Jitter Audibility

[tbewick]

Seeing as this was a subject I was having trouble with, here are some pages from The Digital Interface Handbook, 2nd edition, by Francis Rumsey and John Watkinson, Focal Press 1995. Chris also posted some data on this before, which he kindly gave me a link to in an earlier thread:

http://forums.audioholics.com/forums/showthread.php?t=4547

The attachments are of poor image quality, but they are readable.

 

[mtrycrafts]

Seeing as this was a subject I was having trouble with, here are some pages from The Digital Interface Handbook, 2nd edition, by Francis Rumsey and John Watkinson, Focal Press 1995. Chris also posted some data on this before, which he kindly gave me a link to in an earlier thread:
The attachments are of poor image quality, but they are readable.

Then there is this:

Benjamin, Eric and Gannon, Benjamin ' Theoretical and Audible Effects of Jitter on Digital Audio Quality,' 105th AES Convention, 1998, Print 4826.


Actual listening experiments

I think his conclusions may be a bit different, not sure though. They indicate
that it is not that easy and a lot is needed that is well above what has been actual measurements of players.

 

[tbewick]

Then there is this:

Benjamin, Eric and Gannon, Benjamin ' Theoretical and Audible Effects of Jitter on Digital Audio Quality,' 105th AES Convention, 1998, Print 4826.


Actual listening experiments

I think his conclusions may be a bit different, not sure though. They indicate
that it is not that easy and a lot is needed that is well above what has been actual measurements of players.

I'll try and get hold of that. Unfortunately I only have easy access to The Journal of Audiology.

I haven't had a chance to digest all the information fully, but I was surprised at the reduction in the higher frequency signal-to-noise ratios caused by jitter (fig. 6.4). These measurements were made at high SPL's, so hopefully jitter modulation noise will be inaudible at normal listening levels.

Thanks for the SACD/DVD Audio paper link you gave earlier (http://www.hfm-detmold.de/eti/projek...paper_6086.pdf). I haven't read through it fully yet but it should be interesting.

 

[westcott]

It seems that the first thing that goes in a player is the ability to maintain a constant rpm causing audible jitter.

 

[tbewick]

It seems that the first thing that goes in a player is the ability to maintain a constant rpm causing audible jitter.

I'm sorry, I don't understand what you mean. The data is read off the compact disc and undergoes very effective error correction. This should prevent any issues with the CD player r.p.m. performance.

'CRCC checking is very effective in detection of transmission error. If the number of CRCC bits is n, detection probability is 1 - (2 {raised to the power -n}). If, say, n is 16, as in the case of the Sony PCM-1600, detection probability is:

1 - (2 {raised to the power -16}) = 0.999 985 or 99.998 5 %.

This means that the CRCC features almost perfect error detection capability.'

- Digital Audio Technology, 4th edition, J. Maes & M. Vercammen, Focal Press 2001. Page 96.

 

[westcott]

I'm sorry, I don't understand what you mean. The data is read off the compact disc and undergoes very effective error correction. This should prevent any issues with the CD player r.p.m. performance.

'CRCC checking is very effective in detection of transmission error. If the number of CRCC bits is n, detection probability is 1 - (2 {raised to the power -n}). If, say, n is 16, as in the case of the Sony PCM-1600, detection probability is:

1 - (2 {raised to the power -16}) = 0.999 985 or 99.998 5 %.

This means that the CRCC features almost perfect error detection capability.'

- Digital Audio Technology, 4th edition, J. Maes & M. Vercammen, Focal Press 2001. Page 96.

I have always assumed it was an rpm error. So what does cause jitter to occur so prevalently with older\failing cd players? It must be mechanical because it is inconsistent and is wear related.

 

[tbewick]

I have always assumed it was an rpm error. So what does cause jitter to occur so prevalently with older\failing cd players? It must be mechanical because it is inconsistent and is wear related.

