Jitter article

steveinaz

Interesting:
http://www.audiocraftersguild.com/AandE/cd-jitter.pdf#search='jitter on cd players'

it's in *.pdf format, cool read.

heiney9

Another reason outboard DAC's sound so much better. Somewhere in my favorites I have another link to an excellent jitter article. If I find it I'll post it. For those not familiar jitter has nothing to do with a shaky transport....many associate the two.

FWIW

H9

Tour2ma

http://www.altmann.haan.de/jitter/english/engc_navfr.html

TheReaper

I just don't understand the audio world's problem with jitter. In the computer world data has been serially transmitted jitter free for years using independant transmit and receive clocks (16x receive to eliminate jitter). For audio, the only timing that is really needed off the input signal, is the average bit rate for drift adjustment.

What am I missing?

heiney9

TheReaper wrote:

I just don't understand the audio world's problem with jitter. In the computer world data has been serially transmitted jitter free for years using independant transmit and receive clocks (16x receive to eliminate jitter). For audio, the only timing that is really needed off the input signal, is the average bit rate for drift adjustment.

What am I missing?

Anyone who is seriously into audio knows a bit about jitter. This area of digital playback can be/is just as contested as the cable/interconnect debate. I understand the princples of jitter, but am not sure if I can detect it or if my efforts to reduce it have been successful.

Here's the article I was speaking of in my earlier post. It's an excellent in-depth view of jitter.

http://www.jitter.de/english/engc_navfr.html

H9

EDIT: Sorry, just realized my link is the same as Tour2ma.

mldennison

TheReaper wrote:

I just don't understand the audio world's problem with jitter. In the computer world data has been serially transmitted jitter free for years using independant transmit and receive clocks (16x receive to eliminate jitter). For audio, the only timing that is really needed off the input signal, is the average bit rate for drift adjustment.

What am I missing?

well the problem i see with this statement is that i dont see where in the computer world you are going digital to analog. if you are just doing a purely digital serial transfer then using the oversampling clock on both sides allows you to get rid of any jitter problems. when you are going d to a though, you have alot less room for error and hence the problems with jitter.

TheReaper

TheReaper wrote:

For audio, the only timing that is really needed off the input signal, is the average bit rate for drift adjustment.

In those articles, what they are claiming is causing the jitter. Is the lock-stepping of the output clocking for the dac, to the input bits.

When there could be a separate clock for the ouput dac, regulated by the average input bit rate, eliminating the input bit jitter. As long as the average output meets the average input, the stream flows.

Sami

mldennison wrote:

well the problem i see with this statement is that i dont see where in the computer world you are going digital to analog.

If that's the only problem then I don't see the point for expensive CDP's connected digitally. A digital output to a DAC is a digital output to a DAC, bit for bit. My $17 computer soundcard with optical out should be identical in sound when playing ape-files from HD.

mldennison

sami, i agree with you 100% i really dont see the need for an expensive transport. if you are using an outboard dac then the only thing that i can see that a nicer transport is doing is reducing the bit error rate which i cant see being a very big deal unless you are dealing with a really scratched disc.

of course i have not used an external dac so i really cant comment from experience but that is the way i see it. here is a thread of someone else who seemed to come to the same conclusion though:
http://www.polkaudio.com/forums/showthread.php?t=35100

steveinaz

Sami wrote:

If that's the only problem then I don't see the point for expensive CDP's connected digitally. A digital output to a DAC is a digital output to a DAC, bit for bit. My $17 computer soundcard with optical out should be identical in sound when playing ape-files from HD.

While true that the output is bit-for-bit, even from a $19 CD walkman (if it had a digital output); where jitter comes in is from imperfect clock timing, not bit errors. There is no such thing as a 100% accurate crystal clock, therefore we have jitter. However, the better the transport design and power supply section, the more jitter is reduced.

The claims as to why different transports can sound different is based on this premise. As always, you get what you pay for; more accurate clock circuit=more expensive; beefed-up, clean power supply=more expensive.

This is how I understand it from the reading.

