Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Jitter question (Read 13266 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Jitter question

Hi All,

I know this topic has been touched many times here and on other forums as well, not even speaking about the many articles that can be found online, but the more I read, the more confused I get. 

If I am understanding things correctly, jitter basically is a timing issue: the sender transmits pulses (bits) in another tempo than the receiver expects them to be.  I know this is a very simplified view.

I guess my basic question is why jitter still is an issue?  We have been living in a world (for quite a while now) where, for example, processors can communicate with other devices, transmitting data back and forth at rates that are much, much higher than a relatively slow audio-stream without a glitch, but when it comes down to offering a DAC-chip the correct timing and bits, things start to get troublesome. 

I anyone can give some "not-too-technical" explanation for this I would really appreciate it!

Thanks,
Peter

Jitter question

Reply #1 –
I'm not the one to offer the best explanation, but at least I would like to correct de description.

jitter is not  a clock deviation, but a clock swing. (i.e. not a different frequency, but an oscillation of the frequency over time).
In most cases, the measured swings are quite small, much lower than audibly detectable.

I've found this thread form 2006 with almost your question, but maybe the answer is not complete: http://www.hydrogenaudio.org/forums/index....showtopic=49795


Anyway, the idea is that a digital system does not need to be at an exact clock and is able to retrasmit if necessary. In fact, the jitter problem could be considered moot in any transmission that buffers on the receiver end. (With a couple or so samples would be enough).

Jitter question

Reply #2 –
Quote from: thesurfingalien on
If I am understanding things correctly, jitter basically is a timing issue: the sender transmits pulses (bits) in another tempo than the receiver expects them to be.  I know this is a very simplified view.


You have things right. Note that jitter is an inteface problem, it only actually exists in ADCs and DACs, In the digital domain there is a real, often manifest possibility of zero added distortion and noise.


Quote
I guess my basic question is why jitter still is an issue?


If you are talking jitter related to the digital domain, the answer is that is tha jiiter is not a serious problem in gear with reasonable quality (mid-fi and better) and never was, The original CD players had good jitter performance by modern standards. Jitter in audiophile CD players did not become a problem until they *improved* them by using external DACs.

Quote
We have been living in a world (for quite a while now) where, for example, processors can communicate with other devices, transmitting data back and forth at rates that are much, much higher than a relatively slow audio-stream without a glitch, but when it comes down to offering a DAC-chip the correct timing and bits, things start to get troublesome.


No. Audible jitter related to digital equipment exists only when there are serious mistakes and incompetency.

A few months back I looked at the specs of the best analog tape machines that have ever existed. This was the best sound we had before digital.

While audiophiles are obsessing over dozens or hundreds of picoseconds of jitter, the best analog tape machines ever made had millions of picoseconds of jitter. Millions! Now the nature of that jitter may have been a little different, but that can't overcome the fact that analog tape is, at the best, thousands of time worse than even mediocre digital gear.

If jitter is such a problem in digital equipment, why aren't all these audiophiles running screaming, with blood trickling out of their ears, when they listen to analog playback? We now know that up to 50% of all SACDs and DVD-A titles that were ever released came from legacy sources, including analog master tapes. Why aren't people complaining about all the audible jitter in their new SACDs amd DVD-As?

The simple fact of the matter is that jitter associated with digital equipment was always an overblown issue I'm not saying that there is *no* digital equipment with audible jitter, but I'm telling you that you have to look long and hard to find it among equipment with any pretenses of quality at all.

Quote
I anyone can give some "not-too-technical" explanation for this I would really appreciate it!


The non-technical explanation is that fears of jitter associated with digital audio gear has always been about hype and fear-mongering. The audiopile ragazines have been glorifying analog equipment with millions of picoseconds
of jitter in analog gear and recordings while teaching people to obsess over at most something like a thousand  picoseconds of jitter in digital gear. 

If  jitter were the problem that they pretend it to be, they should have been ranting and raving about the baked-in jitter in all the SACDa, LPs, and DVD-As that were reissues of legacy analog recordings.

We just witnessed a sort of a blind test of the self-proclaimed golden-eared audiophile reviewers. In many cases they did not know that these so-called hi rez recordings were, according to their standards for digital gear, fatally flawed. They were blinded to sighted evidence of this purported problem. Yet they ooohed and ahhhed over them and made them out to be the second coming. 

