Dolby Digital (AC3) - digital cable question(s)
John Muskara
greater than 1 meter. Any suggestions for sources?
What is the longest cable I can use without degradation
to signal itegrity?
Can I use a standard RCA shielded hookup cable or
RCA "video cable" vs a "digital" cable without degradation?
My receiver is a Denon 3600 and only has an RCA jack for
AC3 input. There is another RCA "coaxial" jack and optical jack for
non-AC3 digital (So I cannot use an optical cable for AC3).
John Muskara
Rick
George Mills
Don't spend to much because it will either work or it won't
and just about anything that conducts will work.
There isn't much room in between as far as performance goes on
AC-3.
John Muskara wrote:
>
> I am having a hard time finding a cable (specified for digital)
> greater than 1 meter. Any suggestions for sources?
> What is the longest cable I can use without degradation
> to signal itegrity?
>
> Can I use a standard RCA shielded hookup cable or
> RCA "video cable" vs a "digital" cable without degradation?
>
> My receiver is a Denon 3600 and only has an RCA jack for
> AC3 input. There is another RCA "coaxial" jack and optical jack for
> non-AC3 digital (So I cannot use an optical cable for AC3).
>
> John Muskara
--
===============================================================
George Mills
email: mi...@softronix.com
http://www.softronix.com
The www page contains some very powerful educational software.
Our single most important investment is our kids.
BowHunter
using RG-9 coax cable with F-connectors. I then went to radio shack
and got 2 RCA adapters ($4). After swap testing both cables with both
Dolby Digital and DTS DVDs for about 5 hours I returned the Monster
Cable. NO audible difference. BTW I know that Monster Cable is far
from state of the art, but it was the best that was available at the
time in my area.
Tom Caines
plugs) works just as good as Monster, costs less, and from what I have heard
is manufactured by Monster. Your interconnects and speaker cables are the
place where you can suffer some degradation if you go too cheap.
And even there stay within reason. Your cables shouldn't cost more than your
components.
~~~~~~~~~~~~~~~~~~~
Tom Caines
The Answer Is Always Yes
~~~~~~~~~~~~~~~~~~~
BowHunter wrote in message <37becfeb...@news.pipeline.com>...
:After buying a MONSTER CABLE digital coax cable ($60), I made one
:
Brad Marcus
connection between your DVD and receiver. This means that the cable
you need should be 75-ohm with 75-ohm connections. I often see people
erroneously post answers like "use any cable with RCA connectors and it
works fine." This is not true. The more you deviate from the required
spec, the more performance suffers. With a high quality digital
coaxial cable that is a true 75-ohm cable with true 75-ohm connectors,
you can have a cable that is quite long.
Brad
--
Better Cables LLC
http://www.bettercables.com
br...@bettercables.com
In article <37add39a....@news.gte.net>,
mus...@gte.net (John Muskara) wrote:
> I am having a hard time finding a cable (specified for digital)
> greater than 1 meter. Any suggestions for sources?
> What is the longest cable I can use without degradation
> to signal itegrity?
>
> Can I use a standard RCA shielded hookup cable or
> RCA "video cable" vs a "digital" cable without degradation?
>
> My receiver is a Denon 3600 and only has an RCA jack for
> AC3 input. There is another RCA "coaxial" jack and optical jack for
> non-AC3 digital (So I cannot use an optical cable for AC3).
>
> John Muskara
>
>
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
Joe Shovlin
Brad Marcus <br...@bettercables.com> wrote in message
news:7pp9s8$gnf$1...@nnrp1.deja.com...
> The spec for AC-3 (now known as Dolby Digital) calls for a 75-ohm
> connection between your DVD and receiver. This means that the cable
> you need should be 75-ohm with 75-ohm connections. I often see people
> erroneously post answers like "use any cable with RCA connectors and it
> works fine." This is not true. The more you deviate from the required
> spec, the more performance suffers. With a high quality digital
> coaxial cable that is a true 75-ohm cable with true 75-ohm connectors,
> you can have a cable that is quite long.
>
> Brad
People don't understand what DIGITAL means. That means that you are passing
an "on or off"
signal over the cable. The processor doesn't care if the signal is a little
high or a little low.
Given a 0-10 volt signal, anything below 5V will be interpreted as "0V" (off
bit) and anything above
5V will be intrepreted as "10V" (on bit). If your cable shifts the bit's
voltage by a small amount
it doesn't matter as far as the processor goes -- until the deviation
becomes big enough to change
bits from off to on (or vice versa). If "my bits", being carried over the $3
audio/video cable are
slightly shifted in voltage with respect to "your bits" carried over the $60
monster digital, the processor
cares not a whit. That's the idea of digital -- small deviations in signal
cause NO difference in
the end result.
Performance should be identical -- until the distortion in the signal
becomes big enough to matter.
You will know it right away when that happens.
Martin Raul Loin
wrote:
>The spec for AC-3 (now known as Dolby Digital) calls for a 75-ohm
>connection between your DVD and receiver. This means that the cable
>you need should be 75-ohm with 75-ohm connections. I often see people
>erroneously post answers like "use any cable with RCA connectors and it
>works fine." This is not true. The more you deviate from the required
>spec, the more performance suffers. With a high quality digital
>coaxial cable that is a true 75-ohm cable with true 75-ohm connectors,
>you can have a cable that is quite long.
>
>Brad
Specs are one thing, but digital cable is digital.
So it is either working or not. And when it is not working you hear
audible clicks and hisses or you don't hear anything at all.
Any other sound quality isssues are produced by the listeners
imagination. These include richer or fuller sound and other nonsense
like that.
- M
Gordon Fraser
Raul Loin) writes:
>
>Specs are one thing, but digital cable is digital.
>So it is either working or not. And when it is not working you hear
>audible clicks and hisses or you don't hear anything at all.
>
This is actually quite wrong.
Read articles on how cable design and termination can affect digital signal
transmission in TAG McLaren Audio's white papers on their cable design.
Gordon
StereoStereo: Intelligent Solutions for Intelligent Homes
260 Saint Vincent St
Glasgow 0141 248 4079
10-6.00pm Closed Tuesday and Sunday
KHanawalt
>imagination. These include richer or fuller sound and other nonsense
>like that.
Thank you, M.
Finally, someone is willing to call a spade a spade.
Ken
You can tell a lot from a man's toys & rides: DVD/Dolby Digital Home theater
'85 Z-28 305 TPI '85 Honda VF1000R '81GoldWing Interstate '89 Safari
(wife) '88 Grand Am (daughter) For off-roading: '74 Wagoneer & '84 Cherokee
KHanawalt
>transmission in TAG McLaren Audio's white papers on their cable design.
>
Say, TAG McLaren Audio wouldn't be trying to sell high quality (read expensive)
cables, would they?
Gordon Fraser
(KHanawalt) writes:
>
>Say, TAG McLaren Audio wouldn't be trying to sell high quality (read
>expensive)
>cables, would they?
>
Yes they do manufacture their own cables as well as their own high quality
hi-fi and home cinema equipment. I don't know whether you would call them
expensive though.
The guys that were part of the team designing these cables and components are
also the guys who design the telemetry systems used in West McLarens F1 racing
cars, as well as around 75% of the formula 1 grid. They send digital
information about the cars set up etc to the pits by rf while they are racing
round the track. I believe they call this a "hostile enviroment", although it
doesn't appear to be as hostile as this newsgroup : )
These guys know a bit about digital signal transmission! Whether you believe
them or want to put their equipment to the test is entirely up to you of
course.
It is extremely easy to show a difference in sound quality between two
interconnects carrying digital information. When used in a dvd/cd player to
transport situation.
Rather than criticise why don't you go and read the articles. They try to give
a scientific reason as to why these differences are heard.
Murray Peterson
wrote:
>...
>Rather than criticise why don't you go and read the articles. They try to give
>a scientific reason as to why these differences are heard.
>...
OK -- I went and read these guy's articles -- what a waste of time.
Yet another case of the "golden ears" hearing things that don't exist,
and then inventing pseudo-scientific reasons for these imaginary
differences.
However, I guess that people who believe this stuff deserve to be
parted from their money. Maybe I should start up a new business :-)
Murray Peterson
Murray Peterson
Homepage: http://members.home.net/murraypeterson
Gordon Fraser
>
>OK -- I went and read these guy's articles -- what a waste of time.
Well done for at least bothering to go and read it. Now you should try and hear
them and see if there is a difference.
Dave Beal
Gordon Fraser wrote:
> In article <19990822163404...@ng-fm1.aol.com>, khan...@aol.com
> (KHanawalt) writes:
>
> >
> >Say, TAG McLaren Audio wouldn't be trying to sell high quality (read
> >expensive)
> >cables, would they?
> >
>
> Yes they do manufacture their own cables as well as their own high quality
> hi-fi and home cinema equipment. I don't know whether you would call them
> expensive though.
