IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
S/PDIF Cable and Jitter
Speedskater
post Mar 31 2011, 17:53
Post #1





Group: Members
Posts: 138
Joined: 12-December 07
From: Cleveland, OH
Member No.: 49500



It seems that several audio forums and now a home theater forum have had recent threads about how S/PDIF interconnect cables cause jitter. I thought that S/PDIF was a rather robust system and only poorly designed equipment was affected by the interconnect cable. So what is the true story?


--------------------
Kevin
Go to the top of the page
+Quote Post
DVDdoug
post Mar 31 2011, 18:37
Post #2





Group: Members
Posts: 2566
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



Jitter is one of those made-up problems for people who just can't believe how audibly perfect simple & cheap digital systems are! wink.gif

I don't understand all of the details, but Ias I understand it, the S/PDIF receiver has a buffer and it "cleans-up" jitter on the receiving end.

Ethan Winer has an article where discusses jitter. (He's also got a lot of other information debunking audiophile nonsense.) Or you can search HydrogenAudio for "jitter" and find lots of good info.

QUOTE
I thought that S/PDIF was a rather robust system
It is! It would take a very poor or defective, or very long cable, to cause problems.

And generally, when things go wrong digitally they go horribly wrong! It's not like analog where noise & distortion creep-in slowly. If you scramble or corrupt a digital number, there's nothing to keep the errors small. A number can get changed by 10,000% just as easily as it can get changed by 0.01%.

This post has been edited by DVDdoug: Mar 31 2011, 18:42
Go to the top of the page
+Quote Post
AndyH-ha
post Mar 31 2011, 19:53
Post #3





Group: Members
Posts: 2207
Joined: 31-August 05
Member No.: 24222



Even if cables introduced extreme jitter (which they donít) it would not make any difference to the end result. A properly working DAC re-times the data to its own clock, so that clock determines the limit of how much jitter can be introduced into playback.
Go to the top of the page
+Quote Post
googlebot
post Mar 31 2011, 20:05
Post #4





Group: Members
Posts: 698
Joined: 6-March 10
Member No.: 78779



Reclocking is actually pretty rare. Normally the clock is recovered from the S/PDIF signal. But the answers were correct anyway. Jitter is no deal nowadays.
Go to the top of the page
+Quote Post
pbelkner
post Apr 1 2011, 06:43
Post #5





Group: Members
Posts: 412
Joined: 13-June 10
Member No.: 81467



QUOTE (googlebot @ Mar 31 2011, 20:05) *
Reclocking is actually pretty rare.

If this is true, then that's exactly what I understand by the term jitter (a random drift in the frequency by which the samples are delivered, the bits themselfs are nor disturbed, of course).
Go to the top of the page
+Quote Post
cliveb
post Apr 1 2011, 06:59
Post #6


WaveRepair developer


Group: Developer
Posts: 838
Joined: 28-July 04
Member No.: 15845



Any system where the DAC is fed via SPDIF, and which does not have the DAC sending a clock back to the transport (ie. 99%+ of all consumer setups were a separate DAC is in use), must by definition require the DAC to recover a clock from the SPDIF signal. Most DACs these days use one of a variety of strategies to smooth out any short scale timing inaccuracies. The majority will use a long timebase PLL. Some (the Benchmark DAC1 being the best known example) use asynchronous sample rate conversion. A very small number use extensive buffering. But all of these approaches boil down to the same result: the DAC ends up using its own clock that runs at the same average rate as that of the clock in the incoming SPDIF stream, and where any timing inaccuracies are "smoothed out" over a much longer period. But they are all still affected to some degree by variations in that incoming clock.

And it is certainly the case that different cables used for the SPDIF connection can affect the jitter on the signal.

All that said, there have been properly conducted listening tests to determine the audibility of jitter. Two obvious examples are:

Benjamin & Gannon.
Theoretical and audible effects of jitter on digital audio quality.
105th AES Convention, 1998
Jitter added to digital signal between transport and DAC with a hardware device.
Conclusions: uncorrelated jitter inaudible below 10nS rms on pure tones;
uncorrelated jitter inaudible below 20nS rms on music signal

Ashihara, Kiryu et al.
Detection threshold for distortions due to jitter on digital audio.
Acoust. Sci. & Tech. 26, 1 (2005)
Jitter simulated in the digital domain.
Conclusions: uncorrelated jitter inaudible below 250nS on music signal.

