jacobo2u
jacobo2u's picture
Offline
Last seen: 9 years 6 months ago
Joined: Feb 22 2013 - 7:22pm
Sound Cards
struts
struts's picture
Offline
Last seen: 2 years 10 months ago
Joined: Feb 1 2007 - 12:02pm

...the answer is no.  They may all send the same bits but they have different amounts and types of jitter (timing error) which can significantly influence the sound quality at recording and playback time (although not in between, when the bits are just "sitting there", doing nothing).  I am not saying a more expensive digital source will always sound better than a cheaper one, but one with less jitter will, all else being equal, always sound better than one with more.

Apologies if that wasn't your question.  If not, please could you clarify?

jazzfan
jazzfan's picture
Offline
Last seen: 5 years 2 weeks ago
Joined: Sep 8 2005 - 8:55am
struts wrote:

...the answer is no.  They may all send the same bits but they have different amounts and types of jitter (timing error) which can significantly influence the sound quality at recording and playback time (although not in between, when the bits are just "sitting there", doing nothing).  I am not saying a more expensive digital source will always sound better than a cheaper one, but one with less jitter will, all else being equal, always sound better than one with more.

Apologies if that wasn't your question.  If not, please could you clarify?

@struts:

Please leave all the nonsense about the big bad jitter monster to the hacks who write for the audio magazines. In almost all present day digital audio playback devices, in other words devices manufactured within the past two to three years, ALL jitter is well  below the audible range. The jitter monster is presently being used to scare unsuspecting consumers into spending their money trying to solve a problem that doesn't exist.

Now there above statement does not mean that all digital sources sound equally good just that jitter IS NOT a factor regardless of what is published in the audio magazines.

struts
struts's picture
Offline
Last seen: 2 years 10 months ago
Joined: Feb 1 2007 - 12:02pm

Hi jazzfan!

I think you prove my point with an apparent contradiction:

"ALL jitter is well below the audible range"

" Now there above statement does not mean that all digital sources sound equally good.."

A digital signal only has two variables, bit value and time.  If the sound is different yet the bits are the same then I would assert that there can only be one explanation*.

Or, turn it around.  If they don't sound equally good but jitter isn't the explanation what is?  

* of course others would assert that the explanation is that I am imagining things and that of course they sound the same because the bits are the same.  

One of my electronics professors had a great axiom to the effect that "if it sounds different but measures the same then either you're measuring the wrong thing or your measuring equipment isn't resolving enough".  I wrote it down on a scrap of paper that I've saved all these years.  If only I could find it! 

struts
struts's picture
Offline
Last seen: 2 years 10 months ago
Joined: Feb 1 2007 - 12:02pm

"If there is an observable difference yet the measurements are the same, then either the instrumentation is insufficiently sensitive to an important parameter to detect the difference, or there is an important parameter that you either know nothing about, or have no instrument to measure."

jazzfan
jazzfan's picture
Offline
Last seen: 5 years 2 weeks ago
Joined: Sep 8 2005 - 8:55am
struts wrote:

Hi jazzfan!

I think you prove my point with an apparent contradiction:

"ALL jitter is well below the audible range"

" Now there above statement does not mean that all digital sources sound equally good.."

A digital signal only has two variables, bit value and time.  If the sound is different yet the bits are the same then I would assert that there can only be one explanation*.

Or, turn it around.  If they don't sound equally good but jitter isn't the explanation what is?  

* of course others would assert that the explanation is that I am imagining things and that of course they sound the same because the bits are the same.  

One of my electronics professors had a great axiom to the effect that "if it sounds different but measures the same then either you're measuring the wrong thing or your measuring equipment isn't resolving enough".  I wrote it down on a scrap of paper that I've saved all these years.  If only I could find it! 

struts wrote:

"If there is an observable difference yet the measurements are the same, then either the instrumentation is insufficiently sensitive to an important parameter to detect the difference, or there is an important parameter that you either know nothing about, or have no instrument to measure."

Taking your professor's quote I would say that in the case of jitter the instrumentation is more than capable of measuring jitter down to levels humans can not hear so therefore it must be that in the case of digital sources "there is an important parameter that you either know nothing about, or have no instrument to measure."

But whatever it is IT AIN'T JITTER! Unless of course you are a cable manufacturer selling $3,000 USB cables or a writer for a magazine in which that cable manufacturer advertises.

jacobo2u
jacobo2u's picture
Offline
Last seen: 9 years 6 months ago
Joined: Feb 22 2013 - 7:22pm

sounds like jitter, that is wrong timing, is the major thing i need to worry about.  Optical out is not as reliable, correct?  Also, how do card manufactures measure their jitter?  Is there any specs that I need to find, or is it kind of "by the ear"?

struts
struts's picture
Offline
Last seen: 2 years 10 months ago
Joined: Feb 1 2007 - 12:02pm
jazzfan wrote:

Taking your professor's quote I would say that in the case of jitter the instrumentation is more than capable of measuring jitter down to levels humans can not hear so therefore it must be that in the case of digital sources "there is an important parameter that you either know nothing about, or have no instrument to measure."