>
> The guys that were part of the team designing these cables and components are
> also the guys who design the telemetry systems used in West McLarens F1 racing
> cars, as well as around 75% of the formula 1 grid. They send digital
> information about the cars set up etc to the pits by rf while they are racing
> round the track. I believe they call this a "hostile enviroment", although it
> doesn't appear to be as hostile as this newsgroup : )
>
> These guys know a bit about digital signal transmission! Whether you believe
> them or want to put their equipment to the test is entirely up to you of
> course.
>
> It is extremely easy to show a difference in sound quality between two
> interconnects carrying digital information. When used in a dvd/cd player to
> transport situation.
>
OK, please post a reference to a blind trial that demonstrates
an audible difference between digital interconnect cables.
ggh
>
> This is actually quite wrong.
>
> Read articles on how cable design and termination can affect digital signal
> transmission in TAG McLaren Audio's white papers on their cable design.
>
This is all bullSSSt. If the goddamn cable brings all its bits down
the path in the right order within the right voltage tolerances, there
is ABSOLUTELY no difference. This is digital! Hello? Anyone home?
Anyone selling anything should not be allowed to right a "whitepaper"
(read "advertisement"). They should be called "blackpapers", because
of the lies built into them.
Double blind testing is, and always will be the ONLY way to determine
differences. End of story.
I think a fool and his money are lucky enought to find each other in
the first place.
GG
Rick Nelson
here. Jitter is real, and it has real consequences.. ask any EE out there who
specializes in data communications (like me).
> If the goddamn cable brings all its bits down
> the path in the right order within the right voltage tolerances, there
> is ABSOLUTELY no difference. This is digital! Hello? Anyone home?
There are several requirements for digital transmissions, correct bit order and
voltage tolerances are certainly among the most important. However, for
applications like real-time audio, you need to make sure that the timing is
correct. Real-time transmissions need to be synchronized (e.g., via a
Phase-locked loop), because there is no major buffering or data storage. There
are other considerations to be be made if the cable is of low quality -- bits
could be lost, sampling rates could become irregular, inductance of the wire can
cause all sorts of neat effects...
>
> Double blind testing is, and always will be the ONLY way to determine
> differences. End of story.
Not sure about the ONLY way... but it's certainly the best way in this case..
Oh yeah... go here... http://cyclone.stanford.edu/~davros/jitter/jitter.html
Rick
--
Chesapeake Sciences Corp. <-- a real life engineering company!
1127B Benfield Blvd Millersville, MD 21108
Tel: (410) 923-1300 x3430 Fax: (410) 923-2669
Robert A. King
>ggh.n...@wwa.com (ggh) wrote:
>
>>This is digital! Hello? Anyone home?
>
>I must you the same questions.Helllooooo? Are we awake?
Yes, but did you show up for class?
>Not only does the cable need to convey the bits in the correct order and
>voltages, but it also cannot introduce "ringing" that could be sensed as
>additional bits or noise,
True, but do you know how to prevent this?
>but the rise and fall times of the signal must be
>retained. An overly reactive cable can destroy a digital signal. It can
>turn a square wave into a sine wave if it is really poor quality.
When's the last time you looked at a digital signal on an ocilloscope?
They ARE square waves. You make it sound like the digital ssignal is
a sine wave that *might* be dostorted into a square wave.
>You're a bit too cynical for my money. You tend to remind me of Einstein's
>famous quip, "Mediocre minds condemn that which is beyond their
>comprehension." Your opening paragraph proves your knowledge is, at best,
>mediocre.
Your's too, buddy. Perhaps you need a refresher course in digital
data communications...
-- Robert
bk...@my-deja.com
consensus (IIRC) was
for AC3/DTS soundtracks, where the information is compressed anyway,
and you would need to do some digital signal processing to get back the
PCM data, any ordinary 75ohm cable will do
for stereo (CD audio) reproduction however, little changes in timings
(jitter) of the PCM signal will have an impact on the final analog
signal
I would have to agree
in the past, i used to use good analog interconnects between DVD and
amp, but an elcheapo 75ohm video cable for AC3 until today.
I junked the analog interconnects and just compared the difference
between the the QED DS and the elcheapo 75ohm.
for DVDs, I couldn't tell much difference, honestly.. but with the CD
audio, bass came across tigher and more powerfully and the treble was
just incredible
System Tested
QED DS digital interconnect
normal 75ohm composite video cable
Sony STR-DB930 AC3/DTS amp
Sony DVP-725D DVD player
B&W 601 S2s (front and rear)
B&W CC6 S2 (centre)
B&W ASW1000
In article <37AF66B0...@softronix.com>,
George Mills <mi...@softronix.com> wrote:
> Use a 75ohm "Video" cable.
> Don't spend to much because it will either work or it won't
> and just about anything that conducts will work.
>
> There isn't much room in between as far as performance goes on
> AC-3.
>
> John Muskara wrote:
> >
> > I am having a hard time finding a cable (specified for digital)
> > greater than 1 meter. Any suggestions for sources?
> > What is the longest cable I can use without degradation
> > to signal itegrity?
> >
> > Can I use a standard RCA shielded hookup cable or
> > RCA "video cable" vs a "digital" cable without degradation?
> >
> > My receiver is a Denon 3600 and only has an RCA jack for
> > AC3 input. There is another RCA "coaxial" jack and optical jack
for
> > non-AC3 digital (So I cannot use an optical cable for AC3).
> >
> > John Muskara
>
> --
> ===============================================================
> George Mills
> email: mi...@softronix.com
> http://www.softronix.com
> The www page contains some very powerful educational software.
> Our single most important investment is our kids.
>
Robert A. King
wrote:
>There are several requirements for digital transmissions, correct bit order and
>voltage tolerances are certainly among the most important. However, for
>applications like real-time audio, you need to make sure that the timing is
>correct. Real-time transmissions need to be synchronized (e.g., via a
>Phase-locked loop), because there is no major buffering or data storage. There
>are other considerations to be be made if the cable is of low quality -- bits
>could be lost, sampling rates could become irregular, inductance of the wire can
>cause all sorts of neat effects...
But at what bandwidths do these effects become *significant*. I think
that's the key here. For the relatively narrow bandwidth applications
of PCM data, are these effects significant?
>Oh yeah... go here... http://cyclone.stanford.edu/~davros/jitter/jitter.html
Just curious, how do 100mhz T1 cables deal with jitter?
-- Robert
Rick Nelson
> But at what bandwidths do these effects become *significant*. I think
> that's the key here. For the relatively narrow bandwidth applications
> of PCM data, are these effects significant?
...and that's the other side of the coin... I can't really tell you when these
effects become signifigant, and it could very well be that they are not. I was just
informing everyone that jitter exists. Whether or not jitter causes cheaper cables
to not make your DVD's sound good is another question.
> Just curious, how do 100mhz T1 cables deal with jitter?
I don't know that one... if you are really curious, ask the folks at www.ieee.org
I know jitter can be reduced by using high-bandwith cables (fiber optic being the
obvious choice), and by using better synchronization methods.
Rick
--
Chesapeake Sciences Corp.
Kerry Shetline
> In article <37c08fb7.18667465@news>, m...@home.com (Murray Peterson) writes:
> >OK -- I went and read these guy's articles -- what a waste of time.
>
> Well done for at least bothering to go and read it. Now you should try and hear
> them and see if there is a difference.
You're letting yourself off easy here. You told everyone to go read this
white paper, as if it was so impressive. It's not. It's marketing hype
with pretty pictures, big talk laced with technical language, making
lots of claims but offering no proof. Other than a single footnote
mentioning blind testing (not double-blind) pertaining to only one of
their many claims, they never said anything about the methodology of the
listening tests they supposedly make.
If this "white paper" is so impressive, why don't they pass it on to a
peer-reviewed scientific journal?
Their point that experimental physics strikes a good balance between
ivory tower theorizing and crude empiricism is indeed a good point. Too
bad it has so little to do with what they seem to be doing. Too bad they
so blithely bypass the issue of the quality of experimental observations
and the importance of experimental controls.
If all you're going to end up telling someone is "Now you should try and
hear them [the cables in question] and see if there is a difference",
why bother trying to back your opinions with a bogus paper in the first
place?
-Kerry
Kerry Shetline
> Not only does the cable need to convey the bits in the correct order and
> voltages, but it also cannot introduce "ringing" that could be sensed as
> retained. An overly reactive cable can destroy a digital signal. It can
> turn a square wave into a sine wave if it is really poor quality.
I think the whole point of this discussion is not whether a digital
connection can go bad or not -- certainly they can, be they optical or
coaxial. The point is whether more subtle differences, that end up
translating into real audible effects, can exist between cables that
aren't failing in an obviously way with clicks, pops, chirps, or dead
silence.
Within in a fairly wide range of tolerance for cable connections, cable
types, impedance matching, etc., completely error-free transmission of
bits happens routinely with ordinary consumer gear and cables. Nothing
fancy is required. Whatever noise and distorted wave shapes and ringing
you're worrying about, it's rare that those effects are of great enough
magnitude to cause failure. When the effects are that large, failure
tends to be obvious.