Just goes to show where sighted bias leads you - nonsense and self-contradiction.

Jitter question

Reply #3 –
Look at it this way... When digital pulses are being transmitted, they are temporarily analog. Each pulse has a starting time and an ending time, and the transition between 0 and 1 is not instantaneous.

At the receiving end these pulses must be converted from analog back to digital, through the use of a voltage comparator. This comparator has inherent noise, and noise may be picked up in the act of transmission as well. The combination of voltage noise plus the finite rise and fall times of the pulses results in timing jitter.

Jitter question

Reply #4 –
Quote from: pdq on
Look at it this way... When digital pulses are being transmitted, they are temporarily analog. Each pulse has a starting time and an ending time, and the transition between 0 and 1 is not instantaneous.

At the receiving end these pulses must be converted from analog back to digital, through the use of a voltage comparator. This comparator has inherent noise, and noise may be picked up in the act of transmission as well. The combination of voltage noise plus the finite rise and fall times of the pulses results in timing jitter.



The above explanation is so good that I used it to build a "jitterizer" that worked well. I used a high speed comparator to turn the input digital audio signal into a square wave, followed by a low pass filter that turned the square wave into a triangular wave. I mixed in various signals that were at a far lower frequency and of a lower amplitude. I then used another high speed comparator to recreate the square wave that was the jitterized digital audio signal.. The low frequency signals I mixed in shifted the decision points of the second comparator.  The end result was a digital audio signal with variable frequency, variable amplitude jitter that was just about completely under my control.  Analysis of the sidebands created by the resulting FM distortion confirmed that my jitterizer was working as desired. I had two DACs on hand, one that was highly susceptible to jitter and one that was almost completely resistant to it.

Jitter question

Reply #5 –
While the responses explain some things about jitter, and correctly proclaim its unimportance in any reasonable audio gear, they don't really address the question: Why does jitter still exist in today's world?

You might as well ask why quantum uncertainty still exists. It is an intrinsic part of this universe. Digital sampling depends upon a clock, which is to say, upon measuring that which you are sampling at specified intervals. There is a limit to the precision at which the intervals themselves may be measured or generated. Any variation in the interval between taking one sample, taking the next, taking the next, is jitter.

I don't know the numbers, but even the best time standard, the atomic clock at the national bureau of standards (or perhaps some experimental improvement I don't know about) is a statistical determination. There is some amount of jitter in the best time measurement stream.

In audio, the timing variations in sampling, in the DAC, are a non-correctable part of the audio data created by the DAC. The timing variations in the ADC add to the jitter inherent in the data from original sampling time. This final step produces the analogue signal that you will hear, so its added jitter is non-correctable (but will, of course, be different the next time the same audio data is played)

Any jitter added between DAC and ADC can be eliminated (to the limit of the ADC's inherent jitter) even if made exceptionally great, as in Arnold's experiments. Think of this in the extreme case: The digital audio data starts from where ever it is created or stored, is transmitted over a very jittery circuit, and is recorded on a hard drive (or any other kind of stable media).

The audio data was not changed in the transmission, only time interval between samples varied. Later, play the audio data from its storage locations, as millions of people do ever second of every day from their computers, optical media, personal players, etc.. Only the samples themselves are stored, along with a single control that tells the output circuit the sample rate. The transmission interval variations, the added jitter, are not part of the record and thus cannot effect the output.

In a real time playback, without that intermediate hard stop, various means can be employed to markedly reduce or eliminate the transmission jitter, including the aforementioned buffer. Whether or not it is completely eliminated depends on the particular playback equipment, but virtually all such equipment does at least an adequate job.

CHAOS investigations have shown there are random variations in any process. There are also complex patterns in these variations. Those patterns frequently, or generally, change abruptly with conditions, but the change is usually only a matter of scale. Feedback controls can determine the change points but not eliminate the variations beyond intrinsic limits.

Some of the conditions in audio gear that are never completely controllable, and thus produce jitter during sampling or playback, are temperature and voltage. Voltage regulation can be very elaborate but the regulating components can never be beyond the influence of the rest of the universe nor of their own intrinsic structure. Temperature control "ovens" can be used around critical components, but heat must be supplied and heat will escape, and there is some limit below which the controls cannot respond.

Jitter question

Reply #6 –
@All,

Thanks (so far) for your replies.  They have been more understandable for me than most other information I read.

So, if I understand correctly, we are talking about 3/4 different types of jitter (again, put very simple):

A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.

B: jitter caused by a difference in time between clock pulses (not a steady beat) of the clock that controls the DAC chip

C: jitter caused by a clock that is overall too slow or too fast (beat may be steady, but time between beats may be too long or too slow) of the clock that controls the DAC chip.

D: jitter caused by unclear (distorted) pulses in the analog signal transfer (the receiving end has trouble determining weather the analog pulse represents 1 or 0)

For B and C (maybe to be regarded as sort of the same)...  Is this behavior not sometimes referred to as digital "wow / flutter"?  If so, I have read this more, but the quarter never fell 

Please correct me if I did not understand correctly.

Many thanks,
Peter

Jitter question

Reply #7 –
Quote from: thesurfingalien on
@All,

Thanks (so far) for your replies.  They have been more understandable for me than most other information I read.

So, if I understand correctly, we are talking about 3/4 different types of jitter (again, put very simple):

A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.

B: jitter caused by a difference in time between clock pulses (not a steady beat) of the clock that controls the DAC chip

C: jitter caused by a clock that is overall too slow or too fast (beat may be steady, but time between beats may be too long or too slow) of the clock that controls the DAC chip.

D: jitter caused by unclear (distorted) pulses in the analog signal transfer (the receiving end has trouble determining weather the analog pulse represents 1 or 0)

For B and C (maybe to be regarded as sort of the same)...  Is this behavior not sometimes referred to as digital "wow / flutter"?  If so, I have read this more, but the quarter never fell


Jitter is always about timing variations while data is being handled. Jitter is older than digital, and has been around at least as long as man has been writing and reading. The first recorded instance of jitter was when the weekly letter from the king was delivered a day late. IOW, it was supposed to show up every Monday, but one week it arrived on Tuesday.  This made the king's faithful followers jittery. ;-)

Audible jitter has been around as long as the phonograph. You can call it flutter, and you can call it wow, and you might even call a speed variation, but it is all jitter.

As you point out in (1) jitter can cause problems in a data link. Timing variations in a data link are only a problem when they cause data to be received incorrectly. If the timing of the samples is so far off that the data can't be properly recovered from them, then you have a problem. This sort of problem is actually very common, and the means for addressing it have been discussed.

(B) is correct but incomplete, because clock jitter can affect data conversion by both ADCs and DACs.

IMO © should be (A) because it is the root cause for everything that follows.

(D) is an effect of jitter. Jitter in a data link or during the reading of properly-recorded media can cause data to be received in error. Jitter during the recording of digital media can cause the media to be permanently wrong.

In general the effects of jitter can be almost completely nullified unless it is so bad that there are unrecoverable data errors.

As has been pointed out, jitter is a consequence of living in a world that is full of undesired influences. 

To show how well jitter can be nullified, we have the humble CD player. For a variety of reasons the digital data stream generally comes out of the optical pickup with a large amount of jitter.  The data is read into a buffer as at whatever rate it is actually read, and clocked out of that buffer with a steady clock. As long as there is data in the buffer, the process works well. CD players have a connection between the data buffer and the disc spindle motor. Whenever the buffer gets too empty, the motor is instructed to run faster. Whenever the buffer is in danger of getting too full, the motor is instructed to slow down.  Just about every other digital medium and data link can be viewed as if it was a variation on this basic scheme.

Another example is playing a digital music file on a PC. The data is supplied to the audio interface in blocks of data - a clear example of data showing up in fits and spurts.  Jitter is maximized.There is a data buffer in the audio interface that provides audio data to the DAC. The DAC is clocked by a stable oscillator. When the buffer gets too empty the PC is instructed to obtain more data from the hard drive or other storage. When it gets too full, requests for data are throttled back. One of the key assumptions is the idea that the PC can supply data to the audio interface on very short notice,  whenever needed.

Jitter question

Reply #8 –
Quote from: thesurfingalien on
A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.

This is only half right. In the case of a USB DAC, there is jitter in the DAC's clock, but the PC is not a source of jitter.

Jitter question

Reply #9 –
Quote from: pdq on
Quote from: thesurfingalien on
A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.

This is only half right. In the case of a USB DAC, there is jitter in the DAC's clock, but the PC is not a source of jitter.


But then what about the clock in the USB port in the computer?  Or is it always the receiving end that is "responsible", so to speak.

Thanks,
Peter

Jitter question

Reply #10 –
There are basically two ways of sending audio data to a DAC. One way the sender controls the data rate, as in S/PDIF and the receiver acts as sort of a slave device.

The other way the receiver has an internal buffer and the sender only sends data as needed to prevent over/underflow of the buffer. The latter describes (most) USB DACs as well as CD players.

Jitter question

Reply #11 –
Quote from: pdq on
Quote from: thesurfingalien on
A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.

This is only half right. In the case of a USB DAC, there is jitter in the DAC's clock, but the PC is not a source of jitter.


There are a number of different methods that USB DACs use to manage their data link with the PC.

Synchronous
Adaptive
Asynchronous

The Aysnchronous method most closely maps to the way that CD players and PCI PC sound cards work.

This article explains this in more detail:

One of many USB DAC articles

Jitter question

Reply #12 –
I tend to forget that the high-end audio market, due to its low volume, tends to cut corners in product development. 

Jitter question

Reply #13 –
Quote from: thesurfingalien on
A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.


I probably should have written:

A: jitter caused by clock differences between sending and receiving end,  for example a sending PC and a receiving USB DAC

Jitter question

Reply #14 –
On any properly designed digital audio playback device the digital data is buffered (stored in memory) before being fed into the converters, so any clock jitter and/or speed differences are eliminated and only the clock in the final digital to analog converter is relevant to playback.

I've experienced a poorly buffered audio playback device before and it's not pretty or subtle. It was a original Sound Blaster Live! which when combined with a motherboard using a certain Via chipset could render the audio a garbled, scratchy mess. The motherboard chipset let the AGP slot gobble up all the bandwidth on the bus which would starve the sound card of audio data.

Of course, you'll still have whatever clock jitter was present in the source ADCs clock present, but there's nothing you can do about that so you might as well just listen to the music and enjoy it.

Jitter question

Reply #15 –
@All,

I am just asking these questions because I want / need to get correct information on this topic as I am working on a (as far as possible) non-technical explanation (for a Dutch website) of jitter so everybody can at least understand the basics and understand why there is nothing to worry about. 

If it were not for that, I would not waste a minute thinking about these effects. 



Jitter question

Reply #16 –
I have no practical experience on this subject, but as I recently browsed book about jitter, noise and signal integrity, IMHO pdq's post #4 and first reply in [JAZ]'s linked topic says it nicely in term that layman can understand. If you want to go further in various types of jitter then their symmetry with noise (as statistical phenomenons) or influence on medium type etc then you are going couple of levels further and would not gain anything for general non-technical report IMHO

Jitter question

Reply #17 –
For some perspective on the amount of jitter that audiophiles have been putting up with for decades, check out this post and the one after it. By comparison, the jitter resulting from S/PDIF transmission is miniscule.

Jitter question

Reply #18 –
Quote from: thesurfingalien on
Quote from: thesurfingalien on
A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.


I probably should have written:

A: jitter caused by clock differences between sending and receiving end,  for example a sending PC and a receiving USB DAC


One associated item about this statement about data links is that if a clock difference  between sending and receiving end persists, then there is data loss which manifests itself audibly as clicks, thunks, and pops.

So far we haven't said a lot about what jitter sounds like.

If the jitter is random then it manifests itself as a heightened noise floor. To put this into perspective, the noise floor that is inherent in 16 bits is low enough that it is rarely if ever audible. People have thundered about jitter so bad that it actually raises the 16 bit noise floor.  This represents pretty severe jitter, but it is still low enough to have a very, very low probability of actually being audible.

If the jitter is not random then it is almost always periodic. Periodic jitter affects musical sounds differently depending on the  nature of the musical sound, and also the nature of the jitter.

Low frequency jitter is best known as wow and we commonly hear it from LPs and analog tape. Wow is a self-describing word, as wow it makes steady tones go woo-uhh-woo-uhh-woo, intstead of the desired wooooooo.

Mid frequency jitter is again best known from the world of analog recodring  as flutter. It turns steady tones into tones with vibrato. Another word that is used to describe vibrato is roughness.

High frequency jitter makes music sound sour. It actually adds tones that are sufficiently displaced in the frequency domain from the original tone that they are not masked. These added tones are generally non-harmonic and so they add a slightly out-of-tune flavor to the music.

Actually hearing jitter is easiest when you can compare music with no jitter to music with the jitter that you are investigating. This is because the unpredictable and varied nature of the way that jitter can affect the sound of the music. Jitter can have have at lkeast 5 different sonic signatures:


(1) Clicks and pops
(2) Wow
(3) flutter, vibrato or roughness
(4) sourness
(5) a raised noise floor

Until you have your reliable diagnosis, you don't know which one or which ones are going to be your problem.

Jitter question

Reply #19 –
Quote from: Arnold B. Krueger on
Quote from: thesurfingalien on
Quote from: thesurfingalien on
A: jitter caused by clock differences on both sending and receiving end,  for example a sending PC and a receiving USB DAC.


I probably should have written:

A: jitter caused by clock differences between sending and receiving end,  for example a sending PC and a receiving USB DAC


One associated item about this statement about data links is that if a clock difference  between sending and receiving end persists, then there is data loss which manifests itself audibly as clicks, thunks, and pops.



Slightly off-topic, but last week I spend some time arguing with someone that wrote a review about "Audiophile Grade" USB cables on his audio-website.  His conclusion was that they all had a different sound when used on his USB DAC.  Obviously, this person had not done any ABX / DBT to support his claims.  As the used USB DAC uses Adaptive Isochronous Transfer, which carries CRC data, packets that have errors get dropped, causing exactly the audible result you just described, not a "warmer" or a "fresh" sound.  I got him to do a test with a USB data analyzer on all cables, testing data-transmission to a USB-harddisk to show the unlikeliness of errors occurring in the first place.  After I and some other people also persuaded him to perform an ABX, the expected result came out, and he consequently removed his previous article, replacing it with an article describing his new insight and warning people not to buy into the USB cable myth.

Jitter question

Reply #20 –
Quote from: thesurfingalien on
Slightly off-topic, but last week I spend some time arguing with someone that wrote a review about "Audiophile Grade" USB cables on his audio-website.  His conclusion was that they all had a different sound when used on his USB DAC.  Obviously, this person had not done any ABX / DBT to support his claims.  As the used USB DAC uses Adaptive Isochronous Transfer, which carries CRC data, packets that have errors get dropped, causing exactly the audible result you just described, not a "warmer" or a "fresh" sound.  I got him to do a test with a USB data analyzer on all cables, testing data-transmission to a USB-harddisk to show the unlikeliness of errors occurring in the first place.  After I and some other people also persuaded him to perform an ABX, the expected result came out, and he consequently removed his previous article, replacing it with an article describing his new insight and warning people not to buy into the USB cable myth.


Good job!

Jitter question

Reply #21 –
Quote from: Arnold B. Krueger on
Quote from: thesurfingalien on
...


Good job!


Thanks!  I got help from others in the technical explanation department...

Anyway, I think most of the credits should go to the guy who wrote the article.  He is a rather public figure in The Netherlands and announcing publicly that  spending EUR 300 on an AudioQuest USB cable is a waste is pretty gutsy!

I recall AQ to be a rather respectable brand that did not have too much BS products, but if you want to have a good laugh, visit their website and pay special attention to the power-cords... 

@All:

Thanks for the clear and understandable replies!  I will complete my "article" in Dutch, translate it to English and post it here for final remarks (if you do not get sick of me and my questions :-))

Jitter question

Reply #22 –
Quote from: thesurfingalien on
I recall AQ to be a rather respectable brand that did not have too much BS products, but if you want to have a good laugh, visit their website and pay special attention to the power-cords...


IME AQ is so far out that they make MC look almost sane. ;-)

Jitter question

Reply #23 –
I have two quick questions...

Do I understand correctly that clock jitter (the picosecond value) refers to the inaccuracy between clock-pulses, and it can go either way?

What would be an average ps for normal, consumer-grade CD-players...  Something between 500 - 1000 ps?

Thanks,
Peter

Jitter question

Reply #24 –

I have read in some postings on audiophile sites - and even on diy audio, a slightly saner forum usually - that various pc players like foobar, VLC,  media monkey, j rivers, winamp, db poweramp etc. sound different, the difference explained by "jitter" introduced by the player. Is that possible, and what would be the mechanism?.