Something about " reclock " & "dejitter ", which I couldn't comprehend...

Preston8452

Member
Jan 5, 2022
12
2
5
33
Hi y'all,

Firstly, sorry to bring out this old question again, but recently I just received this inquiry from my friend who is also on his way to digital streaming though, as we were discuss whether a network switch would help or not like everyone else, he sorta struck me with this, " Based on some researches, I found that most hi-end or high quality digital audio devices would reclock and dejitter signals on the arrival anyway, so what's the possible benefit of doing that in a network switch prior to arrival? "

I was speechless lol, because I didn't know about this, and I bough a network switch myself actually, so I seriously would like to prove my point to him!

If you guys have some insightful opinions regarding this question, please do share, I'm dying to learn...

Best,
 

BlueFox

Member Sponsor
Nov 8, 2013
1,709
406
405
How does the signal receiver know if the signal has been previously affected by jitter? How does it ‘dejitter’ the received signal?
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
Most DACs include asynchronous clock recovery and data buffering to isolate the actual DAC inside the box from the incoming data stream. This isolates the DAC itself (the thing inside the box that does the actual digital to analog conversion) from any jitter on the incoming data. A phase-locked loop (PLL) with a separate reference clock is used to pull the data from the internal data buffer and provide a "clean" stream to the DAC. The PLL is designed to lock onto a clock and filter out any noise outside its bandwidth. It has been done this way for many years, and for systems well beyond audio. (For example, the PCIe and SAS/SATA busses inside your computer that handle billions of bits per second with a bit error rate per spec of better than 1e-12, -240 dB error rate if you express it that way.) This applies to any data coming into the DAC no matter the input (USB or Ethernet). About twenty years ago when I first started looking more deeply into audio DACs, and especially as HDMI band USB became popular, it was obvious that the incoming data stream had substantial jitter, and different DACs did better or worse jobs of rejecting that jitter. Since then, buffering and reclocking has become universal AFAIK. It is included in all the modern chip sets and has been for many years now.

If there was jitter imposed upon the source, e.g. by the ADC used to convert the original analog signal to digital, then that is part of the data and the DAC cannot do anything about that. But neither will anything else -- it's in the recording.

Ethernet data is packetized -- it comes in "bursts" that include the data and a bunch of additional info for framing and error correction, and any Ethernet link adheres to that standard. That means the data ultimately provided to the DAC is recovered, buffered, error-corrected, reclocked, and all that jazz so there is essentially no way for jitter on the Ethernet wires to reach the DAC. The Ethernet receiver must capture all those bursts, strip off the framing and other protocol information, do the error recovery (and request another packet if it cannot recover the data), create a continuous stream of data, and deliver the recovered data bits to the DAC. Ethernet also includes galvanic isolation per the standard so there is normally no chance of forming a ground loop or introducing ground noise from the Ethernet cable. Not to say some bad designs might mess that up, but it is very rare IME. We use Ethernet all over at work to connect million-dollar test equipment far more sensitive and with much greater bandwidth than our hearing and it all works just fine.

That said, I have seen noise added by some power over Ethernet (PoE) devices that did not properly implement the standard. As PoE becomes more common, those "bad" devices are becoming less common, natch, and most consumer switches do not include PoE as it is still pretty rare in the home AFAIK.

HTH - Don
 

rando

Well-Known Member
Sep 22, 2019
1,701
1,234
245
Online
In some ways it represents the same challenge as finding your way inside a bake sale to select blue ribbon winners. Reaching your personal best in show with palate and stomach heightened by a combination of decisions that ensures immense enjoyment at the high end of your ability to consume. There is an almost chemical audio reaction taking place along the trail of what stands in for tables and sales pitches encountered. A single vigorous handshake or thump on the back can upset the works. :p

Sometimes you need to feed a problem. Force out all but that which makes the act itself enjoyable and relatable. Slow down and make time to see what a few sittings inspire in the way of important decisions. It's hard work building up to those bake sales you hope to bring something home from. Even a simpler grasp something can be sugar free but isn't free of sugars is better than losing out deciphering this manner of information to an inconclusive end.

 

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
Most DACs include asynchronous clock recovery and data buffering to isolate the actual DAC inside the box from the incoming data stream. This isolates the DAC itself (the thing inside the box that does the actual digital to analog conversion) from any jitter on the incoming data. A phase-locked loop (PLL) with a separate reference clock is used to pull the data from the internal data buffer and provide a "clean" stream to the DAC. The PLL is designed to lock onto a clock and filter out any noise outside its bandwidth. It has been done this way for many years, and for systems well beyond audio. (For example, the PCIe and SAS/SATA busses inside your computer that handle billions of bits per second with a bit error rate per spec of better than 1e-12, -240 dB error rate if you express it that way.) This applies to any data coming into the DAC no matter the input (USB or Ethernet). About twenty years ago when I first started looking more deeply into audio DACs, and especially as HDMI band USB became popular, it was obvious that the incoming data stream had substantial jitter, and different DACs did better or worse jobs of rejecting that jitter. Since then, buffering and reclocking has become universal AFAIK. It is included in all the modern chip sets and has been for many years now.