Sami

steveinaz wrote:

While true that the output is bit-for-bit, even from a $19 CD walkman (if it had a digital output); where jitter comes in is from imperfect clock timing, not bit errors.

The question is in the design then, is the information transferred to the DAC as it is in the disc? If it is then there is no difference at all. If the transfer isn't identical then it's a different story. If jitter is a major problem to hi-fi crowd then I'm sure someone would have already developed equipment that would get rid of it, at least until it reaches DAC.

steveinaz

There are expensive add-on products out there to reduce jitter, but jitter will always be there (well, at least until the perfect crystal is made). They way I understand it, jitter is what causes alot of the complaints with digital reproduction, flat sound stage, glare, harshness, etc.

Sami

steveinaz wrote:

There are expensive add-on products out there to reduce jitter, but jitter will always be there (well, at least until the perfect crystal is made). They way I understand it, jitter is what causes alot of the complaints with digital reproduction, flat sound stage, glare, harshness, etc.

But how does it affect the transport stage from the disc to the DAC? To make that exactly the same should not be a problem (but are there any equipment that do it is the real question).

heiney9

If you read the link from my earlier post. You'll realize that jitter can be introduced during the recording process if this is the case it can never be corrected. Jitter is a real quantifiable issue. It can be reduced but never eliminated. Jitter has to do with the sending and receiving clocks, during conversion, not syncing up correctly. There are many other area of design in an audio playback system that need attention before jitter suppression becomes an issue. The overall design and quality of the unit are much more important factors than just managing jitter al by itself.

Simply...a cheap poorly designed unit will not benefit from reduced jitter. The rest of the pieces of the puzzle have to be in place to benefit.

H9

Sami

heiney9 wrote:

If you read the link from my earlier post. You'll realize that jitter can be introduced during the recording process if this is the case it can never be corrected. Jitter is a real quantifiable issue. It can be reduced but never eliminated.

For what I read in that article is that while a digital transport could be identical, in todays designs it isn't. That is why I always posted my remarks with conditions as I suspected this was the case.

heiney9

Sami wrote:

For what I read in that article is that while a digital transport could be identical, in todays designs it isn't. That is why I always posted my remarks with conditions as I suspected this was the case.

Why do you keep referencing transport? Jitter is incorrect syncing between the sending and receiving clocks during conversion. The transport really has very little effect on jitter. Pleas explain what you mean. (perhaps reference parts of the article)

H9

Sami

heiney9 wrote:

Why do you keep referencing transport? Jitter is incorrect syncing between the sending and receiving clocks during conversion. The transport really has very little effect on jitter. Pleas explain what you mean. (perhaps reference parts of the article)

H9

Because some people are saying that the transport matters from the disc to the DAC. It shouldn't unless the hardware specs are designed in a way that introduces jitter. You should be able to design hardware where it isn't an issue.

http://www.jitter.de/english/engc_navfr.html
"Some digital transmission formats like S/PDIF(by Tomi Engdahl) or AES/EBU(crystal an22.pdf) carry clock and data in one signal."

No jitter there I assume so transporting the signal through digital connection should have no effect on the sound. But it doesn't explain why there is this mentioned as it shouldn't matter at all if it truely was identical transport:

http://www.jitter.de/english/engc_navfr.html

"Use the best digital interconnects:

Use SToptical instead of Toslink.
Use AES3 interconnects (XLR) instead of S/PDIF (RCA). "

heiney9

Ok I see where the confussion lies. I think of transport as the part of the unit where you put the disc to be read. You use the word transport as a verb as in transporting info from one place to another.

Your assumption is incorrect. Just by using the "best" digital interconnects or using the S/PDIF or AES/EBU connection will not supress jitter. Again jitter is a byproduct of incorrect syncing between the sending and receiving clock during the conversion from digital to analog or visa/versa. You could have the "best" interconnects made and if the clocks are poor and out of sync then it still will suffer from jitter.

I'm going to try this analogy, it might make sense or completely miss the point.