I think you are confusing jitter with read error. Read error with older CD players is most likely the result of the lens and focus mechanisms deteriorating with time.

 

[westcott]

If jitter is audible, then it really does not matter how it is produced.

In fact, if you want to use the term read error, it seems far more likely that this is the cause of most audible jitter than any sampling rate or error correction method.

 

[WmAx]

I think you are confusing jitter with read error. Read error with older CD players is most likely the result of the lens and focus mechanisms deteriorating with time.

The paper above claims some rather high jitter amounts. These levels are not approached in properly operating modern equipment of good design.

-Chris

 

[tbewick]

If jitter is audible, then it really does not matter how it is produced.

In fact, if you want to use the term read error, it seems far more likely that this is the cause of most audible jitter than any sampling rate or error correction method.

Jitter is a precisely defined term in electronics. The jitter you are describing uses the word in the looser, general sense. As I said earlier, the data read from the CD undergoes very effective error correction. Only a CD player on its last legs would give rise to large amounts of read errors, and even then, this is totally unrelated to jitter audibility.

 

[tbewick]

The paper above claims some rather high jitter amounts. These levels are not approached in properly operating modern equipment of good design.

-Chris

I can only go on my own equipment's specs. My DVD player has a jitter of around 0.2 ns. Assuming the interface transition jitter is 'well below 20 ns' (p 157 of the book I referenced, which is image 2), and that my receiver has as good a jitter performance as my DVD player, then figure 6.4 (p 155, image 1) should cover all the bases with regard to signal-to-noise performance.

It's a shame that receiver specifications don't list jitter or full-bandwidth signal-to-noise data on digital inputs/outputs. The specs given aren't really that useful. In general, I don't really know how well modern equipment performs. The full-bandwidth signal-to-noise ratio tests that some reviewers do should, I presume, include the modulation noise caused by jitter.

 

[WmAx]

I can only go on my own equipment's specs. My DVD player has a jitter of around 0.2 ns. Assuming the interface transition jitter is 'well below 20 ns' (p 157 of the book I referenced, which is image 2), and that my receiver has as good a jitter performance as my DVD player, then figure 6.4 (p 155, image 1) should cover all the bases with regard to signal-to-noise performance.

Also note that the reference is over a decade old. Since the paper is based on other studies/references, the actual equipment it refers to is probably on the order of up to 15 years old. In cases where it discussed any sort of 'typical' equipmet jitter amounts, it should be noted that the paper can not refer to modern equipment. ADC/DAC systems have improved considerably, so far as measured performance, over the last 15 years.

The paper also refers to 'worst case' audio signal; which is BTW, an isolated sine wave. The thresholds as defined in the Dolby paper which I referred, that you linked to earlier, required over 2 ns for an isolated sine wave for the most sensitive trained test subject. For the most sensitive music signals, threshold raised into the tens of ns, for the most sensitive trained test subject. The typical critical listener would not begin to be able to detect the jitter until it raised to even higher levels. The tests were also done on closed studio monitor headphones to increase the SNR; so on speakers in a normal environment(not a special acoustically isolated room), it would be even less detectable.

-Chris

 

[mtrycrafts]

I have always assumed it was an rpm error. So what does cause jitter to occur so prevalently with older\failing cd players? It must be mechanical because it is inconsistent and is wear related.

No, not RPM. CD rpm is not constant at all as it needs a relatively constant data downloading to the buffer to keep it stocked, clocked and processed.
In the inner circle it spins fast and slows to the outside. The data goes to a buffer, so speed is not an issue, withing limits as if the buffer runs out, no data and silence

Jitter is timing errors in the clock and perhaps elsewhere. There is no timing info on the CD.

 

[mtrycrafts]

I'll try and get hold of that. Unfortunately I only have easy access to The Journal of Audiology.

I haven't had a chance to digest all the information fully, but I was surprised at the reduction in the higher frequency signal-to-noise ratios caused by jitter (fig. 6.4). These measurements were made at high SPL's, so hopefully jitter modulation noise will be inaudible at normal listening levels.