If there was jitter imposed upon the source, e.g. by the ADC used to convert the original analog signal to digital, then that is part of the data and the DAC cannot do anything about that. But neither will anything else -- it's in the recording.

Ethernet data is packetized -- it comes in "bursts" that include the data and a bunch of additional info for framing and error correction, and any Ethernet link adheres to that standard. That means the data ultimately provided to the DAC is recovered, buffered, error-corrected, reclocked, and all that jazz so there is essentially no way for jitter on the Ethernet wires to reach the DAC. The Ethernet receiver must capture all those bursts, strip off the framing and other protocol information, do the error recovery (and request another packet if it cannot recover the data), create a continuous stream of data, and deliver the recovered data bits to the DAC. Ethernet also includes galvanic isolation per the standard so there is normally no chance of forming a ground loop or introducing ground noise from the Ethernet cable. Not to say some bad designs might mess that up, but it is very rare IME. We use Ethernet all over at work to connect million-dollar test equipment far more sensitive and with much greater bandwidth than our hearing and it all works just fine.

That said, I have seen noise added by some power over Ethernet (PoE) devices that did not properly implement the standard. As PoE becomes more common, those "bad" devices are becoming less common, natch, and most consumer switches do not include PoE as it is still pretty rare in the home AFAIK.

HTH - Don
So simply said bits are bits right. No way Jose man and term jitter is not even one part of noise or the total effect of how the dac analog conversion will sound.
for voltage variation and potential from one device to the next is basically a ground loop.
when most claim isolation again fake news.
only galvanic , fiber or WiFi offer true isolation in one domain like ground loops , voltage variations
now some will say this ground or varying voltage potential does not matter.
cause this is digital but all we do in most all audio is voltage based manipulation so I think it can’t not affect things.
not all dacs inputs are reclocked. most newer dacs do reclock spidif but still may not usb
people like taiko em Eli have there systems setup to make sense of a minefield of issues to make there servers work well. Now the post above does have many truths but not all in the right context.
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
So simply said bits are bits right. No way Jose man and term jitter is not even one part of noise or the total effect of how the dac analog conversion will sound.
for voltage variation and potential from one device to the next is basically a ground loop.
when most claim isolation again fake news.
only galvanic , fiber or WiFi offer true isolation in one domain like ground loops , voltage variations
now some will say this ground or varying voltage potential does not matter.
cause this is digital but all we do in most all audio is voltage based manipulation so I think it can’t not affect things.
not all dacs inputs are reclocked. most newer dacs do reclock spidif but still may not usb
people like taiko em Eli have there systems setup to make sense of a minefield of issues to make there servers work well. Now the post above does have many truths but not all in the right context.
You are making a leap and attacking things well beyond what I said. Signal integrity of Gb/s systems is my day job so what the bits are doing is critical and we measure things like jitter (many types), noise, bit error rate, various types of distortion, and so forth. The question here was about the impact an Ethernet switch can have so that is all I addressed, not other noise sources. The way Ethernet works the DAC is isolated from the bit stream until the data is completely lost, which will result in quite audible issues like dropouts and such. But that is very rare; consider how many bits travel just to get this nice forum to display on your screen and how rarely it gets corrupted (and usually when it does it has nothing to do with the actual Ethernet network but some other failure).

Jitter from the Ethernet just has no way to get directly to the DAC; the data must be extracted and converted to a proper format (bit depth and rate) for the DAC itself to convert to analog. A typical Ethernet control chip uses about a 156 MHz clock IIRC, much higher than what your DAC needs even for high-res signals, and it is multiplied up to reach the Gb/s or so data rate over the Ethernet cable.

Ethernet is galvanically isolated per spec, so no ground or power connection to the receiver, unless the design is non-compliant. There can be other noise and leakage paths, of course. But again for it to get to the DAC's analog output implies a bad design.

I certainly do not know all DACs, but I would be really surprised if a USB DAC is not asynchronous. USB data comes in bursts and also has protocol-level data that must be stripped off to provide the final bits to the DAC.