When talking about home theater, and therefore AC-3 or DTS audio signals
most of the time, even the dread audiophile issue of jitter should have
very little to do with the cables used, given the buffering,
calculating, and re-clocking needed to process these compressed audio
bit streams.
I'm very skeptical about claims of "tighter bass", "greater air",
"improved imaging", and so forth, made about digital cables.
-Kerry
Brian L. McCarty
----------
In article <19990822172436...@ngol01.aol.com>,
gdfr...@aol.comRemoveMe (Gordon Fraser) wrote:
> It is extremely easy to show a difference in sound quality between two
> interconnects carrying digital information. When used in a dvd/cd player to
> transport situation.
Name ONE blind study that showed this!
You can't, because all the published blind studies I have seen have NEVER
shown any difference, as long as neither of the cables is DEFECTIVE.
---
Steve Zipser (Sunshine Stereo) is a proven:
Zipser is a liar http://dejanews.com/=dnc/getdoc.xp?AN=369217967
Zipser is a scammer http://dejanews.com/=dnc/getdoc.xp?AN=368363274
Zipser is a cheater http://dejanews.com/=dnc/getdoc.xp?AN=374900703
Zipser is a THIEF http://dejanews.com/=dnc/getdoc.xp?AN=509980240
Murray Peterson
Murray Peterson
On Mon, 23 Aug 1999 15:33:24 -0400, Rick Nelson <ri...@csciences.com>
wrote:
>I'm the last person to start any sort of war on NG's, but I really must inteject
>here. Jitter is real, and it has real consequences.. ask any EE out there who
>specializes in data communications (like me).
I also specialize in digital communications (low bandwidth, < 1MHz)
>There are several requirements for digital transmissions, correct bit order and
>voltage tolerances are certainly among the most important. However, for
>applications like real-time audio, you need to make sure that the timing is
>correct. Real-time transmissions need to be synchronized (e.g., via a
>Phase-locked loop), because there is no major buffering or data storage. There
>are other considerations to be be made if the cable is of low quality -- bits
>could be lost, sampling rates could become irregular, inductance of the wire can
>cause all sorts of neat effects...
>
Ah, but the nice thing about digital audio is that we only need to
re-create the digital bit stream, not an analog bit stream. Which
means that your link below has no meaning in this application
(re-creating an analog square wave). All we need to do is get the bit
value correct, and we don't care if it has been degraded to a sine
wave (or any other wave form). The harmonics are of no importance
whatsoever; in fact, a sine wave actually causes less problems than a
square wave. When sampling a digital bit stream, the <only>
requirement for perfect recreation of it is that we get each bit
correct (and error-correction codes even relax that requirement).
Over a 2 meter length, and at the bandwidth in question, I would be
quite happy with the proverbial coat hanger (I would like it
shielded).
>> Double blind testing is, and always will be the ONLY way to determine
>> differences. End of story.
>
>Not sure about the ONLY way... but it's certainly the best way in this case..
>
For arguments about subjective "sound", I don't know of any other way.
If I put a scope on the incoming signal and prove that the bit stream
has been perfectly re-created, would that convince anyone of the
validity of my argument (that cables make no difference)? I will
probably just get someone asserting that my measurements must be
meaningless because <they> can hear a difference. I think the only
real proof is a double blind test; that way the person can be proven
incorrect in a conclusive manner. Of course, they then start whining
about double blind tests being faulty, at which point we might as well
all go home -- superstition cannot be reasoned with...
>
>Oh yeah... go here... http://cyclone.stanford.edu/~davros/jitter/jitter.html
Went there, see above.
>
>Rick
Gordon Fraser
I admit I cannot post any references to double blind tests of interconnect
cables and frankly I don't care.
I trust my own judgement and believe that I have a longer memory span than a
Goldfish, thus enabling me to make comparisons myself. All I suggest is that
there are audible differences and that the digital engineers employed in a high
tech electronics company have tried to suggest reasons as to why these
differences exist.
To dismiss a phenomenon because it is beyond your understanding is your
perogative. I have an open mind about these things and this has allowed me to
enjoy my films and music more, without having to worry about whether I am
actually hearing soundtracks more clearly or whether I am just hearing those
voices inside my head again…..
It costs nothing to listen!
Gordon Fraser
<shet...@cyberzone.net> writes:
>
>You're letting yourself off easy here. You told everyone to go read this
>white paper, as if it was so impressive.
I never said anything about it being impressive. I claim I can percieve a
difference between cables carrying digital information. I claim that many of
our customers can do so as well.
I am not a digital engineer and therefore cannot explain why this may be. All I
suggest is that here is a company that specialises in digital signal transfer
technology and their engineers have come up with possible reasons for this
phenomenon. The fact that you see it as marketing hype is fine. I dont care!
I do not need to have a scientific paper presented to me to explain why I can
hear these differences. Even if you could prove that it was impossible for
there to be a diference I wouldn't care.
Effective cabling need not be expensive. I would respectfully suggest that it
costs nothing to ask to hear the difference between two cables carrying digital
information, in your local specialist retailer. If you hear no difference,
great! If you hear a difference, great! At least you will have experienced the
demonstration for yourself.
If you do not trust your own judgement, well that is another matter.....
Robert A. King
>no_...@no.where.com (Robert A. King) wrote:
>
>> ed...@aes-intl.com (Ed Jay) wrote:
>>
>>>ggh.n...@wwa.com (ggh) wrote:
>>>
>>>>This is digital! Hello? Anyone home?
>>>
>>>I must you the same questions.Helllooooo? Are we awake?
>>
>>Yes, but did you show up for class?
>>
>I did, but obviously you weren't there.
Really? What leads you to that conclusion?
>There are several techniques employed to reduce jitter and ringing. For
>someone who pretends to know it all enough to criticize others, you sure do
>ask a lot of questions.
I'm merely asking you to give us some of your obvious knowledge. Will
you?
>>Your's too, buddy. Perhaps you need a refresher course in digital
>>data communications...
>>
>I'll take one as soon as you have taken a course in reading.
I'll ask you again, how does one prevent ringing and reflections in a
digital cable? (I'll give you a hint -- it doesn't have anything to
do with the construction of the cable.)
I'll bet I don't get an answer...
-- RK
Robert A. King
<shet...@cyberzone.net> wrote:
>Ed Jay wrote:
>> Not only does the cable need to convey the bits in the correct order and
>> voltages, but it also cannot introduce "ringing" that could be sensed as
>> additional bits or noise, but the rise and fall times of the signal must be
>> retained. An overly reactive cable can destroy a digital signal. It can
>> turn a square wave into a sine wave if it is really poor quality.
>
>I think the whole point of this discussion is not whether a digital
>connection can go bad or not -- certainly they can, be they optical or
>coaxial. The point is whether more subtle differences, that end up
>translating into real audible effects, can exist between cables that
>aren't failing in an obviously way with clicks, pops, chirps, or dead
>silence.
But in a data stream, there cannot be any "changes" smaller than a bit
inversion. If a bit gets inverted (or a series of bits,) then the
transmission error is detected through various means. these
transmission errors aren't going to simply slip through and "color"
the sound. If they DO get through, the result is quite audible as a
glitch in the sound. They do NOT result in things such as
"coloration" or loss of "spectral cohesion" or "imaging" as some
golden-ear audiophiles would have you believe. Digital systems just
don't work that way.
>Within in a fairly wide range of tolerance for cable connections, cable
>types, impedance matching, etc., completely error-free transmission of
>bits happens routinely with ordinary consumer gear and cables. Nothing
>fancy is required. Whatever noise and distorted wave shapes and ringing
>you're worrying about, it's rare that those effects are of great enough
>magnitude to cause failure.
Exactly. the effects are there byt they're so minute, the waveform is
still detected and decoded without error. That's the main objective
behind digital transmission -- so that distorions and noise that
affect an analog signal will not affect a digital signal.
> When the effects are that large, failure
>tends to be obvious.
And effects that are that large are almost certainly due to
manufacturing or assembly defects.
>When talking about home theater, and therefore AC-3 or DTS audio signals
>most of the time, even the dread audiophile issue of jitter should have
>very little to do with the cables used, given the buffering,
>calculating, and re-clocking needed to process these compressed audio
>bit streams.
Preceisely.
>I'm very skeptical about claims of "tighter bass", "greater air",
>"improved imaging", and so forth, made about digital cables.
It's more "golden-ear" audiophile stuff. And if mere mortals can't
hear the differnces, it's their own fault, and are lesser men for it.
-- Robert
Robert A. King
wrote:
>Guys,
>
>I admit I cannot post any references to double blind tests of interconnect
>cables and frankly I don't care.
Then what good are you?
>I trust my own judgement and believe that I have a longer memory span than a
>Goldfish, thus enabling me to make comparisons myself. All I suggest is that
>there are audible differences
In what? Different digital cables?
