[JJ15k]

what do you mean by compatibility issues?

 

[Carl]

Quote:

Originally Posted by JJ15k what do you mean by compatibility issues?

I2S uses a funky DIN connector that looks more like a serial cable for a printer than anything audiophiles would be familiar with. You also need a second set of cables for genlocking.

So basically it's too difficult to get people to use it, despite being demonstratably superior to SPDIF. The professional industry is having the same problem with AES/EBU (as an aside, XLR connectors are rather bad for jitter, too).


Too many audiophiles fall for marketing gimmucks over actual performance.

 

[JJ15k]

Oh, I didn t know about that, I understand why it would be an issue

 

[darkless]

Quote:

Originally Posted by Carl I2S uses a funky DIN connector that looks more like a serial cable for a printer than anything audiophiles would be familiar with. You also need a second set of cables for genlocking. So basically it's too difficult to get people to use it, despite being demonstratably superior to SPDIF. The professional industry is having the same problem with AES/EBU (as an aside, XLR connectors are rather bad for jitter, too). Too many audiophiles fall for marketing gimmucks over actual performance.

I'd like to know more about the issue of jitter when using XLR connectors (and therefore AES/EBU) compared to other means of connecting digital gear.

Do you have any links to information supporting these claims?

 

[pawlowski6132]

Quote:

Originally Posted by granodemostasa Okay, so everywhere people talk about jitter...what does it sound like...?

I wouldn't know!!

 

[NotJeffBuckley]

Quote:

Originally Posted by Ferbose Jitter has nothing to do with bits. It is analog in nature. Jitter is the clock inaccuracy in A/D or D/A conversion. Jitter exists because there is no perfect clock and can get progressively worse as clock signal is transmitted in the system, or by interference. Of course there are methods to correct for jitter but complete elimination is impossible.

After all, the simultaneity of two distanced events parallel to the plane of motion can't be objectively evaluated. Jitter correction is relativistic but imperfect, and can't be perfect lacking an absolute reference frame (which, as we know, doesn't exist. No preferred frame, etc.)

Sorry for confusing the terminology.

 

[Beauregard]

For the knowledgeable: I've often seen the statement from AE types that the noise introduced by upstream and internal jitter is part and parcel of the THD+N spec for a digital component. A jitter thread on another forum linked to an engineer's diary of developing a USB audio device. I surely didn't understand the details, but it was clear that he was using the analog THD+N measure to determine the efficacy of his jitter-reduction efforts for each iteration of the project.

How valuable do you think THD+N is for a consumer shopping for a DAC - or deciding which products are worth auditioning? Anyone have any guidelines for what might constitute good vs not-so-good specs that would be reflected as audible "distortion".

Thanks,
Beau

 

[drarthurwells]

Quote:

Originally Posted by Carl Jitter sounds like wrong. How's that for an English sentence. The stuff doesn't have a defining sound colouration that you can put your finger on, it just makes the music harsh and unnatural, and sound less like it is live. The stuff affects the music because delta-sigma DACs work on a high-speed swiching principle (imagine a guy toggling a switch that has only two choices; "up" and "down", really, really fast, trying to recreate a sine wave), and any timing errors confuse the hell out of them (making them go up and down at the wrong times, and thus unable to make the sine wave properly). Analogue music, like from vinyl and cassettes, is always a perfect wave, but CD players, MP3 players, and their ilk can't quite reach that goal in reality. At the very least they can keep jitter to a minimum so that they get very, very close to the goal. There are other kinds of DACs that don't use the high speed switching principle (namely sign-magnitude DACs and PWM converters) that work on slower, but more accurate, techniques, and are thus not affected by jitter as much. However, they aren't used as much as they (in my opinion) should be.

Art: Yes, timing errors. Timing errors create phase distortion - brief timing delays ranging from some glare or hardness to a shattering distortion. Longer delays muddy tones and reduce image focus. Real long dealys of 12 miliseconds or more, can provide some echo effect that simulates hall ambience and can be pleasing.

Now, if these brief delays are a consistent slowing down then immediate renormalization, of the whole bit stream, wouldn't there be turntable "flutter" or "wow" as Ferbose alludes to? Don't know that this is how jitter works in the data stream - it may be a partial effect of some data rather than of the whole data stream.

What has always bothered me about digital is its sparse sampling of analog data - about one of 20 pieces of analog data are sampled.

 

[chris719]

The number of data points sampled at 44.1 KHz is perfectly sufficient to reproduce waveform containing no higher frequencies than 22.050 KHz without any information loss at all. Google the Nyquist-Shannon Sampling Theorem, it will explain how this is possible. Although we do not have perfect lowpass filters in reality, we do a surprisingly good job at reconstruction.

 

[Carl]

Quote:

Originally Posted by darkless I'd like to know more about the issue of jitter when using XLR connectors (and therefore AES/EBU) compared to other means of connecting digital gear. Do you have any links to information supporting these claims?

I've read several articles from audio engineers on it (basically bemoaning how bad AES/EBU is), be I'd be blown if I could find you a link to one. If memory serves it's because the XLR connection has capatitance issues or something like that.

I believe in order of least jitter to most jitter with SPDIF and AES interfaces is STOptical->Single Ended Coaxial->XLR->Toslink.

 

[Carl]

Quote:

Originally Posted by chris719 The number of data points sampled at 44.1 KHz is perfectly sufficient to reproduce waveform containing no higher frequencies than 22.050 KHz without any information loss at all. Google the Nyquist-Shannon Sampling Theorem, it will explain how this is possible. Although we do not have perfect lowpass filters in reality, we do a surprisingly good job at reconstruction.

The nyquest theorum is really intended for multibit DACs. To properly recreate the waveform using single-bit DACs requires quite a lot of oversampling because of the quantisisation noise.

 

[Ferbose]

Quote:

Originally Posted by drarthurwells Art: Yes, timing errors. Timing errors create phase distortion - brief timing delays ranging from some glare or hardness to a shattering distortion. Longer delays muddy tones and reduce image focus. Real long dealys of 12 miliseconds or more, can provide some echo effect that simulates hall ambience and can be pleasing. Now, if these brief delays are a consistent slowing down then immediate renormalization, of the whole bit stream, wouldn't there be turntable "flutter" or "wow" as Ferbose alludes to? Don't know that this is how jitter works in the data stream - it may be a partial effect of some data rather than of the whole data stream.

Jitter is not rally the slowing down of the entire clock, but just that each time point deviates a bit ahead and behind. We are talking about 100-1000 ps errors here. The analogy would be a clock that is accurate over a whole year. But each time it reports a second it is actually 1.00000001 or 0.99999999 seconds.
The artifact of jitter is mainly side bands that are not harmonically related to the signal. The ear is very sensittive to aharmonic distortions.

 

[UezeU]

Quote:

Originally Posted by NotJeffBuckley After all, the simultaneity of two distanced events parallel to the plane of motion can't be objectively evaluated. Jitter correction is relativistic but imperfect, and can't be perfect lacking an absolute reference frame (which, as we know, doesn't exist. No preferred frame, etc.) Sorry for confusing the terminology.

Love to see a bit of relatively. Of course Einstein meant points (in space-time), not actual events.That is where the HUP occurs. To put it practically, if your analogy is true; then CMBR, would give a relitive reference. We all know that is false........DUH!

Edit:
huh...huh..huh, I said ANALogy....