A ground loop is a special case for noise conduction and just one of many. If you do not isolate the digital and analog grounds and supplies inside the DAC box, then you can get internal noise coupling no matter the external interface (Ethernet, USB, WiFi, whatever). And there can be radiated noise sources that find their way into the analog output by various devious means. Again, any good design should avoid that, but I do not know what you or anybody else is actually using and how well it is designed. Nor will an external switch or anything else help since it happens inside the box. So even when the bits are bits, which is the vast majority of the time, the magic in design is all about how you handle them once they get inside the box.
 
Last edited:

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
In some ways it represents the same challenge as finding your way inside a bake sale to select blue ribbon winners. Reaching your personal best in show with palate and stomach heightened by a combination of decisions that ensures immense enjoyment at the high end of your ability to consume. There is an almost chemical audio reaction taking place along the trail of what stands in for tables and sales pitches encountered. A single vigorous handshake or thump on the back can upset the works. :p

Sometimes you need to feed a problem. Force out all but that which makes the act itself enjoyable and relatable. Slow down and make time to see what a few sittings inspire in the way of important decisions. It's hard work building up to those bake sales you hope to bring something home from. Even a simpler grasp something can be sugar free but isn't free of sugars is better than losing out deciphering this manner of information to an inconclusive end.

I am not sure I quite followed the text, but the picture did bring back memories of my biology and organic chem classes in college, not the best of memories alas.
 

rando

Well-Known Member
Sep 22, 2019
1,701
1,234
245
Online
I am not sure I quite followed the text, but the picture did bring back memories of my biology and organic chem classes in college, not the best of memories alas.

For the best, number of edible references is on the rise this week where conversations intersect a well worn topic. Not everyone can stomach the full mix of networking and digital chain component behaviors all at once coming from analog or less technical backgrounds. Even fewer appear capable of letting pass a lower order explanation that might help that person make a positive connection. At risk of being vague, cordial to outright sweet responses offering longer sighted views are harder to offend with.

Funny that got the intended response from you. I think everyone has a subject like that.
 

AMR / iFi audio

Industry Expert
Aug 21, 2019
2,636
1,153
260
43
UK
ifi-audio.com
@Preston8452 it's true that many devices use internal reclocking. It's a powerful technique that really is one of the most important in a contemporary DAC. We use it too in our more advanced products. There is one thing that is important and answers your question. All those systems have a big, but limited ability to remove jitter. While it can really make engineers' lives easier, it's not a perfect solution for everything. There is even a scientific term for that: GIGO. Garbage In, Garbage Out :)
 

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
If bits were just bits and jitter stops at just isolation then why do many use re clockers anyway
ask empirical audio or ideon why. in audio there are many false hoods , cables to fix all.
usb can be great but it seems much cheaper dacs with spidif or hdmi sound better anyway
I’ve read some reclock spidif some don’t. but any ive used seem to benifit from a reclocking device be it usb conversion or card in server
 

AMR / iFi audio

Industry Expert
Aug 21, 2019
2,636
1,153
260
43
UK
ifi-audio.com
If bits were just bits and jitter stops at just isolation then why do many use re clockers anyway
ask empirical audio or ideon why. in audio there are many false hoods , cables to fix all.
usb can be great but it seems much cheaper dacs with spidif or hdmi sound better anyway
I’ve read some reclock spidif some don’t. but any ive used seem to benifit from a reclocking device be it usb conversion or card in server
Isolation stops noise, but not jitter. It actually adds a tiny bit of jitter to the signal. Reclocking usually uses PLL circuits to work, which are like a filter for jitter. They don't cut perfectly everything out. Of course, it can be better or worse in quality. It depends on the application and new clock quality. We use a GMT system with a precision integrated circuit and a clock with jitter in femtoseconds.
 
  • Like
Reactions: Alrainbow

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
A question do you have any tests showing jitter before and after ?
i thought the term jitter is not a real term in that there is no specific meaning
what I thought and I’m not a pro is jitter is many types of noise be it voltage or timing issues. Sone makers have tried to fox the issue with a load the track memory then this is sent to the player
a few do this but it never sounds right to me
many re clockers do as you said fix in part and effect things in others
I own from the past a few of your products
they do work always
but now I’ve kind of given up on usb to network or usb conversion to spidif
I also try I2S over Netwrok or hdmi
to me it sounds the same but others feel it’s worse
I mentioned empirical for a few reasons
one being every product he makes that I use works always a plus
expense dacs do have the edge on usb for sure
lampizator is one that has made usb a go to input
but cheaper dacs to me need spidif and some kind of re clocker
im old on this
had Paul pang pci usb boards bet this rings a bell lol
most people of your products too. All good stuff and very non expensive products that do improve things very Much
I really appreciate your input on these forums kind of you
 