>and that the digital engineers employed in a high
>tech electronics company have tried to suggest reasons as to why these
>differences exist.
And I believe, based on over 12 years experience in the computer and
digital communications industry, that at the bit level (which is what
we're talking about, after all,) there are essentially NO differences
between transmission techniques used in Home Theater. Ditto for the
quality of cables.
In short -- Unless there is a manufacturing defect, the data stream at
the output end of any digital cable, be it a $400 cable or a $4 cable,
Toslink, coaxial, or twisted-pair, will be identical and
undistinguishable from the data stream from any other cable.
>To dismiss a phenomenon because it is beyond your understanding is your
>perogative.
I dismiss the phenomenon because it runs counter to 12 years of
experience and common sense.
>It costs nothing to listen!
And only a little bit more to think. It's OK to listen and observe,
but one must also take into account the organ doing the listening and
observing will often delude itself. One MUST take that into accound,
especially when dealing with somthing as tenuous as music.
>StereoStereo: Intelligent Solutions for Intelligent Homes
I find this slogan very ironic...
-- RK
Martin Raul Loin
>
>
>Gordon Fraser wrote:
>
>> In article <19990822163404...@ng-fm1.aol.com>, khan...@aol.com
>> (KHanawalt) writes:
...
>> It is extremely easy to show a difference in sound quality between two
>> interconnects carrying digital information. When used in a dvd/cd player to
>> transport situation.
>>
>
>OK, please post a reference to a blind trial that demonstrates
>an audible difference between digital interconnect cables.
Yes. I would very much like to see one also.
And if that is not available then please post a reference to a study
where digital information is transmitted via an "cheap" and "hifi"
digital cable and the signal on the other end is compared to the
original signal.
I bet both cables transmit data as well and the differences are left
to the listeners imagination.
One only wonders why no such comparison is available in the press or
at any cable manufacturers site. After all digital data is extremely
easy to compare (to other or the original).
This whole thing is as ridiculous as comparing two different hard
discs and saying that programs and files stored on the other produce
richer and fuller data.
- M
Murray Peterson
wrote:
>Guys,
>
>I admit I cannot post any references to double blind tests of interconnect
>cables and frankly I don't care.
>
Sigh (see my previous post). I should probably quit saying anything
right now, since this thread can only go downhill from here. When one
side insists on strict, repeatable testing regimes, and the other side
doesn't care, then we are into the realm of blind faith and/or
superstition. Discussions across this divide quickly become
pointless.
>I trust my own judgement and believe that I have a longer memory span than a
>Goldfish, thus enabling me to make comparisons myself. All I suggest is that
>there are audible differences and that the digital engineers employed in a high
>tech electronics company have tried to suggest reasons as to why these
>differences exist.
>
Actually, reputable high tech companies first do strict, repeatable
tests to verify if these differences actually exist. Double blind
testing has never shown any of these perceived differences to be real.
The <only> way the "golden ears" can hear a difference is if they know
which cable they are listening to. In other words, I dispute your use
of the word "reputable" with respect to these companies.
>To dismiss a phenomenon because it is beyond your understanding is your
>perogative. I have an open mind about these things and this has allowed me to
>enjoy my films and music more, without having to worry about whether I am
>actually hearing soundtracks more clearly or whether I am just hearing those
>voices inside my head againв─╕..
>
If anyone is dismissing anything, you are (see your first sentence).
And asserting that something is beyond our understanding appears to be
more perjorative than anything the remainder of us have said. Are you
asserting that the rest of us are not "open minded"?
What is beyond our understanding? That double blind tests have shown
that these differences do not exist?
To believe in something with absolutely no hard evidence (except your
personal belief) is your perogative.
>It costs nothing to listen!
It costs a great deal of time and effort to perform a well controlled
listening test where the "placebo effect" can be ruled out. If I plug
in a new cable to my system and listen to it, it may sound better (or
worse) to me, but this is not a controlled test. Because of that, I
have no right to assert my belief to the rest of the world as being
the "right" one.
>
>Gordon
>
>StereoStereo: Intelligent Solutions for Intelligent Homes
>260 Saint Vincent St
>Glasgow 0141 248 4079
>10-6.00pm Closed Tuesday and Sunday
>
>http://members.aol.com/stere0/index.html
Murray Peterson
wrote:
>In article <37C20D0B...@cyberzone.net>, Kerry Shetline
><shet...@cyberzone.net> writes:
>
>>
>>You're letting yourself off easy here. You told everyone to go read this
>>white paper, as if it was so impressive.
>
>I never said anything about it being impressive. I claim I can percieve a
>difference between cables carrying digital information. I claim that many of
>our customers can do so as well.
>
But can they do so in a double blind situation (neither they, nor you,
know which cable is under test)?
>I am not a digital engineer and therefore cannot explain why this may be. All I
>suggest is that here is a company that specialises in digital signal transfer
>technology and their engineers have come up with possible reasons for this
>phenomenon. The fact that you see it as marketing hype is fine. I dont care!
>
But they haven't proven that the difference exists, so it is hype by
definition. Whether you care or not is unimportant.
>I do not need to have a scientific paper presented to me to explain why I can
>hear these differences. Even if you could prove that it was impossible for
>there to be a diference I wouldn't care.
>
Again, can you hear these differences under controlled testing
conditions (double blind)? I assert that you cannot.
>Effective cabling need not be expensive. I would respectfully suggest that it
>costs nothing to ask to hear the difference between two cables carrying digital
>information, in your local specialist retailer. If you hear no difference,
>great! If you hear a difference, great! At least you will have experienced the
>demonstration for yourself.
>
Not exactly a scientific test...
>If you do not trust your own judgement, well that is another matter.....
>
Nobody can trust their own judgement when it comes to "placebo
effect". The medical research community learned that lesson years
ago.
I am not arguing about what cables anyone should buy, I am arguing
about real vs perceived differences in sound. A perceived difference
may lead you to buy a different cable, and I cannot argue with that
assertion (or purchase). However, a perceived difference does not
give one the license to assert to the world that the difference is
real. Proof is now required, and assertions of "I can hear the
difference" are not proof.
Robert A. King
wrote:
>> Just curious, how do 100mhz T1 cables deal with jitter?
>
>I don't know that one... if you are really curious, ask the folks at www.ieee.org
From what I've seen, I've *never* heard jitter mentioned in reference
to any digital sitnals I've worked with (802.3, RS-232, DS0, DS1, TTL,
CMOS, etc...)
>I know jitter can be reduced by using high-bandwith cables (fiber optic being the
>obvious choice), and by using better synchronization methods.
Generally the choice of fiber is made for bandwidth/capacity reasons,
not for improved signal quality.
-- RK
Scott Randolph
> to any digital sitnals I've worked with (802.3, RS-232, DS0, DS1, TTL,
> CMOS, etc...)
I'm only posting this link because I thought paragraph #10 was pretty funny
(and applies to this thread)
http://www.galstar.com/~ntracy/acg/AandE/npt.on.jitter2.htm
Scott
Gordon Fraser
no_...@no.where.com (Robert A. King) writes:
>
>>StereoStereo: Intelligent Solutions for Intelligent Homes
>
>I find this slogan very ironic...
>
I do not.
Gordon Fraser
>If I plug
>in a new cable to my system and listen to it, it may sound better (or
>worse) to me, but this is not a controlled test. Because of that, I
>have no right to assert my belief to the rest of the world as being
>the "right" one.
ok ok . I agree it is not a controlled test. I have only ever taken part in one
double blind test. If someone wants to arrange it I will happily take part in a
double blind test of cable carrying digital data between cd and dac. I am sure
I could percieve the differences.
You are all right. Because I can hear a difference and believe it is nothing to
do with a placebo effect, this does not constitute it as fact rather it is my
opinion. Because you believe that there can be no difference because it is
digital data and all data can be measured as being transmitted and recieved
does not mean that I am wrong in my assertion that I can hear a difference does
it?
If it does then I am happy I am not scientist and do not have to validate my
own experiences to prove they actually happen!
Robert A. King
wrote:
>In article <0E18682EC2FFDE51.76C0D477...@lp.airnews.net>,
>no_...@no.where.com (Robert A. King) writes:
>
>>
>>>StereoStereo: Intelligent Solutions for Intelligent Homes
>>
>>I find this slogan very ironic...
>>
>
>I do not.
Of course you don't. I wouldn't expect less (or more,) from you.
-- RK
Robert A. King
wrote:
>You are all right. Because I can hear a difference and believe it is nothing to
>do with a placebo effect, this does not constitute it as fact rather it is my
>opinion.
OK, I'll agree with that.
>Because you believe that there can be no difference because it is
>digital data and all data can be measured as being transmitted and recieved
>does not mean that I am wrong in my assertion that I can hear a difference does
>it?
We simply fail to understand how there can be a difference in sound
*apart* from the placebo effect, given that the *only* difference is
in the digital transmission cable.