.. but that's fine, neither of us can prove our hypotheses so we'll just have to wait for science to catch up.

I do agree that jitter provides a ready foil for unscrupulous marketeers to latch onto, and I also agree that its effects can easily be blown out of proportion.  So while I do, in point of fact, believe that jitter is the root of most digital evil, I also believe that in most circumstances its audible effects will be so minor that they will be easily masked by other system deficiencies.  

However for highly resolving systems in properly treated rooms I believe the effects can easily be heard - even at levels well below the resolution of the J-test suite (which I think even JA would admit is pretty much obsolete as far as measuring SoA digital equipment is concerned, unfortunately there is nothing better available at anywhere near a reasonable cost).

struts
struts's picture
Offline
Last seen: 2 years 10 months ago
Joined: Feb 1 2007 - 12:02pm
jacobo2u wrote:

sounds like jitter, that is wrong timing, is the major thing i need to worry about.  Optical out is not as reliable, correct?  Also, how do card manufactures measure their jitter?  Is there any specs that I need to find, or is it kind of "by the ear"?

Well, I propose that it's jitter, jazzfan believes that it's "quiff" (a name one of my university professors gave to the parameter you don't know about but which causes the difference).

Regardless of which one of us is right it is nevertheless a fact that jitter is quite difficult to measure.  There is one quite widely-used tool, the J-Test, which is specifically designed to test for jitter in interfaces that embed the clock into the datastream (e.g. AES3 and S/PDIF - not asynch USB).  Unfortunately this tool has a number of limitations and a resolution (approx 120 ps) which have rendered it all but obsolete for measuring today's digital products.  JA now uses a Audio Precision SYS2722 measurement system.  This offers far higher resolution (it has 24-bit A/D converters whereas the NI cards used by the J-Test tool are only 16-bit).  Unfortunately however the Audio Precision does not have the functionality to automatically search the FFT plot for sidebands, so while it presents a lot of high resolution data it doesn't offer much help in analyzing it.

I think the best bet is to proceed by ear.  Jitter specs can be misleading because of weaknesses in the test tools.

Oh, and optical out is every bit as reliable, but in many products seems to sound inferior to the wired connection.  Again, I believe the reason for this is ultimately related to jitter, but jazzfan would probably also assert that it was down to "quiff".

jazzfan
jazzfan's picture
Offline
Last seen: 5 years 2 weeks ago
Joined: Sep 8 2005 - 8:55am
struts wrote:
jacobo2u wrote:

sounds like jitter, that is wrong timing, is the major thing i need to worry about.  Optical out is not as reliable, correct?  Also, how do card manufactures measure their jitter?  Is there any specs that I need to find, or is it kind of "by the ear"?

Well, I propose that it's jitter, jazzfan believes that it's "quiff" (a name one of my university professors gave to the parameter you don't know about but which causes the difference).

Regardless of which one of us is right it is nevertheless a fact that jitter is quite difficult to measure.  There is one quite widely-used tool, the J-Test, which is specifically designed to test for jitter in interfaces that embed the clock into the datastream (e.g. AES3 and S/PDIF - not asynch USB).  Unfortunately this tool has a number of limitations and a resolution (approx 120 ps) which have rendered it all but obsolete for measuring today's digital products.  JA now uses a Audio Precision SYS2722 measurement system.  This offers far higher resolution (it has 24-bit A/D converters whereas the NI cards used by the J-Test tool are only 16-bit).  Unfortunately however the Audio Precision does not have the functionality to automatically search the FFT plot for sidebands, so while it presents a lot of high resolution data it doesn't offer much help in analyzing it.

I think the best bet is to proceed by ear.  Jitter specs can be misleading because of weaknesses in the test tools.

Oh, and optical out is every bit as reliable, but in many products seems to sound inferior to the wired connection.  Again, I believe the reason for this is ultimately related to jitter, but jazzfan would probably also assert that it was down to "quiff".

My understanding of the optical digital audio interface is that the poor performance of many inexpensive devices, such as computer sound card and motherboards, is mainly caused by the use of really cheap transmitters (the little device that the cable plugs into) when a slightly more expensive transmitter would result in much improved performance. Hence the optical interface has gotten a bad reputation.

Log in or register to post comments
-->
  • X