A photograph is a representation of an image the same way a digital "bit" is a representation of part of an analog waveform (in audio). When you take a picture the quality (sharpness, color accuracy etc.) is dependant on the quality of the camera. All cameras do the same thing some just do it better (more correctly)The same goes for components used in the audio conversion process. The better the clock, dac's, power supply, input stage, output stage the better the analog waveform we hear. All these components help reduce jitter, but the main factor are the sending and receiving clocks.

Reduction of jitter eliminates glare, hardness and fatigue.

Comparing a computer which always operates in the digital domain is unfair because it never has to switch/convert to analog.

Hope this helps...because beyond this simple explanation it all gets very complicated very quickly. I'm not prepared to take the time to go much beyond this. However, there are 1000's of articles out there if you do a Google search.

H9

Sami

heiney9 wrote:

Your assumption is incorrect. Just by using the "best" digital interconnects or using the S/PDIF or AES/EBU connection will not supress jitter.

It was not my assumption, it was in the website you posted. I was wondering why it was there as I don't think it should matter.

heiney9 wrote:

Comparing a computer which always operates in the digital domain is unfair because it never has to switch/convert to analog.

I was just wondering the digital part of the equation, whether it introduces jitter or not. This because I have read multiple comments how a $10 computer sound card isn't the same as a $1000 CDP when both are connected to a DAC (external or in receiver/prepro). There is no DA conversion so both should produce identical sound (read my original post).

heiney9

Sami wrote:

I was just wondering the digital part of the equation, whether it introduces jitter or not. This because I have read multiple comments how a $10 computer sound card isn't the same as a $1000 CDP when both are connected to a DAC (external or in receiver/prepro). There is no DA conversion so both should produce identical sound (read my original post).

I'm confused . You say there is no DA conversion. In order to hear audio it has to be an analog waveform. So there HAS to be DA conversion. A $10 computer card will NEVER sound as good as $1000 CDP. On that little soundcard all the processes have to take place just the same as in an expensive CDP or DAC. Poor quality components are used and a computer is a very hostile environment for an audio signal. Poor power supply regulation, lots of EMI interference, ribbon cable connections are not friendly to audio signals. Noisy fans can dirty up the signal because of poor isolation, etc.

Yes, the digital stream from a sound card into to an external DAC will sound much better than the stand alone soundcard. With my 2nd system which is hooked to my computer I can switch between my external DAC and the sound card and the difference is VERY apparent.

Again....Jitter is a byproduct of the conversion process from Digital to Analog or visa/versa and it's created by poor syncing of the sending and receiving clocks. If there is no conversion then there is no associated jitter.

Sami

heiney9 wrote:

I'm confused . You say there is no DA conversion. In order to hear audio it has to be an analog waveform.

No DA conversion before it reaches the DAC (logical), in reference to a cheap soundcard and expensive CDP both connected via digital connection sounding the same (connected to the same device/DAC).

heiney9

Sami wrote:

No DA conversion before it reaches the DAC (logical), in reference to a cheap soundcard and expensive CDP both connected via digital connection sounding the same (connected to the same device/DAC).

Problem is all things aren't equal. The computer is a hostile environment. So I'd have to say NO they won't sound the same. But that's a subjective question as well as a subjective answer.

I will say my computer system sounds much better than I expected using relatively inexpensive components. Comparing the internal soundcard to the outboard DAC is a night and day difference. I hope this helps answer your question.

H9

Sami

heiney9 wrote:

Problem is all things aren't equal. The computer is a hostile environment. So I'd have to say NO they won't sound the same. But that's a subjective question as well as a subjective answer.

And the question really is why not? Is it by design of the CD format? Is it by design of the connection and data package? If there is no DA conversion happening then without a specific reason they should sound the same. What introduces the difference between the two in digital format?

heiney9

Sami wrote:

And the question really is why not? Is it by design of the CD format? Is it by design of the connection and data package? If there is no DA conversion happening then without a specific reason they should sound the same. What introduces the difference between the two in digital format?