>If it does then I am happy I am not scientist and do not have to validate my
>own experiences to prove they actually happen!
Sometimes its hard to divorce yourself from your preconceptions and
observations. Those who can and readily do make for excellent
engineers, scientists, and doctors.
-- RK
Murray Peterson
wrote:
>
>ok ok . I agree it is not a controlled test. I have only ever taken part in one
>double blind test. If someone wants to arrange it I will happily take part in a
>double blind test of cable carrying digital data between cd and dac. I am sure
>I could percieve the differences.
>
And I am sure you couldn't, but hey, that's beside the point. Until
such a test is actually performed, neither one of us can assert your
beliefs as being true or false.
>You are all right. Because I can hear a difference and believe it is nothing to
>do with a placebo effect, this does not constitute it as fact rather it is my
>opinion. Because you believe that there can be no difference because it is
>digital data and all data can be measured as being transmitted and recieved
>does not mean that I am wrong in my assertion that I can hear a difference does
>it?
>
>If it does then I am happy I am not scientist and do not have to validate my
>own experiences to prove they actually happen!
>
No, my personal assertions are not fact either. However (and this is
a big one), there have been such tests performed, and all of them have
shown that there is no perceptible difference in the sound of cables.
Those test results ARE accepted as being factual. Until someone
performs and publishes a contradicting test, AND this test is repeated
by other parties, then the existing conclusion must stand.
Science is not used to validate one's experiences; it is used to
advance the current state of our knowledge about the world. Your own
experiences (and mine) are irrelevant to this process, except as
follows: you have the tools at your disposal to completely prove our
assertions as being incorrect (double blind testing). That's the fun
part about science -- proving someone to be totally wrong. Have fun,
prove us wrong, but until then, don't just claim that we are wrong.
Murray Peterson
It certainly does apply :-) One thing I find notable about that
paper, is that they completely (and maybe even conveniently?) fail to
mention that increased jitter directly causes increased harmonic
distortion. Why spend "five figures US$", when harmonic distortion is
much cheaper to measure, and will immediately show if any significant
jitter is present. Seems like a pretty glaring oversight...
Howard Ferstler
>
> On 24 Aug 1999 18:42:29 GMT, gdfr...@aol.comRemoveMe (Gordon Fraser)
> wrote:
> >You are all right. Because I can hear a difference and believe it is nothing to
> >do with a placebo effect, this does not constitute it as fact rather it is my
> >opinion.
> >digital data and all data can be measured as being transmitted and recieved
> >does not mean that I am wrong in my assertion that I can hear a difference does
> >it?
> *apart* from the placebo effect, given that the *only* difference is
> in the digital transmission cable.
> >own experiences to prove they actually happen!
> observations. Those who can and readily do make for excellent
> engineers, scientists, and doctors.
They are also more complete human beings, because they are more
effectively performing one of the most vital operations a human
being can do: thinking. What's more, they can feel confident that
what they are hearing is real and not something cooked up inside
of their heads. There is no substitute for being sure of
something.
I marvel that some enthusiasts appear to want to live in a
fantasy-audio world, where their imaginations are the reference
standard. You would think that "real" hi-fi enthusiasts would
want to be absolutely and positively sure that the components
they had were as exact in in their reproductive abilities as
possible, and that no superfluous or useless and overpriced
products were involved. Going this route both insures that that
there will be more cash left over for important things like
recordings, and also helps to settle the nerves of buffs who are
forever wondering if what they own is as good as possible.
If you want to be sure, why not do the comparison scientifically,
or at least try to comprehend what others have done
scientifically, and then be done with it. It is a very restful
experience.
Howard Ferstler
Author of The Home Theater Companion
Brian Short
Murray Peterson <m...@home.com> wrote in message
news:37c4228b.140377571@news...
> On 24 Aug 1999 18:42:29 GMT, gdfr...@aol.comRemoveMe (Gordon Fraser)
> wrote:
>
> >
in one
> >double blind test. If someone wants to arrange it I will happily take
part in a
> >double blind test of cable carrying digital data between cd and dac. I am
sure
> >I could percieve the differences.
> >
> And I am sure you couldn't, but hey, that's beside the point. Until
> such a test is actually performed, neither one of us can assert your
> beliefs as being true or false.
I'd definitely be interested in seeing the results of such a test. How are
these usually run? It would seem to me you would have to play multiple
sources through each cable, and then randomize when each cable is used (for
example, for one source, play the high end cable first, then the cheap
cable, then on a later source, use the cheap one first) to eliminate the
chance of simply GUESSING which is the expensive cable.
I did get my coax digital audio cable yesterday for when I get my Kenwood
HTB-502... $5
Brian Short
bds...@hotmail.com
Murray Peterson
<bds...@hotmail.com> wrote:
>I'd definitely be interested in seeing the results of such a test. How are
>these usually run? It would seem to me you would have to play multiple
>sources through each cable, and then randomize when each cable is used (for
>example, for one source, play the high end cable first, then the cheap
>cable, then on a later source, use the cheap one first) to eliminate the
>chance of simply GUESSING which is the expensive cable.
>
>I did get my coax digital audio cable yesterday for when I get my Kenwood
>HTB-502... $5
>
>Brian Short
>bds...@hotmail.com
>
There are many methods of running a double blind test (and I won't get
into many details, since I don't profess to be a expert on the
subject). However, the core method of eliminating someone just
guessing, is to run enough trials to be able to perform statistical
analysis on the results. Anyone can guess which side a coin will land
on correctly once or twice. Nobody can consistently do it 10 times in
a row. So you run mutiple trials, and see if the testee performed any
better (or worse) than the expected average.
As far as switching the cables, it must be more than random; it must
be unknown at the time of testing. That's what "double" blind means;
neither the testee NOR the tester know which cable is currently
plugged in. There is a company (anyone know their web site?) which
builds ABX switches to make this whole process automated: it randomly
selects a cable(X), and the testee has to decide if the cable is
really cable A or cable B. At the end of the test run, you can get a
dump of the actual cable selected each time, and the testee's
"guesses".
However, some purists object to such a box, since it would "colour"
the sound. In cases like these, the ABX box is a convenience, not a
necessity. Just set up the amp and DVD player in a different room,
and have a third person manually switch the cables when told to do so
(tester and testee are in the listening room). I would think a coin
flip is sufficient to decide which cable to switch to. The big
requirement here is that the person switching cables never sees,
hears, or talks to the person being tested (and vice-versa). This is
to prevent any possible passing of information (intentional or
otherwise).
Gary Sanford
wrote:
>In article <19990822163404...@ng-fm1.aol.com>, khan...@aol.com
>(KHanawalt) writes:
>
>>
>>Say, TAG McLaren Audio wouldn't be trying to sell high quality (read
>>expensive)
>>cables, would they?
>>
>
>Yes they do manufacture their own cables as well as their own high quality
>hi-fi and home cinema equipment. I don't know whether you would call them
>expensive though.
>
>The guys that were part of the team designing these cables and components are
>also the guys who design the telemetry systems used in West McLarens F1 racing
>cars, as well as around 75% of the formula 1 grid. They send digital
>information about the cars set up etc to the pits by rf while they are racing
>round the track. I believe they call this a "hostile enviroment", although it
>doesn't appear to be as hostile as this newsgroup : )
>
>These guys know a bit about digital signal transmission! Whether you believe
>them or want to put their equipment to the test is entirely up to you of
>course.
>
>It is extremely easy to show a difference in sound quality between two
>interconnects carrying digital information. When used in a dvd/cd player to
>transport situation.
>
>Rather than criticise why don't you go and read the articles. They try to give
>a scientific reason as to why these differences are heard.
Digital cables can cause problems in poorly designed DACs With a well
made DAC, signal correlated jitter is irrelevant because the bitstream
will be read into a FIFO and asynchronously reclocked with a reference
oscillator.
There's a lot of equipment makers out there that say you need fancy
cables just to fix their crappy designs.
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Gary Sanford
<shet...@cyberzone.net> wrote:
>Ed Jay wrote:
>> Not only does the cable need to convey the bits in the correct order and
>> voltages, but it also cannot introduce "ringing" that could be sensed as
>> additional bits or noise, but the rise and fall times of the signal must be
>> retained. An overly reactive cable can destroy a digital signal. It can
>> turn a square wave into a sine wave if it is really poor quality.
>
>I think the whole point of this discussion is not whether a digital
>connection can go bad or not -- certainly they can, be they optical or
>coaxial. The point is whether more subtle differences, that end up
>translating into real audible effects, can exist between cables that
>aren't failing in an obviously way with clicks, pops, chirps, or dead
>silence.
>
>Within in a fairly wide range of tolerance for cable connections, cable
>types, impedance matching, etc., completely error-free transmission of
>bits happens routinely with ordinary consumer gear and cables. Nothing
>fancy is required. Whatever noise and distorted wave shapes and ringing
>you're worrying about, it's rare that those effects are of great enough
>magnitude to cause failure. When the effects are that large, failure
>tends to be obvious.
