tfarney /img/forum/go_quote.gif Evidently the dismissal will not be summary. Tim

One of the players I saw figures for (Panasonic SL220) had jitter sidebands at -38db, if jitter at -130db is a problem (which jitter-worriers would insist is true) then jitter at -38db, a factor of 32768 worse would be utterly unlistenable and not just in a subjective "wow that is bad" way but in a "we cannot possibly market this player" way.

So in the interest of science I am going to buy a Panasonic SL220 to see just how bad it really is, I have found one for sale in Canada, and I will report back my findings.

 

[tfarney]

Quote:

Originally Posted by nick_charles /img/forum/go_quote.gif One of the players I saw figures for (Panasonic SL220) had jitter sidebands at -38db, if jitter at -130db is a problem (which jitter-worriers would insist is true) then jitter at -38db, a factor of 32768 worse would be utterly unlistenable and not just in a subjective "wow that is bad" way but in a "we cannot possibly market this player" way. So in the interest of science I am going to buy a Panasonic SL220 to see just how bad it really is, I have found one for sale in Canada, and I will report back my findings.

Sir, I admire your dedication to scientific discovery. And your willingness to take one for the team. Now if we could just get those black boxes into some independent testing...

Tim

 

[nick_charles]

Quote:

Originally Posted by tfarney /img/forum/go_quote.gif Sir, I admire your dedication to scientific discovery. And your willingness to take one for the team. Now if we could just get those black boxes into some independent testing... Tim

Easier than you think. Since jitter is inaudible until the signal becomes analog you just record (digitize) the analog out of a DAC twice, once with the jitter reducing box in place and once without it. Then you just look at the resultant waveforms and energy spectrums.

The difference in jitter distortion will show up as noise, also a software spectrum analyzer will show any deviations in min, max and average energy levels and sample values. Also you can use CD rips as benchmarks to show just how far from transparent the DAC/ADC step is in general.

Oh, and then you can do some blind testing as well

 

[b0dhi]

Quote:

Originally Posted by tfarney /img/forum/go_quote.gif Actually, the best test would be an objective one -- a double blind listening test of music being played on a reference system,

I've highlighted a particular word in the quote you were responding to. See if you can figure out which it was, and why.

 

[tfarney]

Quote:

Originally Posted by b0dhi /img/forum/go_quote.gif I've highlighted a particular word in the quote you were responding to. See if you can figure out which it was, and why.

Why did you highlight tfarney?

Tim

 

[nick_charles]

I ran the jitter-free and -80db jittered signals through a spectrum analyzer, the differences over the audible spectrum were quite small. Figures are in db

ave------0.012879333
min------0.000122
max-----0.075943

However when I did the same for the -20db jittered signal the differences were much bigger


ave------0.83444462
min------0.000196
max------22.303299

However please note that the big differences are in the 20K - 22K range, and at energy levels that get no higher than -76db, below 20K the maximum difference is 0.472 db and the average difference is 0.0288 db

 

[minivan]

i found some measurement of my dac ,these sound like what b0dhi is looking for, square wave? figure 1 is the i2s signal after recovering from spdif, figure 2 is the i2s signal after recovering from tg-link.

some1 please enlighten me why i can hear an improvement of sq from tglink over spdif based on these
graphs. i dont know how to read these grahps
another ones

 

[Pedja]

Hi all,

My first post here, though I think I know some of you, and some of you may know me as well.

It was the thread title that pulled me in here, and I've read most of the pages. Initial post assumed that the PC based audio has less jitter than conventional digital audio, and question was on possible sources of jitter in PC audio. The assumption wasn't really correct but the question was very good. I can witness both that different PC interfaces perform different and that different PCs perform different.

Data storage and signal processing are obvious advantages of PC and I believe this will become even more obvious and more important by time. But when you consider PC jitter performance, the things with it are not easy at all. You have firstly to consider particular interface you use. Professionals will slave the PC to the external clock thus (mostly) discarding PC from jitter equation. It is however not how PCs are used in home environment. In fact, not all professionals apply that scheme. Then you have to go through particular interfaces (USB, S/PDIF, Firewire). Finally you have to go through hardware issues. Still, there is more about sound quality from PC than is the jitter. Noise is one if such issues, and it is probably the same important.

As for the jitter measurements, no need to invent the method, it's been done already. You can measure either the clock line directly (i.e. its timing deviation, and you can observe this either in time domain or by spectral analysis), or you can measure the audio output signal observing frequency modulation of high frequency audio carrier.

Pedja

 

[tfarney]

Quote:

As for the jitter measurements, no need to invent the method, it's been done already. You can measure either the clock line directly (i.e. its timing deviation, and you can observe this either in time domain or by spectral analysis), or you can measure the audio output signal observing frequency modulation of high frequency audio carrier.

We know it can be measured. We're debating whether or not it can be heard.

Tim

 

[Pedja]

I understand this difference very well, but if I can notice, majority of discussions on jitter these days I stumble across, and even many things said in this thread, basically assume that it can be heard but it can not be measured, at least not in some practical way.

(BTW, it is not clear what the graphs above show about the jitter... it looks like they show relation between the word clock, supposedly 96kHz and bit clock, supposedly 6.144MHz, when triggered on raising edge of word clock; the difference between the two shots would be a jitter only if 96kHz was some steady and synchronized time base.)

Anyhow, one has to induce jitter and to remove it in a controlled way to get some idea about its audibility. Unfortunately, only a fraction of people has such an experience.

Mind you, audibility of jitter is also not the initial question of this thread, the question was about sources of jitter in PC audio, which can be discussed regardless of its audibility.

And while we are at this, because of its possibilities in signal processing area, PC can be a great tool to simulate and induce jitter intentionally, or more precisely speaking, to induce jitter associated distortion/artifacts.

Pedja

 

[tfarney]

Quote:

I understand this difference very well, but if I can notice, majority of discussions on jitter these days I stumble across, and even many things said in this thread, basically assume that it can be heard but it can not be measured, at least not in some practical way.

Well, I think that's close. If I understand what I'm looking at/reading, there is no question that jitter can be measured, but it is only being detected at levels well below what humans should be able to hear. And many in this thread, and elsewhere, basically assume that they hear it anyway.

Tim

 

[regal]

Pars, how would you measure jitter of a transport/DAC, is there a test method with NIST traceable standards? Is there an ASTM for jitter in a transport/DAC combo? Any measurement is only as good as the test method. What is the 90% CI of these jitter measurements, are they repeatable? Has there ever been a Guage R&R?