Sorry Sami, I've tried to stay with you....but at this point, and really several points before, I'm not sure where you are coming from or trying to go. Things are not equal because of the quality of parts and the intended design to get to the end result (which in this case is analog sound). It seems you want to discuss just the digital side of the equation, but want an answer for both sides (digital and analog). For audio purposes if there is no conversion to analog then it's a moot point. The analog side is primarily where the problems occur (audibly) anyways. Beyond that I have no idea what you want to know. 1 bit of info is 1 bit of info....but 1 bit on its own is nothing. In audio several bits have to be converted back to analog and depending on the quality of the components and design it can make or break the sound (output).

So strictly bits are bits, but it's the process and the environment that determine accuracy, quality, etc.

H9

heiney9

Sami wrote:

If there is no DA conversion happening then without a specific reason they should sound the same. What introduces the difference between the two in digital format?

If there is no DA conversion happening then there is no sound. What does a bunch of 1's and 0's sound like? There is no difference in the digital format. But we don't listen to digits. The digits are a representation of part of the original analog waveform that have to be converted back to analog.

H9

Sami

heiney9 wrote:

Sorry Sami, I've tried to stay with you....but at this point, and really several points before, I'm not sure where you are coming from or trying to go.

You have already stated that the problem of jitter occurs in the conversion to analog, but you also say that you don't think a cheap soundcard sounds the same as an expensive CDP when both are connected digitally to external DAC. Why do you feel they won't sound the same if bits are bits and there is no analog conversion (between the two devices and external DAC)?

heiney9 wrote:

If there is no DA conversion happening then there is no sound.

Of course there is DA conversion before the sound comes out but only after both sources have delivered the digital information to the DAC. If the information is the same and the DAC does the conversion, why would it sound different?

I am only interested in the path to the DAC, not the analog conversion itself. Does the information change somehow in digital domain before reaching DAC depending on the device delivering it?

heiney9

Do you know what a DAC does? Digital to Analog converter. By definition it take the digital info via a cable and converts the digital bits to analog which is how our ears hear sound. You can't stay 100% in the digital domain and still hear sound. A computer sound card can 1 of 2 things. It can convert the digital bits to analog and you hear them thru your speakers or it can be set up to output a digital stream which cna them be hooked to an external DAC to be converted to analog and heard thru speakers hooked ot an amplifier. So in BOTH instances conversion takes place. It HAS to or there is no sound (audio).

heiney9

Sami wrote:

I am only interested in the path to the DAC, not the analog conversion itself. Does the information change somehow in digital domain before reaching DAC depending on the device delivering it?

Theoretically no. But again there are people who feel Coaxial cables are better than Optical cables. Again if you are using a cheap cable other interference can occur which has nothing to do with the digits but is still there. Same in the computer environment, it noisy and unstable and full of EMI/RF noise as well as fan noise and poor power supply isolation which could interfer with the signal, even though the digits are all the same.

H9

Sami

heiney9 wrote:

Do you know what a DAC does? Digital to Analog converter.

That's very obvious, and that's not the issue here. What you haven't explained to me is why would two devices sound different if they are both connected to the same DAC, delivering the same information? You said this earlier:

heiney9 wrote:

So I'd have to say NO they won't sound the same.

heiney9

Sami wrote:

That's very obvious, and that's not the issue here. What you haven't explained to me is why would two devices sound different if they are both connected to the same DAC, delivering the same information? You said this earlier:

Sorry I wasn't trying to sound flippant with that question. Two players hooked to the same DAC via the same cable should sound the same. Let's take the computer out of the equation for reasons I explained before. I have a very cheap DVD player and a relatively expensive CDP hooked to the DAC in my main system and I can't tell the difference other than the scraping sound the DVD player's transport makes. I think I've finally answered the right question. We've both been a bit confused on each other's posts.

H9

Sami

heiney9 wrote:

I think I've finally answered the right question.

Yes you have.

Your post before about the signal interference, while possible, could be avoided by error correction. Whether the design of the transfer protocol takes care of it or not is something I was wondering. If we can transfer bit-perfect data over the air with error correction it certainly would be possible with the computer via a cable. But that all depends on the specifications on how the data is transmitted to the DAC.