>
>When talking about home theater, and therefore AC-3 or DTS audio signals
>most of the time, even the dread audiophile issue of jitter should have
>very little to do with the cables used, given the buffering,
>calculating, and re-clocking needed to process these compressed audio
>bit streams.
That's the problem. A lot of digital gear doesn't do reclocking. A
clock signal is sent along with the data and is "recovered" with
a phase lock loop chip. If the corners of the clock signal get
dispersed in any way by the cable you will get errors at
reconstruction.
Feeding the data into a buffer and reclocking it with a new stable
clock solves the problem. It's immune to transmission problems.
If you get difference in cables (and you can) it's not the cables
fault.
It's the badly designed equipment.
>
>I'm very skeptical about claims of "tighter bass", "greater air",
>"improved imaging", and so forth, made about digital cables.
>
>-Kerry
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Gary Sanford
wrote:
snip...
Here's a good article on jitter:
http://www.digido.com/jitteressay.html
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Murray Peterson
wrote:
Actually, it's a horrible article on jitter. Anyone that starts
making unproven assertions about "reduced soundstage width" and other
such things have destroyed their credibility beyond repair. Read the
other posts in this thread -- I think I said enough to make my point.
Gary Sanford
wrote:
>Comments interspersed below:
>Murray Peterson
>
>On Mon, 23 Aug 1999 15:33:24 -0400, Rick Nelson <ri...@csciences.com>
>wrote:
>
>>I'm the last person to start any sort of war on NG's, but I really must inteject
>>here. Jitter is real, and it has real consequences.. ask any EE out there who
>>specializes in data communications (like me).
>I also specialize in digital communications (low bandwidth, < 1MHz)
>
>>There are several requirements for digital transmissions, correct bit order and
>>voltage tolerances are certainly among the most important. However, for
>>applications like real-time audio, you need to make sure that the timing is
>>correct. Real-time transmissions need to be synchronized (e.g., via a
>>Phase-locked loop), because there is no major buffering or data storage. There
>>are other considerations to be be made if the cable is of low quality -- bits
>>could be lost, sampling rates could become irregular, inductance of the wire can
>>cause all sorts of neat effects...
>>
>Ah, but the nice thing about digital audio is that we only need to
>re-create the digital bit stream, not an analog bit stream. Which
>means that your link below has no meaning in this application
>(re-creating an analog square wave). All we need to do is get the bit
>value correct, and we don't care if it has been degraded to a sine
>wave (or any other wave form). The harmonics are of no importance
>whatsoever; in fact, a sine wave actually causes less problems than a
>square wave. When sampling a digital bit stream, the <only>
>requirement for perfect recreation of it is that we get each bit
>correct (and error-correction codes even relax that requirement).
Not true. Reclocking the data is important too. It's not only data,
it's data and timing. With perfect data and a badly recovered clock,
errors are created. See:
http://www.digido.com/jitteressay.html
Of course the right way is to reclock the data with a new master
clock. Too bad it's not usually done that way in consumer audio gear.
>Over a 2 meter length, and at the bandwidth in question, I would be
>quite happy with the proverbial coat hanger (I would like it
>shielded).
>
>>> Double blind testing is, and always will be the ONLY way to determine
>>> differences. End of story.
>>
>>Not sure about the ONLY way... but it's certainly the best way in this case..
>>
>For arguments about subjective "sound", I don't know of any other way.
>If I put a scope on the incoming signal and prove that the bit stream
>has been perfectly re-created, would that convince anyone of the
>validity of my argument (that cables make no difference)? I will
>probably just get someone asserting that my measurements must be
>meaningless because <they> can hear a difference. I think the only
>real proof is a double blind test; that way the person can be proven
>incorrect in a conclusive manner. Of course, they then start whining
>about double blind tests being faulty, at which point we might as well
>all go home -- superstition cannot be reasoned with...
>>
>>Oh yeah... go here... http://cyclone.stanford.edu/~davros/jitter/jitter.html
>Went there, see above.
>>
>>Rick
>
>Murray Peterson
>Homepage: http://members.home.net/murraypeterson
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Murray Peterson
wrote:
>On Tue, 24 Aug 1999 02:02:01 GMT, m...@home.com (Murray Peterson)
>wrote:
>>square wave. When sampling a digital bit stream, the <only>
>>requirement for perfect recreation of it is that we get each bit
>>correct (and error-correction codes even relax that requirement).
>
>Not true. Reclocking the data is important too. It's not only data,
>it's data and timing. With perfect data and a badly recovered clock,
>errors are created. See:
>http://www.digido.com/jitteressay.html
Went there; awful paper. Makes unfounded assertions about sound stage
width, instruments losing focus, etc. So much for <their>
credibility.
>
I agree that without re-clocking, timing errors can be introduced. If
you feed this signal directly to a D/A converter, then there will be
effects (the rest of this thread talks about their significance).
>Of course the right way is to reclock the data with a new master
>clock. Too bad it's not usually done that way in consumer audio gear.
>
I would actually expect most Dolby Digital receivers and preamps to
reclock the data before it hits the D/A converters. I took a peek
inside mine, and the signal is processed by 2 DSP chips before being
converted to analog (definitely not using the DVDs FIFO clock signal
by that point).
Gary Sanford
wrote:
I don't necessarily agree with that subtle sound staging stuff
either. But, program correlated, jitter induced, IM spuriae is a real
and easily measurable phenomena. My main point is that if
there _are_ any real differences in digital cables, it's not the
cables fault (as asserted by some) but the associated equipments
fault. As long as the data is reclocked at the DAC ,correctly,
no problem.
Yeah, the data is pretty immune to corruption in digital media,
but that's not the whole story. I'm mostly familiar with the S/PDIF
interface used in 2 box CD players. The word clock is multiplexed in
with the data. In some players it is stripped out and used to slave
an oscillator to the DAC. This is not a good idea. The master clock
should be at the DAC (where it counts) and the CD reader should be
slaved to it.
I agree that your DD receiver with the DSP's is probably immune
to cable induced jitter. I'm really not familiar with how all DVD/DD
data streams are moved and decoded. If they all do it that way,
so much for the better.
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Luke Oliver
high speeds, or in a very noisy environment. Anyone who thinks that a
digital cable in a home theater system is going to make an audible
difference needs to put down thier crack pipe for a few minutes and
think about this.
You either have a 1 or a 0. no in betweens, no sine waves, just on or
off.
If you have spent big money on a digital cable then you don't have any
busiess arguing about it with educated people. Anyone who knows
anything at all about digital knows that it doesn't make the slightest
difference at these speeds.
If you bought an expensive cable and you do hear a difference: This is
your brain refusing to belive that you have been ripped off.
Most of what I post in here is my opinion....THIS IS FACT....get a clue
Luke
Robert A. King
wrote:
>On Tue, 24 Aug 1999 14:28:07 GMT, no_...@no.where.com (Robert A. King)
>wrote:
>snip...
>
>Here's a good article on jitter:
>http://www.digido.com/jitteressay.html
Yeah, I've read it. That doesn't apply to low-speed transmission,
such as 96kHz PCM data.
-- RK
Robert A. King
wrote:
>That's the problem. A lot of digital gear doesn't do reclocking. A
>clock signal is sent along with the data and is "recovered" with
>a phase lock loop chip. If the corners of the clock signal get
>dispersed in any way by the cable you will get errors at
>reconstruction.
Not if they get "dispersed in any way" (whatever thay means,), but
rather if the clock waveform is distorted enough to cause it to be
misinterpreted. And in practice, that's pretty difficult to do.
Designing circuitry to accurately receive and interpret the signal has
been pretty highly refined for years. A 2-meter cable, or a 20-meter
cable of decent quality simply isn't going to distort the signal
enough. Shoot, 2 meters of wire-wrap wire isn't going to distort the
signal enough.
>Feeding the data into a buffer and reclocking it with a new stable
>clock solves the problem. It's immune to transmission problems.
That's a pretty powerfull blanket statement.
If you can't trust the incoming clock signal (which is your point,)
how can you trust the data that came in via the same cable?
-- RK
Dave Beal
"Robert A. King" wrote:
> On Wed, 25 Aug 1999 07:16:30 GMT, sanf...@sprynet.com (Gary Sanford)
> wrote:
>
> >That's the problem. A lot of digital gear doesn't do reclocking. A
> >clock signal is sent along with the data and is "recovered" with
> >a phase lock loop chip. If the corners of the clock signal get
> >dispersed in any way by the cable you will get errors at
> >reconstruction.
>
> <snip>
>
> >Feeding the data into a buffer and reclocking it with a new stable
> >clock solves the problem. It's immune to transmission problems.
>
> That's a pretty powerfull blanket statement.
>
> If you can't trust the incoming clock signal (which is your point,)
> how can you trust the data that came in via the same cable?
>
I think you're misunderstanding Gary's point.
The problem isn't that the received data is going to be
in error (i.e., mistaking ones for zeroes).
It's more subtle than that. The problem is that
the clock edges "slide" forward and backward
in time, so that the bits and words don't get clocked
into the receiving device at exactly the right time.
If the receiving device doesn't reclock the words
with an accurate, independent clock, this error
will persist right up to the DAC, causing the
analog output to be distorted, because the samples
appear at the output slightly before or after they should.
Robert A. King
>> >Feeding the data into a buffer and reclocking it with a new stable
>> >clock solves the problem. It's immune to transmission problems.
>>
>> That's a pretty powerfull blanket statement.
>>
>> If you can't trust the incoming clock signal (which is your point,)
>> how can you trust the data that came in via the same cable?
>>
>
>I think you're misunderstanding Gary's point.
>The problem isn't that the received data is going to be
>in error (i.e., mistaking ones for zeroes).
>It's more subtle than that. The problem is that
>the clock edges "slide" forward and backward
>in time, so that the bits and words don't get clocked
>into the receiving device at exactly the right time.
But in a digital signal, it doesn't matter EXACTLY WHEN the bit
arrives, just so that it arrives before the next bit does.
>If the receiving device doesn't reclock the words
>with an accurate, independent clock, this error
>will persist right up to the DAC, causing the
>analog output to be distorted, because the samples
>appear at the output slightly before or after they should.
Nope. Doesn't work that way. The absolute timing of the signals are
irrelevant, so long as they conform to the specification for the
signal, which allows for considerable leeway in the absolute timing.
-- RK
Rick Nelson
> arrives, just so that it arrives before the next bit does.
That type of thing completely depends on the application... but it's my
guess (I could be wrong here) that there isn't any signifigant buffering
on the Rx side -- you want the data in real-time so that it corresponds to
the picture on the screen.
Now, when data falls out of phase, it may take several (or 100) clock
cycles to re-sync. During that time, you lose data -- it's getting
dropped if there are no FIFOs.
>
> Nope. Doesn't work that way. The absolute timing of the signals are
> irrelevant, so long as they conform to the specification for the
> signal, which allows for considerable leeway in the absolute timing.
>
> -- RK
That's true. It all depends on how much throughput you're trying to get
out of it.
Rick
--
Chesapeake Sciences Corp.
1127B Benfield Blvd Millersville, MD 21108
Tel: (410) 923-1300 x3430 Fax: (410) 923-2669
Robert A. King
wrote:
>> But in a digital signal, it doesn't matter EXACTLY WHEN the bit
>> arrives, just so that it arrives before the next bit does.
>
>That type of thing completely depends on the application... but it's my
>guess (I could be wrong here) that there isn't any signifigant buffering
>on the Rx side
In practice there's almost ALWAYS buffering going on for digital data
streams. Even the lowly RS-232 port of your garden-variety has a
16-byte buffer these days (16550N UART). Generally, not buffering
your data stream meand you have to be VERY sure that the OS can
service the interrupt before the next bit (or byte, depending on the
hardware design,) can arrive. Buffering the data is a very
cost-effective way of lessening this time constraint.
>Now, when data falls out of phase, it may take several (or 100) clock
>cycles to re-sync. During that time, you lose data -- it's getting
>dropped if there are no FIFOs.
If you loose sync, you loose data -- period. You never have it, so it
can't be put into a buffer.
Secondly, loosing sync usually means you loose the byte. Depending on
the interface spec, the byte can either be dropped (lost,) or it can
be re-transmitted immediately, or, if you're using a packetized
protocol, the packet fails its checksum, and THAT is retransmitted.
Since the data rate over the cable is much higher than what is
required for realtime playback, any corrupted packets can be
re-transmitted. This is fairly typical for data transmission, and
even goes way back to the edays of Xmodem file transfers (anyone
remember THAT?)
>That's true. It all depends on how much throughput you're trying to get
>out of it.
And remember, at most we're talking about a 96kHz PCM data stream --
pretty low bandwidth.
-- RK
Dave Beal
"Robert A. King" wrote:
<snip>
> But in a digital signal, it doesn't matter EXACTLY WHEN the bit
> arrives, just so that it arrives before the next bit does.
If the variance in interarrival times makes its way
through the DAC and appears in the analog output,
it matters.
>
> >If the receiving device doesn't reclock the words
> >with an accurate, independent clock, this error
> >will persist right up to the DAC, causing the
> >analog output to be distorted, because the samples
> >appear at the output slightly before or after they should.
>
> Nope. Doesn't work that way. The absolute timing of the signals are
> irrelevant, so long as they conform to the specification for the
> signal, which allows for considerable leeway in the absolute timing.
Think about what's happening at the DAC. A digital
sample comes in and some fixed time later, appears
as an analog voltage at the output. The stairstepped
analog output is then smoothed by a low-pass filter
to recreate the original analog signal. If the rate that
the digital samples arrive at the DAC varies, the
corresponding analog output voltages also shift
in time, distorting the recreated signal.
Dave Beal
Rick Nelson wrote:
> > But in a digital signal, it doesn't matter EXACTLY WHEN the bit
> > arrives, just so that it arrives before the next bit does.
>
> That type of thing completely depends on the application... but it's my
> guess (I could be wrong here) that there isn't any signifigant buffering
> on the Rx side -- you want the data in real-time so that it corresponds to
> the picture on the screen.
>
> Now, when data falls out of phase, it may take several (or 100) clock
> cycles to re-sync. During that time, you lose data -- it's getting
> dropped if there are no FIFOs.
We're not talking about clock shifts so large that the PLL
loses sync. We're just talking about small shifts in the clock
derived from the received data stream. If a jittery received
clock is used as input to the DAC, rather than retiming the
data to an accurate independent clock, the signal at the
output of the DAC will be affected.
Robert A. King
>We're not talking about clock shifts so large that the PLL
>loses sync. We're just talking about small shifts in the clock
>derived from the received data stream. If a jittery received
>clock is used as input to the DAC, rather than retiming the
>data to an accurate independent clock, the signal at the
>output of the DAC will be affected.
I see no reason to feed a clock signal into the DAC. Rather, it makes
MUCH more sense to have a timed interrupt feeding data from an input
buffer into the digital-to-analog DSP chip.
To carry this discussion any farther, we're going to need a DETAILED
design spec of the receiving hardware and software for several signal
receivers, BEFORE the data stream hits the DSP chip.
However, I'd be VERY surprised to see anyone using an external clock
signal to drive an internal DSP. It's VERY stupid.
-- RK
Gary Sanford
>
>
>"Robert A. King" wrote:
>
>> On Wed, 25 Aug 1999 07:16:30 GMT, sanf...@sprynet.com (Gary Sanford)
>> wrote:
>>
>> >That's the problem. A lot of digital gear doesn't do reclocking. A
>> >clock signal is sent along with the data and is "recovered" with
>> >a phase lock loop chip. If the corners of the clock signal get
>> >dispersed in any way by the cable you will get errors at
>> >reconstruction.
>>
>> <snip>
>>
>> >Feeding the data into a buffer and reclocking it with a new stable
>> >clock solves the problem. It's immune to transmission problems.
Sorry, I meant comparatively immune. I was posting at around 3AM.;-)
>>
>> That's a pretty powerfull blanket statement.
>>
>> If you can't trust the incoming clock signal (which is your point,)
>> how can you trust the data that came in via the same cable?
>>
>
>I think you're misunderstanding Gary's point.
>The problem isn't that the received data is going to be
>in error (i.e., mistaking ones for zeroes).
>It's more subtle than that. The problem is that
>the clock edges "slide" forward and backward
>in time, so that the bits and words don't get clocked
>into the receiving device at exactly the right time.
>If the receiving device doesn't reclock the words
>with an accurate, independent clock, this error
>will persist right up to the DAC, causing the
>analog output to be distorted, because the samples
>appear at the output slightly before or after they should.
>
Yeah that's pretty much it. If you look at a recovered clock
with a digital storage scope and overlay a bunch of samples,
the transitions can get pretty thick. That's what I meant by
dispersion.
Look, I'm no expert at this, but this is not a black and white
issue. With some folks saying that data is data and incorruptible
and others hearing all sorts of differences all time, the truth is
some where in between, IMO.
The 1's and 0's are pretty robust in transmission. The clock
is easier to corrupt. Even picoseconds of error will cause
differences in the analog output. How audible is it?
I don't know.
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Gary Sanford
wrote:
>On Wed, 25 Aug 1999 10:16:29 -0600, Dave Beal <d...@mcdata.com> wrote:
>
>
>>> >Feeding the data into a buffer and reclocking it with a new stable
>>> >clock solves the problem. It's immune to transmission problems.
>>>
>>> That's a pretty powerfull blanket statement.
>>>
>>> If you can't trust the incoming clock signal (which is your point,)
>>> how can you trust the data that came in via the same cable?
>>>
>>
>>I think you're misunderstanding Gary's point.
>>The problem isn't that the received data is going to be
>>in error (i.e., mistaking ones for zeroes).
>>It's more subtle than that. The problem is that
>>the clock edges "slide" forward and backward
>>in time, so that the bits and words don't get clocked
>>into the receiving device at exactly the right time.
>
>But in a digital signal, it doesn't matter EXACTLY WHEN the bit
>arrives, just so that it arrives before the next bit does.
>
As long as it stays digital, of course it doesn't matter.
>
>>If the receiving device doesn't reclock the words
>>with an accurate, independent clock, this error
>>will persist right up to the DAC, causing the
>>analog output to be distorted, because the samples
>>appear at the output slightly before or after they should.
>
>Nope. Doesn't work that way. The absolute timing of the signals are
>irrelevant, so long as they conform to the specification for the
>signal, which allows for considerable leeway in the absolute timing.
>
>-- RK
Until the data has to be reconverted back to an analog waveform.
The correct data at the wrong time will have amplitude errors.
What you're saying is that a 100 Hz wave has the same timing as
a 1000 Hz wave. 'Taint so. Digital to analog conversion is ultimately
an analog process and that allows the gremlins to creep back in.
~~~~~~~~~~~~~~~~
Gary Sanford
sanf...@ibm.net
Robert A. King
>
>
>"Robert A. King" wrote:
><snip>
>
>> But in a digital signal, it doesn't matter EXACTLY WHEN the bit
>> arrives, just so that it arrives before the next bit does.
>
>If the variance in interarrival times makes its way
>through the DAC and appears in the analog output,
>it matters.
I talk about this in another post.
But I do have a question. Are you SURE the clock frequency on the
cable is the SAME as the sampling frequency multiplied by the
word-length (in bits)?
-RK
Dave Beal
Gary Sanford wrote:
> Yeah that's pretty much it. If you look at a recovered clock
> with a digital storage scope and overlay a bunch of samples,
> the transitions can get pretty thick. That's what I meant by
> dispersion.
>
> Look, I'm no expert at this, but this is not a black and white
> issue. With some folks saying that data is data and incorruptible
> and others hearing all sorts of differences all time, the truth is
> some where in between, IMO.
>
> The 1's and 0's are pretty robust in transmission. The clock
> is easier to corrupt. Even picoseconds of error will cause
> differences in the analog output. How audible is it?
> I don't know.
>
I don't, either, but last night I read a paper by some guys who do.
It was a publication by the Audio Engineering Society, written
by two guys at Dolby Labs. The title was something like
"Theoretical and practical effects of jitter in digital audio
systems". Among other things, they measured the amount of jitter
inserted by a 30 meter 75 ohm digital interconnect cable. It
was less than 1 nsec across the entire audio spectrum.
They also did some subjective tests to determine the threshold
of audibility of jitter in real audio program material. Their
conclusion was that *some* listeners could distinguish as little
as 30 nsec of jitter in *some* material. In other words,
the amount of jitter caused by a very long cable was 1/30 the
amount of jitter that some people can just barely hear when
specifically listening for it.
To each his own, but I'm going to continue to buy my
digital interconnect cables at Radio Shack.
Robert A. King
wrote:
>
> Until the data has to be reconverted back to an analog waveform.
>The correct data at the wrong time will have amplitude errors.
>What you're saying is that a 100 Hz wave has the same timing as
>a 1000 Hz wave. 'Taint so. Digital to analog conversion is ultimately
>an analog process and that allows the gremlins to creep back in.
It all turns on what's done to the data AFTER it is taken off the
cable, but before it's converted to analog.
Unless the interface is EXTREMELY primitive, the data should be
re-timed. I won't say re-clocked because that implies generating a
new clock signal, which I don't believe happens.
Rather, I suspect the data is either being buffered, or is being
received a packet at a time, and then, upon completion of the packet,
the whole packet or buffer is being fed into the DSP. In either case,
that requires independant timing (probably done by the DSP chip.)
-- Robert
Robert A. King
wrote:
>Look, I'm no expert at this, but this is not a black and white
>issue. With some folks saying that data is data and incorruptible
>and others hearing all sorts of differences all time, the truth is
>some where in between, IMO.
>
>The 1's and 0's are pretty robust in transmission. The clock
>is easier to corrupt.
Agreed.
>Even picoseconds of error will cause
>differences in the analog output.
Of that, I'm not convinced, but I can understand how it COULD happen,
given some very specific assumptions about how the digital signal is
handled at the receiving end. Unfortunately it's been ages since I
looked at this stuff in any great detail, and I no longer have a
low-level specification. Can you suggest a source for TOSLINK,
S/PDIF, or anything else?
-- Robert
Robert A. King
>>>The 1's and 0's are pretty robust in transmission. The clock
>>>is easier to corrupt.
>>
>>Agreed.
>>
>>>Even picoseconds of error will cause differences in the analog output.
>>
>>Of that, I'm not convinced, but I can understand how it COULD happen,
>
>I'm sure that if we can understand how it could happen, then it probably
>does. That said, though, the question isn't really whether or not an
>electronic phenomena can exist, but rather if it does, is it perceivable by
>mere humans without sophisticated test equipment or 'golden ears.' How bad
>does it have to be before it has a negative impact on the listener?
Someone referenced a paper presented by Dolby Labs (though I forget
which forum or thread it was in,) that indicated in a typical 2 meter
cable (it was unclear whether this was a coax cable or TOSLINK,)
timing errors were typically 1 picosecond.
the paper apparently also stated that "some" listeners were able to
hear anomalys when the timing errors were as short as 30 picoseconds.
Apparently the paper's conclusion was that while digital jitter did
indeed exist, for all practical purposes, it was a non-issue.
>Mr. King, I recently took issue with what I felt was your categorical
>denial that the problem exists.
I chose my words badly. My take on the situation is the same as
Dolby's. It's there, but it isn't affecting the sound enough to make
it an issue.
From a purely technical side, I can also see how digital jitter can be
eliminated all together by simply buffering or re-clocking the data
before it hits the DAC (which may be happening anyway -- I haven't
been able to get any details on anyone's implementation of the link.)
-- RK
Dave Beal
> Someone referenced a paper presented by Dolby Labs (though I forget
> which forum or thread it was in,)
It was me, in this thread yesterday.
> that indicated in a typical 2 meter
> cable (it was unclear whether this was a coax cable or TOSLINK,)
It was a 30 meter 75 ohm coax.
>
> timing errors were typically 1 picosecond.
Up to one nanosecond.
>
> the paper apparently also stated that "some" listeners were able to
> hear anomalys when the timing errors were as short as 30 picoseconds.
30 nanoseconds, but you've got the idea.
>
> Apparently the paper's conclusion was that while digital jitter did
> indeed exist, for all practical purposes, it was a non-issue.
Jitter caused by cables is a non-issue, at least.
>
> >Mr. King, I recently took issue with what I felt was your categorical
> >denial that the problem exists.
>
> I chose my words badly. My take on the situation is the same as
> Dolby's. It's there, but it isn't affecting the sound enough to make
> it an issue.
>
> From a purely technical side, I can also see how digital jitter can be
> eliminated all together by simply buffering or re-clocking the data
> before it hits the DAC (which may be happening anyway -- I haven't
> been able to get any details on anyone's implementation of the link.)
Sounds like we're now in agreement. Although the fact that
the guys at Dolby chose to evaluate cable jitter seems to imply
that consumer devices don't reclock the data.
Robert A. King
>"Robert A. King" wrote:
>
>> Someone referenced a paper presented by Dolby Labs (though I forget
>> which forum or thread it was in,)
>
>It was me, in this thread yesterday.
>
>> that indicated in a typical 2 meter
>> cable (it was unclear whether this was a coax cable or TOSLINK,)
>
>It was a 30 meter 75 ohm coax.
>
>>
>> timing errors were typically 1 picosecond.
>
>Up to one nanosecond.
>
>>
>> the paper apparently also stated that "some" listeners were able to
>> hear anomalys when the timing errors were as short as 30 picoseconds.
>
>30 nanoseconds, but you've got the idea.
OK, jitter in a 30 meter 75-ohm cable is typically 1 nanosecond.
Audible jitter is typically 30 nanoseconds and usually greater. Sorry
I got my units and other details mucked up.
>Jitter caused by cables is a non-issue, at least.
Which brings up a question -- What other sources of jitter are there?
>Sounds like we're now in agreement. Although the fact that
>the guys at Dolby chose to evaluate cable jitter seems to imply
>that consumer devices don't reclock the data.
Maybe not. I've also read that Meridian is the only company to
re-clock their data.
It seems odd to me that anyone would design an interface that didn't
re-clock the data, or buffer it (which would require reclocking the
data when its removed from the buffer.) I wonder if the DSP chips
being used re-clock the data internally. I'd need to know more about
the inner workings of both the transmission spec and the DSP's (and
how the two are joined,) to make a better analysis.
-- RK