Notice how *massive* are the levels of jitter that proved to be inaudible, and compare that to jitter levels in modern systems (typically below 1nS, good ones below 100pS). Jitter simply isn't the bogeyman it's made out to be.
Go to the top of the page
+Quote Post
Dirk95100
post Apr 1 2011, 09:48
Post #7





Group: Members
Posts: 33
Joined: 15-October 10
Member No.: 84639



Because cables are passive, they can not cause jitter.
They do affect jitter. This is because a cable is a lowpass filter and therefore the signal transition from low to high and visa versa becomes more uncertain. So jitter is increased, but no one can hear the artifacts.

Go to the top of the page
+Quote Post
Notat
post Apr 1 2011, 15:13
Post #8





Group: Members
Posts: 581
Joined: 17-August 09
Member No.: 72373



There have been some documented problems with clock recovery through these digital interfaces. Much of it has now been addressed through more sophisticated clock recovery by receivers and improved DACs with greatly reduced jitter sensitivity. I believe the current consensus is that things are good now.

It is, of course, an unprovable assertion to say that no one can hear any flaws in these systems. Certain listeners will always claim to hear differences. From there you have a long road to hoe to verify that's true and, if so, to determine what they're actually hearing.
Go to the top of the page
+Quote Post
Dirk95100
post Apr 1 2011, 15:32
Post #9





Group: Members
Posts: 33
Joined: 15-October 10
Member No.: 84639



QUOTE (Notat @ Apr 1 2011, 15:13) *
It is, of course, an unprovable assertion to say that no one can hear any flaws in these systems. Certain listeners will always claim to hear differences. From there you have a long road to hoe to verify that's true and, if so, to determine what they're actually hearing.


Jitter raises the noise floor. So it can degrade a 20 bit converter into a converter with 16 bit noise floor.
But listening tests have shown that the amount of jitter that is audible is a factor 1000 times higher than my theoretical example. Most converters are close to the theoretical limmit. So we can safely say: No one can hear it.
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Apr 1 2011, 15:57
Post #10





Group: Members
Posts: 3700
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (Dirk95100 @ Apr 1 2011, 04:48) *
Because cables are passive, they can not cause jitter.


Since a large number of things can be called "cables", this is like saying that no passive component can cause jitter, which is false.

There is an *unwritten* law of electrical engineering that says "Every circuit is, in the final analysis, a low pass filter." That is because it takes unbounded energy to power a real-world circuit that has an unbounded hf bandpass.

QUOTE
They do affect jitter. This is because a cable is a low pass filter and therefore the signal transition from low to high and visa versa becomes more uncertain. So jitter is increased, but no one can hear the artifacts.


Slowing down signal transitions most definitely does not aLWAYS cause jitter. In fact most digital outputs include low pass filters so that they will pass legal requirements for EMI, such as FCC Part 15. Slowing down transitions may make a link more susceptible to jitter, but the jitter has some other direct cause.

Dunn suggests some boundaries that relate to this in a 1993 AES paper about jitter due to digital interfaces.

Saying that jitter can never be audible is false. I've built passive circuits that interacted with certain commercial digital audio gear to add various amounts of jitter to digital audio signals. Extremely large amounts of jitter could be added, and they could be easily heard because jitter is FM distortion which is also known as vibrato in music.

I was able to make good recordings sound undeniably wavery if the equipment receiving the signal lacked modern levels of jitter resistance. I have an older Denon DAC that has almost no jitter resistance. In contrast a more modern digital audio decoder based on technology commonly used in modest but modern surround receivers reduced the added jitter to very, very low levels. The addition of jitter could not be detected in its audio output.

Jitter is one of those things that can be reduced as much as is needed provided that the jitter is not so bad that the signal cannot be received at all. CD players are a good example of this. The digital signal that comes off the disc is usually full of jitter, but a fairly simple buffering arrangement reduces it as low as is desired.

This post has been edited by Arnold B. Krueger: Apr 1 2011, 16:06
Go to the top of the page
+Quote Post
googlebot
post Apr 1 2011, 17:32
Post #11





Group: Members
Posts: 698
Joined: 6-March 10
Member No.: 78779



QUOTE (Arnold B. Krueger @ Apr 1 2011, 15:57) *
Saying that jitter can never be audible is false. I've built passive circuits that interacted with certain commercial digital audio gear to add various amounts of jitter to digital audio signals. Extremely large amounts of jitter could be added, and they could be easily heard because jitter is FM distortion which is also known as vibrato in music.


I doubt that jitter generally causes a vibrato type of effect. Jitter testing is just usually done with periodic signals. Of course, one can catch periodic signals over a wire, as in a ground loop, but that's just one possibility out of many.