  • Like
Reactions: AMR / iFi audio

microstrip

VIP/Donor
May 30, 2010
20,806
4,698
2,790
Portugal
If bits were just bits and jitter stops at just isolation then why do many use re clockers anyway
ask empirical audio or ideon why. in audio there are many false hoods , cables to fix all.
usb can be great but it seems much cheaper dacs with spidif or hdmi sound better anyway
I’ve read some reclock spidif some don’t. but any ive used seem to benifit from a reclocking device be it usb conversion or card in server

I think we should separate USB from SPDIF or AES/EBU when addressing jitter. USB is asynchronous and SPDIF or AES/EBU carry clock information.

The effects of jitter and digital noise must be listened in our particular system configuration. Some great sounding physical CD transports of the past - Mark Levinson, Forsell, Metronome, CEC - could sound great because they had a particular jitter signature. Forcing their digital signals through anti-jitter systems made them sound average at best ... :eek:
 
  • Like
Reactions: Alrainbow

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
what you posted is a fact
many new dacs reclock the spidif where usb no
some dacs spidif ext is a but others a miss
as for transports again I agree hit and miss
 

AMR / iFi audio

Industry Expert
Aug 21, 2019
2,636
1,153
260
43
UK
ifi-audio.com
A question do you have any tests showing jitter before and after ?
i thought the term jitter is not a real term in that there is no specific meaning
what I thought and I’m not a pro is jitter is many types of noise be it voltage or timing issues. Sone makers have tried to fox the issue with a load the track memory then this is sent to the player
a few do this but it never sounds right to me
many re clockers do as you said fix in part and effect things in others
I own from the past a few of your products
they do work always
but now I’ve kind of given up on usb to network or usb conversion to spidif
I also try I2S over Netwrok or hdmi
to me it sounds the same but others feel it’s worse
I mentioned empirical for a few reasons
one being every product he makes that I use works always a plus
expense dacs do have the edge on usb for sure
lampizator is one that has made usb a go to input
but cheaper dacs to me need spidif and some kind of re clocker
im old on this
had Paul pang pci usb boards bet this rings a bell lol
most people of your products too. All good stuff and very non expensive products that do improve things very Much
I really appreciate your input on these forums kind of you
I'll ask the RND if they have anything that I can show publicly. Some of this stuff is proprietary.

Following Wiki:
In electronics and telecommunications, jitter is the deviation from true periodicity of a presumably periodic signal, often in relation to a reference clock signal. In clock recovery applications it is called timing jitter.[1] Jitter is a significant, and usually undesired, factor in the design of almost all communications links.

And translating that into English: jitter is a deviation of timing in digital signals. In other words: the signal is not arriving in the receiver when it should. It can be too early, or too late. It includes every signal we use, not only the clocks.

Your experiences seem to match the fact that resampling is not a cure-for-all. It helps, but it won't fix a poor signal source.

All of our current products reclock both USB and other inputs.
 
  • Like
Reactions: Alrainbow

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
Jitter refers to timing variations, noise refers to amplitude variations. Jitter is divided at the top into random, which is not related to the signal or the clock (or theoretically anything else), and deterministic, which is somehow correlated (related to) some other signal that could be the signal, clock, power supply noise, some other interference source, etc. IIRC one the specs we use at work for high-speed (multi-Gb/s) serial links used in computers for connecting to drives and such defines 27 different specific types of jitter we characterize, but they all fall into the two broad random or deterministic categories.

Random jitter is discussed in a couple of the articles on WBF linked in my sig.

Virtually every DAC these days has very good clock recovery and isolates the output from incoming jitter very well.
 
  • Like
Reactions: AMR / iFi audio

AMR / iFi audio

Industry Expert
Aug 21, 2019
2,636
1,153
260
43
UK
ifi-audio.com
Jitter refers to timing variations, noise refers to amplitude variations. Jitter is divided at the top into random, which is not related to the signal or the clock (or theoretically anything else), and deterministic, which is somehow correlated (related to) some other signal that could be the signal, clock, power supply noise, some other interference source, etc. IIRC one the specs we use at work for high-speed (multi-Gb/s) serial links used in computers for connecting to drives and such defines 27 different specific types of jitter we characterize, but they all fall into the two broad random or deterministic categories.

Random jitter is discussed in a couple of the articles on WBF linked in my sig.

Virtually every DAC these days has very good clock recovery and isolates the output from incoming jitter very well.
We are so happy that in audio we don't have to deal with multi-Gb/s signals... Lenght matching traces is a chore...
 
  • Like
Reactions: DonH50

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing