[Snaques]

This is kinda neewbish question, but it still bothers me

Why should the source have any practical effect on the SQ when it's connected to a DAC with a digital IC?

I mean, optical discs, like CD's, introduce errors once in something like 100 000 bits. Add some error correction and you're somewhere in the region of 1 in 10 000 000 (my approx.). If someone claims to hear this one bit error he must be really something

So why should I buy a good quality CDP if I'd be using a DAC? Usability issues aside.

 

[Nerull]

It's really not the bit errors that are the problem, but rather the termination of the digital out. The jitter induced by some things, like optical, also brings about more problems than bit errors. Proper termination of SPDIF will give better results than that you can get from RCA coaxial or optical.

Proper termination would be BNC (75 ohm) with appropriate transmission / receiver circuits on each side. This link should give you a pretty good idea. http://www.epanorama.net/documents/audio/spdif.html

These things usually require modding though, which may or may not be to your liking.
So what you'd be looking for is a BNC out cd player and BNC in DAC with properly made input / output circuitry.

The jitter present on the master clock in the cd player also gets coupled into the digital out, so replacing that with a low jitter spec clock will also give you some benefits. Perhaps a master / slave set up?

These would be the main reasons I went DIY, because I have quite a bit of control then without having to work with what "they" give you.

~Tom

 

[hciman77]

Quote:

Originally Posted by Snaques /img/forum/go_quote.gif This is kinda neewbish question, but it still bothers me Why should the source have any practical effect on the SQ when it's connected to a DAC with a digital IC? I mean, optical discs, like CD's, introduce errors once in something like 100 000 bits. Add some error correction and you're somewhere in the region of 1 in 10 000 000 (my approx.). If someone claims to hear this one bit error he must be really something So why should I buy a good quality CDP if I'd be using a DAC? Usability issues aside.

This debate has been done to death here. If you do a search and look for terms like jitter you will see a healthy debate. In general few people argue about errors in bits arriving or not as these are rarer than the pope's children unless your source is rubbish.

The big debate is about things such as timing errors (jitter) and the evidence for the effect is widely debated with some sources citing jitter errors of 50 ps (10 to the minus 12) as unacceptable and others saying that jitter up to 30 ns (10 to the minus 9) is undetectable in listening tests when embedded in music. You pays your money and you takes your choice.

I have a low end DVD player connected to an external DAC via optical (supposedly the worst possible method) and it sounds great and identical (to me) to a CD player connected via Coax, of course it also sounds great to me via analog so maybe that does not help too much, sorry.

 

[ezkcdude]

Quote:

Originally Posted by hciman77 /img/forum/go_quote.gif The big debate is about things such as timing errors (jitter)

\begin{rant}

I think using words like "error" inflames the debate. Jitter is not caused by timing errors per se. Let's just call it "timing uncertainty" or "timing precision". People like to think of digital as simply 1's and 0's, but those logical units are generated by analog components, which are subject to the laws of physics. Mathematicians and programmers can expect perfection (in most cases); engineers make a living from the lack of perfection inherent in the real world.

\end{rant}

 

[Snaques]

Thanks a bunch guys.

I did search the forum, but obviously I was looking for the wrong thing. Now it makes much more sense. Not that I'm certain that it makes as huge difference as people claim, but at least they have some reasoning behind their thoughts.

 

[Davesrose]

Seems like it has to do with the DAC too.....there's a thread on here about just how clean DACs like the DAC1 are. Normally, a DAC will also get info on what timing they're supposed to be getting the digital signal. So if there's a mismatch in timing, artifacts can occur. Some DACs like the DAC1 will actually ignore that information (or at least it's not as much of an issue). That's what I gather though.....since so many oposing opinions on this site leaves one confused

 

[audioengr]

Quote:

Originally Posted by Davesrose /img/forum/go_quote.gif Seems like it has to do with the DAC too.....there's a thread on here about just how clean DACs like the DAC1 are. Normally, a DAC will also get info on what timing they're supposed to be getting the digital signal. So if there's a mismatch in timing, artifacts can occur. Some DACs like the DAC1 will actually ignore that information (or at least it's not as much of an issue). That's what I gather though.....since so many oposing opinions on this site leaves one confused

It's true that the DAC-1 and the Lavry are both semi-immune to jitter on the inputs compared to the typical DAC, but nothing I have tested or has been tested by my customers is proving to be totally jitter-resistant. Jitter still seems to have an effect on the output quality. I am hoping that my new Pace Car device is the first to be totally jitter immune, assuming that it works.

Steve N.

 

[Garbz]

Jitter is almost impossible to eliminate completely, and worst of all 20ps of so can cause audible and measurable problems on the output. Searching this forum for "jitter" will show many heated discussions on the topic.

 

[Mr_Sukebe]

People can argue all day long about whether dedicated transports sound better or not.
Personally, I'm a believer and think it does make a difference. Frankly, I don't care why there's a difference, all that matters is that it sounds better.

BTW, my DAC is meant to also be "semi immune" to jitter.

 

[Garbz]

Yes I'm sure all DACs claim to be semi-immune to jitter. It's called marketing. Search for threads on the Larvy DA-10 and Benchmark DAC1 which are probably the best at jitter rejection, and there'll be evidence that even this isn't perfect.

 

[MiChael.]

Have there actually been any ABX tests of jitter ?

Or anything more scientific than just "I feel like this" or "Cause it's measurable, there must be audible difference".

Anyway it's nice there is always jitter to blame of bad listening experience.

Why not syncronating the clocks of recording studio to the source's clocks.

 

[puiah11]

Quote:

Originally Posted by Snaques /img/forum/go_quote.gif This is kinda neewbish question, but it still bothers me Why should the source have any practical effect on the SQ when it's connected to a DAC with a digital IC? I mean, optical discs, like CD's, introduce errors once in something like 100 000 bits. Add some error correction and you're somewhere in the region of 1 in 10 000 000 (my approx.). If someone claims to hear this one bit error he must be really something So why should I buy a good quality CDP if I'd be using a DAC? Usability issues aside.

I think most CD transports work similarly. It's the power supply and other internal components that vary between CDPs.

Sterophile has a good article about how CDP guts affect digital jitter and such.

I'm sure you can find that article with the google.

 

[Garbz]

Lol jitter causes minimal changes, but bad ones all the same. It definitely is not the source of all bad problems, sometimes poor design of the analog stage is the culprit

. I got a far greater quality boost changing my I/V stage than any digital changes to my DAC, and I have done a LOT of them. As far as ABXing no, but so far every case including my own of changes where jitter has improved (re-terminating the digital link in my DAC is one example) has had positive results, even from people who had no idea what I changed and only heard the system before and after.

It is far more measurable than audible, but the fact is a study by ... Julian Dunn I think showed a lovely graph of how much jitter at each frequency is audible compared to measurable. It is far worse in the treble, and shows while it is a small problem it's a problem all the same.

Why not syncronating the clocks of recording studio to the source's clocks? Because in theory it would be better to sync to the DAC's clock

Infact studios often use an external dedicated clock generator.

 

[hciman77]

Quote:

Originally Posted by MiChael. /img/forum/go_quote.gif Have there actually been any ABX tests of jitter ? Or anything more scientific than just "I feel like this" or "Cause it's measurable, there must be audible difference".

There have been (as far as I know properly controlled) only a few listening tests with variable amounts of jitter. Of the actual tests (not mathematical models or anecdotal results) I found three - one (A Japanese research paper) determined that between 30 - 40 ns jitter introduced serious degradation and the other two cited 10ns as detectable in a high frequency sine wave and 20ns as detectable in music, it may have been the same study cited twice. I have not found any other controlled listening tests where the jitter level was controlled or measured.

I ask myself. If Jitter is really such a problem then both my low end sources $230 Marantz multichanger (Coax) and $50 Philips DVD player (optical) would sound ***** via an external DAC and yet both sound perfectly good and frankly identical through the same DAC a non jitter imune 203.2. Marantz players are renowned for their awful jitter figures - typically 500 - 600ps - this is 60 times the supposedly degrading 10ps figure often bandied about. If 10ps was problematical my players would be unlistenable. To my ears the insertion of the DAC actually makes zero difference except that the signal is louder. And goodness knows what the jitter figures for my first gen 14bit Marantz would have been.

I really would not worry about jitter.

For reference Julian Dunn of the AES did a few papers http://www.nanophon.com/audio/1394_sampling_jitter.pdf and http://www.audioprecision.com/bin/AN-5.pdf he came up with the 10ps figure (modelled) but in this paper also doubted whether it was valid in the light of listening tests that he cited.

Audibility considerations
It is one thing to be able to identify and measure sampling jitter. But how can we
tell if there is too much?
A recent paper by Eric Benjamin and Benjamin Gannon describes practical research
that found the lowest jitter level at which the jitter made a noticeable difference
was about 10 ns rms. This was with a high level test sine tone at 17 kHz. With music,
none of the subjects found jitter below 20 ns rms to be audible.7
This author has developed a model for jitter audibility based on worst case audio
single tone signals including the effects of masking.8 This concluded:
“Masking theory suggests that the maximum amount of jitter that will not produce
an audible effect is dependent on the jitter spectrum. At low frequencies this level is
greater than 100 ns, with a sharp cut-off above 100 Hz to a lower limit of approximately
1 ns (peak) at 500 Hz, falling above this frequency at 6 dB per octave to approximately
10 ps (peak) at 24 kHz, for systems where the audio signal is 120 dB above the threshold
of hearing.” In the view of the more recent research, this may be considered to be overcautious.
However, the consideration that sampling jitter below 100 Hz will probably be less audible
by a factor of more than 40 dB when compared with jitter above 500 Hz is useful
when determining the likely relative significance of low- and high-frequency sampling
jitter.
Now in this model 120db above the threshold of hearing at 24Khz is an extraordinary measure of audible, and this is the absolute worst case. Here we are talking about a level in excess of the threshold for pain. being conservative the threshold for hearing at 24kz is approx 80db so we are talking about a 200db signal.

From elsewhere

Once the fact is in sight that a digital signal is just a special kind of analogue signal, it becomes plausible that even small inaccuracies in this analogue signal could lead to inaccuracies in the sonic output of a digital-to-analogue device. A "digital signal," analogue though it is, is nonetheless a special kind of analogue signal, and the errors that are sonically relevant in a "digital signal" are of a special kind. In particular, surprisingly small errors in timing of the digital signal can produce significant, potentially audible alterations in the sound. How small is surprisingly small? Errors on the order of tens of nanoseconds are sufficient to cause audible changes under some conditions.

"Potentially" may well be the operative word in most cases. Certainly, jitter of sufficient amount can be audible. The Stereophile Test CD NO.2 provides a demonstration involving 11 kHz jittered by 10 nanoseconds at 4 kHz frequency. There is an audible roughness to the sound compared with an unjittered tone. But consider what is involved here: Fifty times the 200 picosecond threshold we just discussed; jitter of fixed frequency, which creates sidebands of definite frequencies, 7 kHz and 15 kHz, the 7 kHz one at a region of high hearing sensitivity; and a sustained pure tone at high level as a signal. No wonder the result is audible! The relevance of the jitter levels of contemporary high-quality CD players to audible performance is less obvious. The evidence adduced in the press tends to be of the anecdotal variety. And, since jitter is only one of a considerable number of potential electronic failings of CD players, it is not easy on an anecdotal basis to isolate the sonic effect of jitter alone.

This is an interesting pdf - http://akmedia.digidesign.com/suppor...tter_30957.pdf
the authors duscuss a new clocking protocol and its low jitter capability - typically 50ps. Then they discuss comparative listening tests with it and other (technically inferior) clocks and conclude.

It’s clear from these examples and from discussions with audio engineers from around the world that it is difficult if not impossible to come to a single conclusion when subjectively identifying the audible effects of jitter. While jitter is quantifi-able and relatively easy to measure, how it affects complex waveforms such as music is much harder to quantify. Many consider the 192 I/O to have the finest sounding clocking scheme while others prefer third party dedicated clocks and using the distributed technique in their systems—even though the effect of external clocks often produces higher amounts of jitter. On the bench, the 192 I/O compares favorably with all of its similarly priced competitors producing low jitter across the spectrum. In fact, it has far less jitter than many dedicated clocks and yet many prefer the sound of these clocks over the 192’s internal clock. Because of this apparent lack of consistency between theory and actual experience, more science and measurement is called for, particularly on the perceptual side. Controlled listening tests involving a cross section of participants using established standards need to be done in order to more fully understand the issue. Marketing materials from many manu-facturers use anecdotal evidence and testimony from highly regarded individuals to promote their products and while that is powerful product endorsement, it needs to be recognized as such. Subjective reactions and uncontrolled listening tests should not be used as a substitute for science and it’s clear that more study needs to be done to fully understand the audible effects of jitter on digital audio.

 

[Garbz]

Quote:

Originally Posted by hciman77 /img/forum/go_quote.gif There have been (as far as I know properly controlled) only a few listening tests with variable amounts of jitter. Of the actual tests (not mathematical models or anecdotal results) I found three - one (A Japanese research paper) determined that between 30 - 40 ns jitter introduced serious degradation and the other two cited 10ns as detectable in a high frequency sine wave and 20ns as detectable in music, it may have been the same study cited twice. I have not found any other controlled listening tests where the jitter level was controlled or measured.

The difference is small as I have said, and exhibits only in the treble. Often it can be the cause of slightly bright or sterile sound. Stereophile released a jitter test track which I posted here somewhere. It shows a pure sinewave, and a then again with exhadurated jitter. On a system upstairs from a cheap DVD player using my HD580s and balanced Dynalo I can't hear the difference, but in my main system it is pretty obvious.

Dunn's comments made sense to me, since jitter causes slewing problems it is more prenounced in signals that slew faster (treble). The graph from Dunn's AES paper showed the following jitter to be audible at a given frequency:

<200Hz : unmeasured
250Hz : 1uS
400Hz : 10ns
600Hz : 1ns
1khz : 300ps
4khz : 100ps
20khz : 10ps

Reference: Dunn, J.: “Jitter: Specification and Assessment in Digital Audio Equipement”, AES Convention Paper 3361, October 1992.

The most unfortunate thing about these is all are peer-reviewed, all are probably right, but none seem to indicate an exact number. The numbers seem to change with how they are measured. To be honest I did not care about the topic until someone pointed out to me what I was looking for, the difference was that small. I should go back and say thanks for causing me to spend extra money on something I was enjoying