This post has been edited by googlebot: Apr 1 2011, 17:33
Go to the top of the page
+Quote Post
lessthanjoey
post Apr 1 2011, 19:29
Post #12





Group: Members
Posts: 17
Joined: 23-March 11
Member No.: 89226



QUOTE (googlebot @ Apr 1 2011, 09:32) *
QUOTE (Arnold B. Krueger @ Apr 1 2011, 15:57) *
Saying that jitter can never be audible is false. I've built passive circuits that interacted with certain commercial digital audio gear to add various amounts of jitter to digital audio signals. Extremely large amounts of jitter could be added, and they could be easily heard because jitter is FM distortion which is also known as vibrato in music.


I doubt that jitter generally causes a vibrato type of effect. Jitter testing is just usually done with periodic signals. Of course, one can catch periodic signals over a wire, as in a ground loop, but that's just one possibility out of many.


Jitter caused by crappy clock-recovery circuits is signal dependent and causes variation in the DAC word clock. This causes signal dependent frequency modulation.

(But as stated by others, it's not audible unless your equipment is horribly broken crap)
Go to the top of the page
+Quote Post
Ethan Winer
post Apr 1 2011, 19:48
Post #13





Group: Members
Posts: 248
Joined: 12-May 09
From: New Milford, CT
Member No.: 69730



QUOTE (DVDdoug @ Mar 31 2011, 13:37) *
Ethan Winer has an article where discusses jitter. (He's also got a lot of other information debunking audiophile nonsense.)


Thanks for the referral. There's even more such information and audio examples in my hour-long AES Audio Myths video.

--Ethan


--------------------
I believe in Truth, Justice, and the Scientific Method
Go to the top of the page
+Quote Post
krabapple
post Apr 1 2011, 21:55
Post #14





Group: Members
Posts: 2234
Joined: 18-December 03
Member No.: 10538



don't forget, audibility of jitter is spectrum-dependent too.

http://www.avsforum.com/avs-vb/showpost.ph...p;postcount=163
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Apr 1 2011, 21:59
Post #15





Group: Members
Posts: 3700
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (googlebot @ Apr 1 2011, 12:32) *
QUOTE (Arnold B. Krueger @ Apr 1 2011, 15:57) *
Saying that jitter can never be audible is false. I've built passive circuits that interacted with certain commercial digital audio gear to add various amounts of jitter to digital audio signals. Extremely large amounts of jitter could be added, and they could be easily heard because jitter is FM distortion which is also known as vibrato in music.


I doubt that jitter generally causes a vibrato type of effect.


In the real world of digital audio, that would be a trusim because jitter generally has no audible effects at all. ;-)

QUOTE
According to Zwicker and Fastl, audible FM distortion usually manifests itself as "roughness". The vibrato effect only exists when the modulating frrequency is relatively low.

QUOTE

Jitter testing is just usually done with periodic signals.


As I read the papers, not only is the modulating test signal periodic, it is often a pure tone. This facilitiates characterizing and probably optmizes the sensitivity of the ear.

QUOTE
Of course, one can catch periodic signals over a wire, as in a ground loop, but that's just one possibility out of many.


I measured jitter in a goodly number of sitautions for my old PCAVTech web site. Real world jitter often involved a modulating frequency that was related to the local power line frequency. Other common modulating frequencies were related to data block sizes. Another set of possibilties related to the frequency of the audio signal being transmitted.

In Dunn's 10/1992 (not 1993 as I erroneously mentioned above) AES Conference paper Figure 9 suggests that the ear is much more sensitive to jitter with frequencies below 250 Hz . This tends to agree with Zwicker and Fastl.


Go to the top of the page
+Quote Post
Speedskater
post Apr 1 2011, 22:23
Post #16





Group: Members
Posts: 138
Joined: 12-December 07
From: Cleveland, OH
Member No.: 49500



QUOTE (Ethan Winer @ Apr 1 2011, 14:48) *
QUOTE (DVDdoug @ Mar 31 2011, 13:37) *
Ethan Winer has an article where discusses jitter. (He's also got a lot of other information debunking audiophile nonsense.)

Thanks for the referral. There's even more such information and audio examples in my hour-long AES Audio Myths video.
--Ethan

The people that want to believe that jitter is a very big problem, won't watch an hour (or even a few minutes) long video.
And it's hard to quote a video in a response.


--------------------
Kevin
Go to the top of the page
+Quote Post
DVDdoug
post Apr 2 2011, 00:50
Post #17





Group: Members
Posts: 2566
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



That's a different subject... We were just trying to answer your question...

biggrin.gif If someone "wants to believe something" and they don't have an open mind... Well, what are you gonna' do? biggrin.gif

crying.gif Unfortunately, the "audiophile community" seems to be dominated by people like that.