Thanks for the SACD/DVD Audio paper link you gave earlier (http://www.hfm-detmold.de/eti/projek...paper_6086.pdf). I haven't read through it fully yet but it should be interesting.

I wonder how they accounted for speaker distortion at such high levels? Who listened at 120dB spl? Threshold of pain, or just all calculated?

 

[westcott]

No, not RPM. CD rpm is not constant at all as it needs a relatively constant data downloading to the buffer to keep it stocked, clocked and processed.
In the inner circle it spins fast and slows to the outside. The data goes to a buffer, so speed is not an issue, withing limits as if the buffer runs out, no data and silence

Jitter is timing errors in the clock and perhaps elsewhere. There is no timing info on the CD.

Here is a link to an article I ran across today by accident.

Basically they say to save your money on an expensive CD transport and just slave the transport to a DAC in Digital Master Mode.

http://www.enjoythemusic.com/magazine/manufacture/1104/

Just thought I would share. I am very sensitive to jitter but always assumed it was because the transports were less than ideal.

 

[Rock&Roll Ninja]

And thats why my next system's will do-away with the pre-amp alltogether. I'll run CD transport -> DAC (with level adjust like Benchmark) -> Power Amps. or CD->Benchmark DAC1 in the case of my headphone system.

Now I just need a very large sum of disposable income.

 

[WmAx]

And thats why my next system's will do-away with the pre-amp alltogether. I'll run CD transport -> DAC (with level adjust like Benchmark) -> Power Amps. or CD->Benchmark DAC1 in the case of my headphone system.

Now I just need a very large sum of disposable income.

If you are buying that stuff for decorative or pride of ownership purposes, cool. If you are buying it for some supposed performance advantage; it's nothing more than a big jar of snake oil.

-Chris

 

[mtrycrafts]

Also note that the reference is over a decade old. Since the paper is based on other studies/references, the actual equipment it refers to is probably on the order of up to 15 years old. In cases where it discussed any sort of 'typical' equipmet jitter amounts, it should be noted that the paper can not refer to modern equipment. ADC/DAC systems have improved considerably, so far as measured performance, over the last 15 years.

The paper also refers to 'worst case' audio signal; which is BTW, an isolated sine wave. The thresholds as defined in the Dolby paper which I referred, that you linked to earlier, required over 2 ns for an isolated sine wave for the most sensitive trained test subject. For the most sensitive music signals, threshold raised into the tens of ns, for the most sensitive trained test subject. The typical critical listener would not begin to be able to detect the jitter until it raised to even higher levels. The tests were also done on closed studio monitor headphones to increase the SNR; so on speakers in a normal environment(not a special acoustically isolated room), it would be even less detectable.

-Chris

I believe that paper is a calculation based arrival what might be but no listening tests conducted like the Benjamin paper.

 

[mtrycrafts]

And thats why my next system's will do-away with the pre-amp alltogether. I'll run CD transport -> DAC (with level adjust like Benchmark) -> Power Amps. or CD->Benchmark DAC1 in the case of my headphone system.

Now I just need a very large sum of disposable income.

Before you buy, you should read one DBT on the Benchmark DAC1:

http://www.matrixhifi.com/contenedor_dac1.htm

While it is in Spanish, you may be able to translate it by right clicking on page and look for that 'translate to English' The google toolbar will do that

As wmax said, you'd be wasting good $$$. Spend it on your room acoustics or speakers.

 

[westcott]

Before you buy, you should read one DBT on the Benchmark DAC1:

http://www.matrixhifi.com/contenedor_dac1.htm

While it is in Spanish, you may be able to translate it by right clicking on page and look for that 'translate to English' The google toolbar will do that

As wmax said, you'd be wasting good $$$. Spend it on your room acoustics or speakers.

This one is in English!

http://www.stereophile.com/digitalsourcereviews/886/