The video is for people who have an interest in that kind of thing, or for people who are curious. I really enjoyed it!
Go to the top of the page
+Quote Post
Notat
post Apr 2 2011, 15:48
Post #18





Group: Members
Posts: 581
Joined: 17-August 09
Member No.: 72373



QUOTE (Dirk95100 @ Apr 1 2011, 08:32) *
Jitter raises the noise floor. So it can degrade a 20 bit converter into a converter with 16 bit noise floor.

That is the case for a well-behaved converter fed random jitter (phase noise). Modern converters are generally well-behaved. This was not always the case. Not all jitter is purely noisy.
Go to the top of the page
+Quote Post
Ethan Winer
post Apr 2 2011, 15:50
Post #19





Group: Members
Posts: 248
Joined: 12-May 09
From: New Milford, CT
Member No.: 69730



QUOTE (Speedskater @ Apr 1 2011, 17:23) *
The people that want to believe that jitter is a very big problem, won't watch an hour (or even a few minutes) long video.


Absolutely, you can't change the mind of a believer no matter how compelling your evidence. But I mostly don't write (or make videos) for them. I write for people who ask legitimately if the various advice they read over and over in forums and magazines are true. A lot of people genuinely want to know if they should worry about jitter, or power "conditioner" products etc.

--Ethan


--------------------
I believe in Truth, Justice, and the Scientific Method
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Apr 4 2011, 16:04
Post #20





Group: Members
Posts: 3700
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (Speedskater @ Apr 1 2011, 17:23) *
QUOTE (Ethan Winer @ Apr 1 2011, 14:48) *
QUOTE (DVDdoug @ Mar 31 2011, 13:37) *
Ethan Winer has an article where discusses jitter. (He's also got a lot of other information debunking audiophile nonsense.)

Thanks for the referral. There's even more such information and audio examples in my hour-long AES Audio Myths video.
--Ethan

The people that want to believe that jitter is a very big problem, won't watch an hour (or even a few minutes) long video.
And it's hard to quote a video in a response.


Many of tht people who think jitter is a big problem, don't realize that the analog world was chock-full of it. If you want a real thrill, calculate the jitter in ps that corresponds to 0.05% flutter and wow which is a SOTA number for analog tape and vinyl.
Go to the top of the page
+Quote Post
pdq
post Apr 4 2011, 17:11
Post #21





Group: Members
Posts: 3392
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



As a back-of-the-envelope calculation, 1.8 seconds (33.3 RPM) times 0.05% equals nine hundred million picoseconds. smile.gif
Go to the top of the page
+Quote Post
Ethan Winer
post Apr 4 2011, 19:17
Post #22





Group: Members
Posts: 248
Joined: 12-May 09
From: New Milford, CT
Member No.: 69730



QUOTE (Arnold B. Krueger @ Apr 4 2011, 11:04) *
If you want a real thrill, calculate the jitter in ps that corresponds to 0.05% flutter and wow which is a SOTA number for analog tape and vinyl.


QUOTE (pdq @ Apr 4 2011, 12:11) *
As a back-of-the-envelope calculation, 1.8 seconds (33.3 RPM) times 0.05% equals nine hundred million picoseconds. smile.gif


Indeed. But remember that analog timing errors are always acceptable because, well, because they're analog! laugh.gif

--Ethan


--------------------
I believe in Truth, Justice, and the Scientific Method
Go to the top of the page
+Quote Post
AliceWonder
post Apr 9 2014, 18:01
Post #23





Group: Members
Posts: 119
Joined: 13-July 12
From: California
Member No.: 101393



QUOTE (DVDdoug @ Apr 1 2011, 16:50) *
That's a different subject... We were just trying to answer your question...

biggrin.gif If someone "wants to believe something" and they don't have an open mind... Well, what are you gonna' do? biggrin.gif

crying.gif Unfortunately, the "audiophile community" seems to be dominated by people like that.

The video is for people who have an interest in that kind of thing, or for people who are curious. I really enjoyed it!


I've only watched a little bit so far but already have enjoyed it.

The very beginning where they explain how listening for certain things steers your hearing - that is exactly why double-blind tests are so critical. The way our brains work, we often will hear things differently and be honest when we say so, even if the audio is identical, if we are listening for a specific thing, which we tend to do when the listening test is not blind.
Go to the top of the page
+Quote Post
Martel
post Apr 10 2014, 08:35
Post #24





Group: Members
Posts: 553
Joined: 31-May 04
From: Czech Rep.
Member No.: 14430



In case someone is so worried about jitter, why not use line-out -> line-in?


--------------------
IE4 Rockbox Clip+ AAC@192; HD 668B/HD 518 Xonar DX FB2k FLAC;
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 28th August 2014 - 00:49