Jitter: Why CD Transports Sound Different
lich
distortion, it is a bad thing. The less of it there is
in your digital signal chain, the happier an audiophile
you will be.
The essential purpose of a CD transport is to read data
and deliver it without jitter.
The essential purpose of CD transport manufacturers is
to get you to buy their transport, rather than someone
else's. You can buy a box to read data from a CD for
$200, and you can buy a box to read data from a CD for
$20,000.
Therefore, wishing to be happy audiophiles, we try
different CD transports with our ancillary equipment
and we typically end up buying the transport that
sounds best to our audiophile ears, sometimes spending
a great deal of money, often in the form of credit card
debt.
Inside our new, pricey device may lurk a nasty
surprise: the actual transport mechanism may be the
very same model used inside the three pound Sony your
tin-eared neighbor got for $249 at Circuit City.
You borrow the neighbor's Sony and try it out. It does
not sound as good as yours, the harmonics seem thinned
out, or the soundstage has shrunk, or inner detail is
missing, or you suddenly find yourself hating sopranos
again, etc. The multiple power supplies with their
overzied bank of filter capacitors, the massive
chassis, proprietary circuits, temperature-controlled
crystal oscillator, damping material, special feet, of
your new transport, they all contribute to the effect.
The effect is, of course, multifaceted, unless the
comparison is done blind. One of those facets, the
important one, is the actual sound of the two DUTs. And
I have been reliably informed that the *only* factor
which can account for any difference in sound
is...jitter.
Thus, the difference in sound between the neighbor's
$249 Sony and your thirty-eight pound $6500
Klangmeister Valkyrie is in the amount of jitter
infecting the digital stream at the output.
Any third party device, green pen, antistatic fluid,
cosmic ray shield, rubber mat with squiggley
fluorescent lines, surface polishing liquid, if in fact
they make an audible difference at all, can only do so
if they make a difference in the amount and frequency
of jitter.
This raises a number of interesting points. Is it
really necessary to sand-cast a forty pound chassis, to
place two pound weights on the CD, the over-engineer
power supplies, etc., in order to limit jitter? And how
about that proprietary clocking circuit inside that
massive chassis: couldn't such a circuit find its way
inside a transport costing less than $1000? And how
much jitter is audible? (One technical editor suggested
to me that levels up to several thousand picoseconds
were, in his opinion, probably inaudible.) What does
the *frequency* of the jitter have to do with how the
music sounds? Is *some* degree of jitter inevitable
simply do to the real-world limitations imposed by
voltage regulators and ground bounce? And why, why are
jitter specifications for CD transports not published?
--
"The dream of reason produces monsters"
-=Goya
Arny Krueger
Subject: Jitter: Why CD Transports Sound Different
>This raises a number of interesting points. Is it
>really necessary to sand-cast a forty pound chassis, to
>place two pound weights on the CD, the over-engineer
>power supplies, etc., in order to limit jitter?
IME, no
>And how
>about that proprietary clocking circuit inside that
>massive chassis: couldn't such a circuit find its way
>inside a transport costing less than $1000?
Circuits with low jitter find their way elsewhere for far less than
$1k
> And how much jitter is audible? (One technical editor suggested
>to me that levels up to several thousand picoseconds
>were, in his opinion, probably inaudible.)
You can investigate that for yourself at
http://www.pcabx.com/technical/jitter_power/index.htm
>What does the *frequency* of the jitter have to do with how the
>music sounds?
Zwicker and Fastl suggest that the lower the frequency of the jitter
the more audible it is. That means that not only do vinyl players and
analog tapes have orders of magnitude more jitter than good digital
equipment, they produce jitter at frequencies where it is more
audible.
>Is *some* degree of jitter inevitable imply do to the real-world
>limitations imposed by voltage regulators and ground bounce?
There is no such thing as zero distortion.
> And why, why are jitter specifications for CD transports not
> published?
One just does not see a lot of objective information about jitter, do
they. Other than tests in HFN&RR, Stereophile and test reports posted
at the www.pcavtech.com web site one does not see very much
objective figures for jitter, do they? Of the three only
www.pcavtech.com has measured jitter using a properly-dithered test
signal.
Justin Sullivan
>Jitter is a subject we've all heard about. It is
>distortion, it is a bad thing. The less of it there is
>in your digital signal chain, the happier an audiophile
>you will be.
>
>The essential purpose of a CD transport is to read data
>and deliver it without jitter.
>
>The essential purpose of CD transport manufacturers is
>to get you to buy their transport, rather than someone
>else's. You can buy a box to read data from a CD for
>$200, and you can buy a box to read data from a CD for
>$20,000.
For $20,000, you can get a box that can read and decode the data from
a CD. The Linn Sondek CD12 is $20K and it has a built-in DAC. The Mark
Levinson 31.5 Reference (at $17,000) is just a transport, though.
As one reviewer put it (when talking about the Sondek CD12): "For
12,000 pounds, you can either have a fairly crappy car, or one of the
best CD players made."
I would like to shake that man's hand.
>Inside our new, pricey device may lurk a nasty
>surprise: the actual transport mechanism may be the
>very same model used inside the three pound Sony your
>tin-eared neighbor got for $249 at Circuit City.
>
>You borrow the neighbor's Sony and try it out. It does
>not sound as good as yours, the harmonics seem thinned
>out, or the soundstage has shrunk, or inner detail is
>missing, or you suddenly find yourself hating sopranos
>again, etc. The multiple power supplies with their
>overzied bank of filter capacitors, the massive
>chassis, proprietary circuits, temperature-controlled
>crystal oscillator, damping material, special feet, of
>your new transport, they all contribute to the effect.
Most transports themselves are made by Pioneer and Philips, and the
odds are good that even in a $1,000 transport, it's the same one used
in the $249 model from Circus City (no, no a typo).
One thing that gets me is that even in some expensive transports, they
have cheap plastic drawers. The solid aluminum drawer found in
Proceed/Mark Levinson transports is, by itself, a powerful, seductive
force. (okay, I'm pushing it a little)
>The effect is, of course, multifaceted, unless the
>comparison is done blind. One of those facets, the
>important one, is the actual sound of the two DUTs. And
>I have been reliably informed that the *only* factor
>which can account for any difference in sound
>is...jitter.
Not the only one. The signal is still subject to RF noise and other
interference. An AES/BEU format signal will sound better (less noise)
than an RCA coaxial or Toslink optical output (probably the worst
connection available).
A fully balanced (in the XLR sense) system is more than just the
transport, of course, but my favorite test to show my tin-eared
neighbors the folly of their Circus City paperweight is to pause the
source and turn the volume way up.
"I don't hear anything" they say.
EXACTLY right. No snap crackle pop when it should be quiet. It only
takes one component to inject noise to ruin the magic spell.
That's because the only sound your system should make is music.
>Thus, the difference in sound between the neighbor's
>$249 Sony and your thirty-eight pound $6500
>Klangmeister Valkyrie is in the amount of jitter
>infecting the digital stream at the output.
>This raises a number of interesting points. Is it
>really necessary to sand-cast a forty pound chassis, to
>place two pound weights on the CD, the over-engineer
>power supplies, etc., in order to limit jitter? And how
>about that proprietary clocking circuit inside that
>massive chassis: couldn't such a circuit find its way
>inside a transport costing less than $1000? And how
>much jitter is audible? (One technical editor suggested
>to me that levels up to several thousand picoseconds
>were, in his opinion, probably inaudible.) What does
>the *frequency* of the jitter have to do with how the
>music sounds? Is *some* degree of jitter inevitable
>simply do to the real-world limitations imposed by
>voltage regulators and ground bounce? And why, why are
>jitter specifications for CD transports not published?
Jitter is present in every digital system to some degree. It's just
easier to notice in audio and video applications because everything
has to happen right on time and right in step. Several thousand
picoseconds is probably not noticeable. 50ms would be very noticeable.
That's enough to make a motion picture and its soundtrack fall out of
sync. It isn't the frequency of jitter, it's that each step in the
process of recreating music from a CD adds to it. The best a system
can do is reduce it to the lowest level possible.
There are two main sources of jitter in most transports: internally
generated and how well the output signal is timed. Free of internal
jitter would mean that it did a bang-up job of spinning the disc and
got all the data off of it exactly on time. A well-timed output would
mean that it was still pretty much on time when it left the unit. From
there, you can expect it to get worse. Even the time it takes to
travel along the cable to the DAC introduces jitter.
The quality of power in ANY system that uses it can make a profound
difference. In this case, it might be more important that the power
source is also shielded enough to keep from interfering with the
clocking circuit(s).
Is it necessary to put that big heavy weight on it? I'm no expert, but
it might make a difference in the quality of data recovered from the
CD if you reduced vibrations as much as possible.
I suppose the custom-made clocks used to drive the servo that spins
the disc and the electronics that handle the signal might find its way
into a sub-$1,000 transport. But there are other economic realities...
Sony (if memory serves) led the way to introducing cheaper components
into consumer gear. Part of it is always progress: better quality in
the future can always be had cheaper than it was before. Remember when
Sony introduced 1-bit DACs and said it was the best way? It isn't, but
it's good quality for the money: a high-quality multibit DAC can be
expensive. High-precision clocks can be expensive. When you're
building a player that a dealer can mark up to $249, you have to pinch
pennies.
As to why nobody seems to publish just HOW good they deal with jitter:
good question. I suspect it's because even though any reduction in
jitter in today's systems is quite an amazing engineering achievement,
it wouldn't look so drastic when held up against our low-end
"reference" (your tin-eared neighbor's $249 Sony).
WARNING: made up numbers follow
Why would you pay $20,000 instead of $249 when you look at a sheet of
paper and it says "0.0002%" for the expensive one and "0.0005%" for
the cheap one? One manufacturer even claims that the jitter in all of
their digital components is "virtually unmeasurable."
There are other benefits to consider in a high-end CD system, though
most of these aren't evident until we get the data to the DAC where we
can decode HDCD and other formats. I am reluctant to mention anything
that does this onthe same unit, because, like your tin-eared
neighbor's Sony, it's not a transport: it's a CD player/processor.
Think about it: even your TEST EQUIPMENT suffers from jitter!
- Justin
Richard D Pierce
lich <ob...@earthlink.net> wrote:
>Jitter is a subject we've all heard about. It is
>distortion, it is a bad thing. The less of it there is
>in your digital signal chain, the happier an audiophile
>you will be.
Only to a point.
>The essential purpose of a CD transport is to read data
>and deliver it without jitter.
100% absolutely WRONG. The essential purpose of a CD transport
is to read data and deliver it without DATA ERROR. The ONLY
place where jitter is of ANY relevance is at the moment of
conversion back to the analog domain. More specifically, the
ONLY place where the jitter is relevant is in the DAC's sample
and hold clock. Noweher else does jitter have ANY relevance.
>Thus, the difference in sound between the neighbor's
>$249 Sony and your thirty-eight pound $6500
>Klangmeister Valkyrie is in the amount of jitter
>infecting the digital stream at the output.
No, the difference may be due to the fact that your DAC is badly
designed and does NOT properly isolate itself from transport
jitter.
Other than these points, I don's necessarily disagree with the
rest of your post.
--
| Dick Pierce |
| Professional Audio Development |
| 1-781/826-4953 Voice and FAX |
| DPi...@world.std.com |
Stewart Pinkerton
>This raises a number of interesting points. Is it
>really necessary to sand-cast a forty pound chassis, to
>place two pound weights on the CD, the over-engineer
>power supplies, etc., in order to limit jitter?
No. In particular, the jitter of the signal actually coming off the
disc is irrelevant, since the data are totally reclocked in the FIFO
RAM buffer which is an essential part of every CD player and transport
ever made.
> And how
>about that proprietary clocking circuit inside that
>massive chassis: couldn't such a circuit find its way
>inside a transport costing less than $1000?
They do - the cheapest Arcam CD-7ES player has one of the lowest
jitters of anything on the market.
> And how
>much jitter is audible? (One technical editor suggested
>to me that levels up to several thousand picoseconds
>were, in his opinion, probably inaudible.)
A vexed question, and frequency-dependent.
> What does
>the *frequency* of the jitter have to do with how the
>music sounds?
LF jitter is more audible.
> Is *some* degree of jitter inevitable
>simply do to the real-world limitations imposed by
>voltage regulators and ground bounce?
Yes, but it can be very low. It's also important to remember that
jitter in the transport really doesn't matter, since jitter does not
propagate in an all-digital environment. The *only* place that jitter
matters is at the point of conversion to analogue, which is why it's a
*very* good idea to use the same low-noise free-running master clock
for both transport and DAC - which generally means a single-box
player.
> And why, why are
>jitter specifications for CD transports not published?
Hit squads from Levinson and CEC are everywhere! :-)
--
Stewart Pinkerton | Music is art, audio is engineering
jj, curmudgeon and tiring philalethist
Hold ON a minute here. HF jitter, jitter that is well
above the range of 20kHz, introduces a gigantic amount of
signal-correlated NOISE, that can most certainly be very
VERY audible.
I've been watching the jitter audibility comments here for
a while, and it's clear to me that people aren't considering
the whole show here. Jitter can create some remarkably
unpleasant sounds, indeed, but it's not all due to LF
jitter.
--
Copyright j...@research.att.com 2001, all rights reserved, except transmission
by USENET and like facilities granted. This notice must be included. Any
use by a provider charging in any way for the IP represented in and by this
article and any inclusion in print or other media are specifically prohibited.
Arny Krueger
in message news:945bf...@news2.newsguy.com...
> In article <944eo...@news1.newsguy.com>,
> Stewart Pinkerton <pat...@popmail.dircon.co.uk> wrote:
> >LF jitter is more audible.
> Hold ON a minute here. HF jitter, jitter that is well
> above the range of 20kHz, introduces a gigantic amount of
> signal-correlated NOISE, that can most certainly be very
> VERY audible.
I mentally rattled through some Bessel functions and I can see how
that COULD happen.
In what context can such a thing be observed?
> I've been watching the jitter audibility comments here for
> a while, and it's clear to me that people aren't considering
> the whole show here. Jitter can create some remarkably
> unpleasant sounds, indeed, but it's not all due to LF
> jitter.
I've done technical measurements on over 100 different digital audio
products ranging from soup to nuts, and never knowingly observed such
a thing. In what context can it be observed?
MT...@webtv.net
I have found that my JISCO, jitter attenuator & decorrelator, is an
awesome answer to the problem of jitter reduction in a transport/ Dac
system. It plugs directly into the Dac, which means there is no
additional jitter added from anadditional cable, & shifts the jitter
to a higher frequency where the imput reciever is VERY effective at
attenuating the incoming jitter.
It works, I have & use one. You can argue theory until you are blue
in the face, but your ears will tell you very quickly what is real.
The music is relaxed, liquid, & more natural sounding ... to me. Are
there better systems, I'm sure there are & will be in the future. If
you want to read some technical info about it, go to www.jitter.de &
click on what is JISCO. I don't want to argue, I've tried it on a
number of expensive systems & it has always made a positive
difference. ... just my 2 cents worth.
Richard :)
Heijligers Marc J.M.
> 100% absolutely WRONG. The essential purpose of a CD
> transport is to read data and deliver it without DATA
> ERROR. The ONLY place where jitter is of ANY relevance
> is at the moment of conversion back to the analog
> domain. More specifically, the ONLY place where the
> jitter is relevant is in the DAC's sample and hold
> clock. Noweher else does jitter have ANY relevance.
From a conceptual point of view, this is absolutely true,
but unfortunalely it does not hold for a full 100% if you
consider (also state of the art) implemenations.
Implementations of convertor chips suffer from crosstalk
effects between data signals and clock signals (e.g. due to
ground bounce effects). This means that a perfect clock
signal can be disturbed by for instance a jittery data
stream at another terminal of your DA-convertor chip. You
cannot cope with this effect as a designer (even a chip
designer can only prevent this effect only to a certain
extend), except by preventing jittery signals around the
DA-convertor chip.
In other words, it is always a good idea to synchronize /
stabilize all signals around a DA-convertor chip (let's call
it reclocking). Depending on other practical reasons (your
supply topology, or your PCB layout) it might even be a good
idea to prevent jittery signals to appear in other places of
your design as well.
We've built, and tested such a situation, and were able to
measure (which is a science on its own) differences
w.r.t. jitter, in favour of the reclocking circuit (the
interpretation of jitter figures is not straighforward,
jitter is a statistical process, which can be characterized
in many many ways, so "10ps of jitter" doesn't tell you
anything without knowing the backgrounds of such a
measurement). We've also listened (non-blind, so some
reservations should be made) to a non-reclocked and
reclocked situation, and we observed (well...) that given
the same (very stable) conversion clock, reclocking
circuitry made audible differences between different
transports less obvious.
Marc
PS: My opinion, not my employer's
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
Stewart Pinkerton
>In article <944eo...@news1.newsguy.com>,
>Stewart Pinkerton <pat...@popmail.dircon.co.uk> wrote:
>>LF jitter is more audible.
>
>Hold ON a minute here. HF jitter, jitter that is well
>above the range of 20kHz, introduces a gigantic amount of
>signal-correlated NOISE, that can most certainly be very
>VERY audible.
Sorry, I should have been clearer. Jitter within the audible range
tends to be more audible when it comprises largely LF sidebands.
Jitter in the stop band can indeed have unpredictable effects,
depending on the amplifier in use. This is especially true of
'single-bit' designs with their much higher RF noise.
>I've been watching the jitter audibility comments here for
>a while, and it's clear to me that people aren't considering
>the whole show here. Jitter can create some remarkably
>unpleasant sounds, indeed, but it's not all due to LF
>jitter.
Agreed.
--
Stewart Pinkerton | Music is art, audio is engineering
Richard D Pierce
It, sorry, but it IS the frequency of the jitter. The audibility
of jitter is dependent, among other things, on the jitter
spectrum.
>it's that each step in the
>process of recreating music from a CD adds to it. The best a system
>can do is reduce it to the lowest level possible.
Why must this myth continue to be promulgated as some Universal
TRVTH (tm). It's just plain false. The ONLY point where jitter
is of ANY relevance is at the very moment of conversion from the
digital to the analog domains. Jitter ANYWHERE ELSE in the
system is completely and totally irrelevant. In fact, if you
look at the data coming of the CD, the jitter is IMMENSE as an
intrinsic part of the design of the system! (The data is
interleaved, it comes off the disk out of order and scrambled!).
>There are two main sources of jitter in most transports: internally
>generated and how well the output signal is timed.
There is only ONE source of jitter that is relevant: at the
output of the DAC. The jitter everywhere else has no relevance
on the sound. You can send the data to the DAC with jitter
measured in kiloseconds: as long as the DAC gets it right, it
makes no difference.
>Free of internal
>jitter would mean that it did a bang-up job of spinning the disc and
>got all the data off of it exactly on time.
And here's where you are, in fact, completely wrong. The data
CONNOT come off the disk "exactly on time" because the data WAS
NEVER PUT ON THE DISK in struct time order. The data is heavily
interleaved as an intrinsic part of the design of the system.
And, as such, this demonstrates well the complete irrelevancy of
the jitter argument when it comes to transports.
>A well-timed output would
>mean that it was still pretty much on time when it left the unit. From
>there, you can expect it to get worse. Even the time it takes to
>travel along the cable to the DAC introduces jitter.
And if the DAC is dependent upon such timing, the designer of
the DAC should be taken out and publically humiliated for such
an incredibly stupid design.
>Is it necessary to put that big heavy weight on it? I'm no expert, but
>it might make a difference in the quality of data recovered from the
>CD if you reduced vibrations as much as possible.
And if it does make a difference, it's only because of
extraordinary incomptetence on the part of the designer who did
the DAC.
>I suppose the custom-made clocks used to drive the servo that spins
>the disc and the electronics that handle the signal might find its way
>into a sub-$1,000 transport. But there are other economic realities...
ANd those are totally overshadowed by the one physical
reality: JITTER IS IRRELEVANT ANYWHERE BUT AT THE VERY POINT OF
CONVERSION FROM THE DIGITAL TO THE ANALOG DOMAIN.
>As to why nobody seems to publish just HOW good they deal with jitter:
>good question.
And the answer is simple: many of the designs in the high-end
industry are just plain stupid. The high-end audio industry is
decades behind understanding of the role of jitter and where it
is important. As long as the engineering is driven by the high
end cranks, marketeers and magazine wonks, this situation will
remain.
>There are other benefits to consider in a high-end CD system, though
>most of these aren't evident until we get the data to the DAC
Not "evident," rather "relevant."
--
| Dick Pierce |
| Professional Audio Development |
| 1-781/826-4953 Voice and FAX |
| DPi...@world.std.com |
jj, curmudgeon and tiring philalethist
Arny Krueger <ar...@pop3free.com> wrote:
>"jj, curmudgeon and tiring philalethist" <j...@research.att.com> wrote
>in message news:945bf...@news2.newsguy.com...
>> above the range of 20kHz, introduces a gigantic amount of
>> signal-correlated NOISE, that can most certainly be very
>> VERY audible.
>I mentally rattled through some Bessel functions and I can see how
>that COULD happen.
>In what context can such a thing be observed?
In DAC's used in the very first MPEG-Audio tests run at
Swedish Radio, for instance, on common source material.
>I've done technical measurements on over 100 different digital audio
>products ranging from soup to nuts, and never knowingly observed such
>a thing. In what context can it be observed?
Interesting. It's what I've observed more than any sort of simple
modulation artifacts. If you have a lot of HF jitter, AND you have
a signal with substantial HF content, the jitter maps to amplitude
error (well, that's hardly surprising) that is much wider band
than the signal. In an ADC, this aliases RIGHT into the
audio band, thud. In a DAC, what happens depends a lot on
how the DAC works, but I've seen anharmonic narrow band noise,
wideband noise modulated at the signal slew rate, and some other
icky stuff. When I say "seen" sorry, I should say "heard". Ecch.
Stewart Pinkerton
>For $20,000, you can get a box that can read and decode the data from
>a CD. The Linn Sondek CD12 is $20K and it has a built-in DAC. The Mark
>Levinson 31.5 Reference (at $17,000) is just a transport, though.
And inside it is just a Philips CD12 transport mechanism!
>As one reviewer put it (when talking about the Sondek CD12): "For
>12,000 pounds, you can either have a fairly crappy car, or one of the
>best CD players made."
>
>I would like to shake that man's hand.
I would like to remind that man that for 12 *hundred* pounds, you can
either have a decent bicycle or one of the best CD players made, the
Arcam CD-23......
>>The multiple power supplies with their
>>overzied bank of filter capacitors, the massive
>>chassis, proprietary circuits, temperature-controlled
>>crystal oscillator, damping material, special feet, of
>>your new transport, they all contribute to the effect.
Nope, only the power supplies and crystal oscillator, and even then,
only if the same oscillator is clocking the DAC!
>Most transports themselves are made by Pioneer and Philips, and the
>odds are good that even in a $1,000 transport, it's the same one used
>in the $249 model from Circus City (no, no a typo).
Correct, although Sony transports are also used (Arcam and Rega)
>One thing that gets me is that even in some expensive transports, they
>have cheap plastic drawers. The solid aluminum drawer found in
>Proceed/Mark Levinson transports is, by itself, a powerful, seductive
>force. (okay, I'm pushing it a little)
You certainly are, since this is just the loader which carries the
disc to the transport, it has no function while the disc is spinning.
>>And
>>I have been reliably informed that the *only* factor
>>which can account for any difference in sound
>>is...jitter.
>
>Not the only one. The signal is still subject to RF noise and other
>interference. An AES/BEU format signal will sound better (less noise)
>than an RCA coaxial or Toslink optical output (probably the worst
>connection available).
Unless of course you have an AT&T optical connection, which is in
theory even better, with 50MHz bandwidth and zero susceptibility to
noise. Also, an AES/EBU connection will only be better *if* there is
noise around to be picked up by your system.
>A fully balanced (in the XLR sense) system is more than just the
>transport, of course, but my favorite test to show my tin-eared
>neighbors the folly of their Circus City paperweight is to pause the
>source and turn the volume way up.
>
>"I don't hear anything" they say.
>
>EXACTLY right. No snap crackle pop when it should be quiet. It only
>takes one component to inject noise to ruin the magic spell.
Yup, and this also works extremely well with my one-box player which
all runs off one low-noise free-running master clock.
>There are two main sources of jitter in most transports: internally
>generated and how well the output signal is timed. Free of internal
>jitter would mean that it did a bang-up job of spinning the disc and
>got all the data off of it exactly on time.
Actually, this doesn't matter at all, since the data are totally
reclocked by the FIFO RAM buffer.
>A well-timed output would
>mean that it was still pretty much on time when it left the unit. From
>there, you can expect it to get worse. Even the time it takes to
>travel along the cable to the DAC introduces jitter.
Only to a totally negligible amount. The problem lies at the receiver
end, since the clock is now embedded in the data stream and has to be
reconstructed, a fudamentally inferior and internally broken process.
>The quality of power in ANY system that uses it can make a profound
>difference. In this case, it might be more important that the power
>source is also shielded enough to keep from interfering with the
>clocking circuit(s).
Very true, the purity of the master clock is critical to accurate D/A
conversion. So why put the darn thing in the transport, when where it
*should* be is right next to the DAC chips?
>Is it necessary to put that big heavy weight on it? I'm no expert, but
>it might make a difference in the quality of data recovered from the
>CD if you reduced vibrations as much as possible.
>I suppose the custom-made clocks used to drive the servo that spins
>the disc and the electronics that handle the signal might find its way
>into a sub-$1,000 transport.
As noted above, irrelevant to jitter at the point of conversion, which
is the *only* place where it matters.
> But there are other economic realities...
>
>Sony (if memory serves) led the way to introducing cheaper components
>into consumer gear. Part of it is always progress: better quality in
>the future can always be had cheaper than it was before. Remember when
>Sony introduced 1-bit DACs and said it was the best way? It isn't, but
>it's good quality for the money: a high-quality multibit DAC can be
>expensive. High-precision clocks can be expensive. When you're
>building a player that a dealer can mark up to $249, you have to pinch
>pennies.
OTOH, the mid-price Arcam CD-23 has excellent power supplies, very low
jitter, and arguably the best CD DAC available at *any* price.
>There are other benefits to consider in a high-end CD system, though
>most of these aren't evident until we get the data to the DAC where we
>can decode HDCD and other formats. I am reluctant to mention anything
>that does this onthe same unit, because, like your tin-eared
>neighbor's Sony, it's not a transport: it's a CD player/processor.
Yes, and it will probably have significantly less jitter than any
two-box solution!
There's only one DAC which properly reclocks the incoming data stream
to break the jitter link, and that's the Meridian 861. It has *much*
lower jitter than most of the high-end combos, although still no lower
than good one-box players like the Arcams, the Krell and the Linn
CD12.
--
Stewart Pinkerton | Music is art, audio is engineering
Nicolas Hodges
>In article <942ii5$4ch$1...@bourbaki.localdomain>,
>lich <ob...@earthlink.net> wrote:
>>The essential purpose of a CD transport is to read data
>>and deliver it without jitter.
>
>100% absolutely WRONG. The essential purpose of a CD transport
>is to read data and deliver it without DATA ERROR.
This is what I was trying to get at in my posting about transport
accuracy. How much of an issue is data error and the resulting
interpolation (which we heard a lot about at the start of CD, I seem
to remember)? By how much, I mean what percentage of data is lost and
interpolated on a $250 player compared to a $1000 one, or compared to
a $5000 transport. And what measurable effect does this have.
--
Nic
jj, curmudgeon and tiring philalethist
Richard D Pierce <world!DPi...@uunet.uu.net> wrote:
>In article <943a7f$cd7$1...@bourbaki.localdomain>,
>Justin Sullivan <jus...@netapp.com> wrote:
>>it's that each step in the
>>process of recreating music from a CD adds to it. The best a system
>>can do is reduce it to the lowest level possible.
>Why must this myth continue to be promulgated as some Universal
>TRVTH (tm). It's just plain false.
Dick, I think it's because people don't understand the whole
issue of digital signals, and are trying to apply voltage-analog
and current-analog rules to sampled, quantized analogs, which
is to say digiatl signals.
Justin, you're dead wrong here. A digital signal has one very
redeeming thing that what is usually called "analog" doesn't have,
which is the ability to be exactly recovered, again and again. This
is due to the quantization as well as the encoding into a plurality
of large, binary values. This means that there is NO, repeat NO,
zero, ZIP, NADA accumluation in each step of the CD player's
processing in terms of jitter or noise (the CD player does not do
things that would lead to additive noise, either, from quanitization
as long as it's well designed, leaving some older designs, at the
very least, out there).
The ONLY jitter in the final signal is the jitter in the DAC clock.
The statements you make afterwards, as Dick has pointed out,
also appear to assume incorrect methods for CD storage and
recovery.
Steven Sullivan
transports here recently. They prompt me to ask the following.
My current CD player -- a years-old Pioneer F109 100 disc changer, is
fed into a Musical Fidelity X-DAC (purchased so I could have HDCD
decoding). Now the 'transport' is acting funny after years of heavy
use and it may be time to retire it (unless cleaning the lens
restores it to full vigor this weekend). I have no idea what the
jitter characteristics are of this system, but I have read from the
usual reliable sources here (Stewart, jj, Dick Pierce, etc) that an
all-in-one box would tend to have less DAC jitter than separates --
something I've also read on other forums. A such, I'd be interested
to know if any jukebox-type players (50 or more disc capacity)
feature a good DAC *and* HDCD encoding.
--
-S.
Richard D Pierce
Nicolas Hodges <n...@nicolashodges.demon.co.uk> wrote:
>Richard D Pierce <world!DPi...@uunet.uu.net> writes
>>In article <942ii5$4ch$1...@bourbaki.localdomain>,
>>lich <ob...@earthlink.net> wrote:
>>>The essential purpose of a CD transport is to read data
>>>and deliver it without jitter.
>>
>>100% absolutely WRONG. The essential purpose of a CD transport
>>is to read data and deliver it without DATA ERROR.
>
>This is what I was trying to get at in my posting about transport
>accuracy. How much of an issue is data error and the resulting
>interpolation (which we heard a lot about at the start of CD, I seem
>to remember)?
The entire issue of data recovery and interpolation is handle
LONG before the data EVER finds its way out of the transport
electronics.
The data is interleaved when it is put on the disk. If you tried
to play back the raw data, it would not make any audible
sense. Immediately after the data comes out of the pickup, it
MUST be buffered so it can be deinterleaved. And an integral
part of the process is the error correction. Thus, well before
the data EVER reaches the S/P-DIF port on the back of a
transport, it has already gone through decoding, RAM buffering,
deinterleaving, error detection and correction and, if the
correction failed, interpolation. The DAC has NO role in any of
this.
>By how much, I mean what percentage of data is lost and
>interpolated on a $250 player compared to a $1000 one, or compared to
>a $5000 transport.
The data suggests that while there may be dozens of raw errors
per minute in the data coming off the disk, the number of
uncorrectable errors amounts to maybe a dozen or less PER
DISK. And every ony of the corrected errors were corrected
EXACTLY. Not interpolated, not guessed at, but corrected to the
EXACT value they should be. It is these last small handfull of
errors that go through interpolation, and that is always done in
the transport electronics themselves.
The exact numbers depend upon the condition of the disk itself
(is it REALLY badly damaged?) and of the transport (cat lint on
the lens?). The entire point is that the process of reading,
decoding and correcting the data is NOT dependent upon the
transport.
>And what measurable effect does this have.
The measurable effect is that the measured differences between
transports is much less than people claim or think it should be.
Dr. Bob
People have been exposed to the idea of digital information for
decades. I wonder, how old *are* some of these folks?
Dr. Bob
---------------------------------------------
In article <947jr...@news1.newsguy.com>, j...@research.att.com (jj,
curmudgeon and tiring philalethist) wrote:
[...]
>
>Dick, I think it's because people don't understand the whole
>issue of digital signals, and are trying to apply voltage-analog
>and current-analog rules to sampled, quantized analogs, which
>is to say digiatl signals.
>
lich
>
>> [Jitter is] Not the only one. The signal is still subject to RF noise
>> and other
>> interference. An AES/BEU format signal will sound better (less noise)
>> than an RCA coaxial
>
Second, it's a generalization for which you will find as many in agreement
as in disagreement (a coax connection is not inherently noisy). Third,
if any "noise" introduced in the digital cable does not directly effect
jitter, or the functioning of circuits inside the DAC, how can it have any
audible effect? I am, to echo your comment, no expert either, but I have
been repeatedly told by experts: jitter is it.
>
>> [snip]
>> Several thousand
>> picoseconds is probably not noticeable.
>
You have hit the bull point: *if* several thousand picoseconds of jitter
is
not noticable, and if jitter is the sole factor which could account for
the differences in sound between transports, then the Levinson, the Krell,
the Linn, the CEC, must all have jitter levels in the many thousands of
pico seconds, and at different frequencies, if they sound different. On
the face of it it does not seem likely these units would have jitter
anywhere near that range.
>
>> [snip]
>>
>> The quality of power in ANY system that uses it can make a profound
>> difference. In this case, it might be more important that the power
>> source is also shielded enough to keep from interfering with the
>> clocking circuit(s).
>>
>> Is it necessary to put that big heavy weight on it? I'm no expert, but
>> it might make a difference in the quality of data
>
What, pray, is "quality of data"? There are only two "qualities" of
concern: do the 1s and 0s of the sample match the 1s and 0s on the CD, and
how accurately are they timed?
>
>> [snip]
>>
>> There are other benefits to consider in a high-end CD system, though
>> most of these aren't evident until we get the data to the DAC
>
Huh? The same data bits at the same level of jitter *will sound exactly
the same* whatever the source, no? If, hypothetically but not
unrealistically,
an Arcam 7 were blind tested against a Linn and you couldn't tell the
difference, how would you have 'benefitted' from spending the additional
$19,000+?
ni...@media.mit.edu
I really appreciate your commitment to taking the junk "science" out of
high fidelity audio.
I would love to hear recommendations from you guys about specific low-
cost high-quality high end products. So far, I think I have only
really heard concrete suggestions from Stewart (Excuse me if I am wrong
here).
I know you guys "complain" a lot about the junk in high end audio and I
can't help but agree but I would love to see some constructive
suggestions on what we can buy to avoid getting ripped off.
Thanks very much!
nivi
Sent via Deja.com
http://www.deja.com/
Stewart Pinkerton
Given that they all use the same basic transport mechanisms and
error-correction circuitry, the answer should be obvious!
In absolute terms, average figures are one interpolated 'best guess'
error every twenty million samples (i.e. one sub-millisecond event
every five minutes or so), and less than one 'muting' error per disc.
In other words, data recovery is essentially perfect in *any* modern
CD player or transport at *any* price - as you'd expect when you
consider that cheap plastic CD-ROM drives can pull bit-perfect files
into your computer at 40x normal speed!
--
Stewart Pinkerton | Music is art, audio is engineering
Heijligers Marc J.M.
> The measurable effect is that the measured differences
> between transports is much less than people claim or think
> it should be.
I'm not sure what people claim or think, so I can't judge
about this proposition, but my own experiences show that
measured differences between transports are quite drastic.
Together with some friends we've build a PLL, that locks on
the clock signal of a SPDIF input receiver. This PLL
contains a VCXO, which shows very good specs. Measurements
on a spectrum analyser (140dB deep) or time domain analyser
show that this VCXO is very stable. The loop bandwidth of
this PLL is very very low (about 0.1Hz). This means that the
phase comparator of the PLL gets a very stable input clock
from the VCXO. The other clock signal is directly
constructed from the SPDIF, and measurements show quite some
jitter, and transport sensibility.
If you don't have expensive equipment in your reach, you can
try out the following funny idea. Just 'listen' to the
output of the phase comparator. This gives funny results,
some of which are mentioned below:
- The output signal consists of audible noise.
- Different transports give VERY different types of audible
noise (e.g sssss as in 'she', or ffffffff as in 'the').
- With some transports, only noise can be heard, with other
transports you can hear (very clearly) audio data
'inside' the noise.
- One transport had the opportunity to switch on/off the
display. Switching on the display resulted in an audible
buzz (electricaly, transport and DAC/PLL are decoupled).
Opening the drawer of a player gives quite some
interesting sounds as well. This probably tells a lot
about the supply rejection or EMC-related aspects in such
a (so-called high-end!!!!) transport.
- With the same transport, but with different brands of
CD-Rs with the same music, differences can be heard quite
easily.
- Measurements on this output with a spectrum analyzer also
give quite different spectra.
In all these conditions, measurements show that the VCXO
signal remains stable. So all these audible effects are
mainly due to the reconstructed clock from the SPDIF, and
are effects originating from the transport.
Whether these effects are 'relevant', is another issue.
Marc
PS: My opinion, not my employer's
Stewart Pinkerton
>world!DPi...@uunet.uu.net (Richard D Pierce) writes:
>
>> The measurable effect is that the measured differences
>> between transports is much less than people claim or think
>> it should be.
>
>I'm not sure what people claim or think, so I can't judge
>about this proposition, but my own experiences show that
>measured differences between transports are quite drastic.
>
>Together with some friends we've build a PLL, that locks on
>the clock signal of a SPDIF input receiver. This PLL
>contains a VCXO, which shows very good specs. Measurements
>on a spectrum analyser (140dB deep) or time domain analyser
>show that this VCXO is very stable. The loop bandwidth of
>this PLL is very very low (about 0.1Hz). This means that the
>phase comparator of the PLL gets a very stable input clock
>from the VCXO. The other clock signal is directly
>constructed from the SPDIF, and measurements show quite some
>jitter, and transport sensibility.
Some numbers for this jitter would be nice. Name names!
Andre Yew
: I would love to hear recommendations from you guys about specific low-
: cost high-quality high end products. So far, I think I have only
: really heard concrete suggestions from Stewart (Excuse me if I am wrong
: here).
I think I can predict what Dick and JJ are going to say already
(should this be an FAQ?). But if you want to know what JJ uses
in his research, you could just read his papers. For example,
in AES preprint 5202, Perceptual Soundfield Reconstruction,
he uses Apogee 20-bit DACs fed from a computer (I assume), a
passive 10-channel volume control, Halfer P7000 amps, and Snell
C/V speakers. Knowing that, I still don't see how you can assemble
a system without listening to it in your own home, and deciding
for yourself if it sounds good to you.
--Andre
marcos...@yahoo.it
> >Not the only one. The signal is still subject to RF noise and other
> >interference. An AES/BEU format signal will sound better (less noise)
> >than an RCA coaxial or Toslink optical output (probably the worst
> >connection available).
>
> Unless of course you have an AT&T optical connection, which is in
> theory even better, with 50MHz bandwidth and zero susceptibility to
> noise. Also, an AES/EBU connection will only be better *if* there is
> noise around to be picked up by your system.
>
Why do not use interfaces like Ethernet 10BASE-T or USB with proper FIFO
buffering? They deliver enough bandwidth and can make the connection
between transport and DAC virtually perfect and should not bee too
expansive
Regards
Stewart Pinkerton
Why? Because there's no need for wide bandwidth in the link, which has
a 2Mbit/sec transfer rate, and because there is only one DAC on the
market (Meridian) which fully reclocks the incoming datastream via a
FIFO buffer.
Stewart Pinkerton
>I would love to hear recommendations from you guys about specific low-
>cost high-quality high end products. So far, I think I have only
>really heard concrete suggestions from Stewart (Excuse me if I am wrong
>here).
Yup, a concrete slab floor and walls are a good start to getting clean
bass with good 'slam'! :-)
>I know you guys "complain" a lot about the junk in high end audio and I
>can't help but agree but I would love to see some constructive
>suggestions on what we can buy to avoid getting ripped off.
Excuse me while I don my nomex racing suit.......
Here be dragons, and speakers are *very* personal, but a few
suggestions at various price points:
Rotel RCD-971
Rotel RA-971
KEF Q35.2
The cables that came in the box, with freebie speaker cable from the
store. Ah yes, always buy from a dealer, otherwise in a few years
you'll have nowhere to hear new gear!
Rotel RCD-971
Roksan Caspian
B&W 604S2
Rat Shack 'Gold' cables, and 12-14AWG zipcord
Arcam CD-23
Roksan Caspian
JMlab Mini Utopia
REL Stentor or Velo HGS-18 sub
Rat Shack 'Gold' cables, and 12-14AWG zipcord
Arcam CD-23
Musical Fidelity A3CR preamp
Krell KAV-250a
B&W N801
Rat Shack 'Gold' cables, and 12-14AWG zipcord
Arcam CD-23
Musical Fidelity A3CR preamp
Krell FPB-300
JMlab Utopia
Rat Shack 'Gold' cables, and 12-14AWG zipcord
Or, for something entirely different, of reasonable cost but decidedly
high end and needing a *big* room:
Arcam CD-23
Unison Research Simply 845
Avantgarde Uno
Rat Shack 'Gold' cables, and 12-14AWG zipcord
Happy listening!
--
Stewart Pinkerton | Music is art, audio is engineering
Terry Demol
> Why? Because there's no need for wide bandwidth in the link, which has
> a 2Mbit/sec transfer rate, and because there is only one DAC on the
> market (Meridian) which fully reclocks the incoming datastream via a
> FIFO buffer.
>
You missed a few. dB Technologies, Weiss, Levinson 30.6.
There are others but can't think of them.
Regards,
Terry
Justin Sullivan
Pinkerton) wrote:
>ni...@media.mit.edu writes:
>
>>I would love to hear recommendations from you guys about specific low-
>>cost high-quality high end products. So far, I think I have only
>>really heard concrete suggestions from Stewart (Excuse me if I am wrong
>>here).
>
>Yup, a concrete slab floor and walls are a good start to getting clean
>bass with good 'slam'! :-)
>
>>I know you guys "complain" a lot about the junk in high end audio and I
>>can't help but agree but I would love to see some constructive
>>suggestions on what we can buy to avoid getting ripped off.
>
>Excuse me while I don my nomex racing suit.......
>
>Here be dragons, and speakers are *very* personal, but a few
>suggestions at various price points:
Well, I show my ignorance a time or two, but it is very educational
(fire breathing aside).
I've said it before, but it bears repeating. For the money, Rotel
simply cannot (in my opinion) be beat. The RCD-971 is about $700.
Arcam makes excellent gear as well, for a little more (right around
$1,000). Arcam comes up here quite a bit - not in the "affordable"
sense, but in the "excellent performance" sense. It just so happens to
be both.
For a lot of enthusiasts, the performance of Rotel and Arcam is more
than satisfactory, even before looking at the price tag. I have not
heard the CD-23 (is it still even available?), but the Alpha 7 starts
around the same price as the Rotel. Unless I'm mistaken, the Alpha 9
is under $2,000. I believe it is the only current Arcam player that
includes HDCD decoding.
The Linn Classik is an excellent completely integrated piece, but it
lacks HDCD decoding and it's only 75w/channel, but hey - it's only
$2,000.
My favorites are Linn, Proceed and Mark Levinson, brands that are
probably not well-liked here in tubeland, but Rotel gear may make you
wonder why you need gear that costs ten times as much - it doesn't
sound ten times better. I find that the Mark Levinson gear has my kind
of sound (transparent), it looks nice and it is a sheer pleasure to
use, so I'm sure I'll be the envy of all at the alms house.
Just my 0.02.
- Justin
Stewart Pinkerton
>"Stewart Pinkerton" <pat...@popmail.dircon.co.uk> wrote in message
>
>> Why? Because there's no need for wide bandwidth in the link, which has
>> a 2Mbit/sec transfer rate, and because there is only one DAC on the
>> market (Meridian) which fully reclocks the incoming datastream via a
>> FIFO buffer.
>>
>You missed a few. dB Technologies, Weiss, Levinson 30.6.
>There are others but can't think of them.
The Meridian genuinely buffers and totally reclocks the data from a
large RAM. This is *not* true of the Levinson, despite what they'd
like you to think. I'm not familiar with the Weiss or dB Tech models.
Lots of companies make vague reference to jitter reduction and
re-clocking (Linn even make the ludicrous claim that the CD12 has *no*
clock!), but AFAIK the Meridian 861 is the only model which is a
*true* reclocker, using a local free-running clock with no reliance
whatever on the embedded clock in the incoming datastream.
Stewart Pinkerton
>For a lot of enthusiasts, the performance of Rotel and Arcam is more
>than satisfactory, even before looking at the price tag. I have not
>heard the CD-23 (is it still even available?), but the Alpha 7 starts
>around the same price as the Rotel. Unless I'm mistaken, the Alpha 9
>is under $2,000. I believe it is the only current Arcam player that
>includes HDCD decoding.
FYI, the CD-23 is the 'FMJ' upgrade of the Alpha 9, is the current top
of the Arcam range at $1800, and does include HDCD. IMNVHO, it
represents the state of the art in CD replay, regardless of price.
vince c
state of the art. Top offerings from Linn and Naim outdo it, but, of
course, at a price.
Vince Cannistraro
William A. Dirks
wrote:
>The CD-23 is good value for the money, but it hardly represents the
>state of the art. Top offerings from Linn and Naim outdo it, but, of
>course, at a price.
>
>Vince Cannistraro
For anyone who has significant listening experience with both the
CD-23 and well-reviewed models at a slightly higher price range--I'm
thinking especially of the Meridian 508-24, since that's what I
have--I'd be curious to hear about your comparisons.
--Bill Dirks
Stewart Pinkerton
>The CD-23 is good value for the money, but it hardly represents the
>state of the art. Top offerings from Linn and Naim outdo it, but, of
>course, at a price.
An interesting claim, presumably based only on cost and 'high end'
reputation. The plain fact is that the Naim CDSII, like all Naim
products, is *at least* five years out of date, while the Linn is
certainly a very fine implementation of conventional technology, with
a Philips reading head and 20-bit Burr-Brown converters. Elegant and
superbly built? Certainly, but hardly 'state of the art' against the
genuine 24-bit linear dCS RingDAC in the Arcam, or even the excellent
and elegant Meridian 508.24.
Note that the above has nothing to do with putative differences in
sound quality, just with which player truly represents the 'state of
the art' in CD replay.
Stewart Pinkerton
>For anyone who has significant listening experience with both the
>CD-23 and well-reviewed models at a slightly higher price range--I'm
>thinking especially of the Meridian 508-24, since that's what I
>have--I'd be curious to hear about your comparisons.
As it happens, I have directly compared these two, and I'd say there's
little or no detectable difference in a level-matched blind comparison
(certainly nothing I or two friends could hear via my Krell amp and
Apogee Duetta Signature speakers). The higher RF noise of the Meridian
*might* be an issue with some sensitive amplifiers, but this is
unlikely. One of my co-listeners in the test subsequently bought the
Arcam, the other bought the Meridian...........
vince c
reproduction of the recorded music on the disk. Good engineering with
conventional devices often surpasses mediocre engineering with the latest
chips. I think both the latest NAIM (not the five year old model cited) and
the CD 20 have better sound quality. Again, at a price.
Vince Cannistraro
"Stewart Pinkerton" <pat...@popmail.dircon.co.uk> wrote in message
news:958ifq$ul8$1...@bourbaki.localdomain...
Stewart Pinkerton
>State of the art refers to the engineering responsible for the most accurate
>reproduction of the recorded music on the disk.
Quite so, and the Arcam CD-23 uses a 24-bit linear DAC from dCS, while
the Linn and Naim players share the same Burr-Brown 20-bit DAC.
Whether all this is *necessary* with a 16-bit medium is perhaps
another question... :-)
> Good engineering with
>conventional devices often surpasses mediocre engineering with the latest
>chips. I think both the latest NAIM (not the five year old model cited) and
>the CD 20 have better sound quality. Again, at a price.
And your evidence to back your opinion of their sound quality is, what
exactly?
The Naim CDSII I mentioned was launched in 1999, and is their flagship
model. The technology inside is the ubiquitous Philips CDM12 transport
that's been around for a decade, and the same Burr-Brown 20-bit DACs
used by Linn in their CD12. Note that Kenwood were offering *eight* of
the same PCM1702 DACs in their excellent $500 DP-7090 CD player five
years ago!
You mention the 'latest' Naim. That would be the CD5, using the usual
Philips transport and 18-bit Philips DACs. State of the art? Maybe
*ten* years ago...............
Yes, as I said before, the Naim and Linn flagship players are good
implementations (in the case of the Linn, *very* good) of conventional
technology from about five years ago (which is a very long time in
digital audio!), but 'state of the art'? Hardly, except of course in
their price tags............
Justin Sullivan
wrote:
>The CD-23 is good value for the money, but it hardly represents the
>state of the art. Top offerings from Linn and Naim outdo it, but, of
>course, at a price.
>
>Vince Cannistraro
Well, especially when it comes to asking for opinions and
recommendations on particular equipment and components, "value" can
be crucial. Of course, value has many interpretations: sound/build
quality for the dollar or pounds sterling, features ("bang for the
buck"), upgradability to ward off the inevitable (obsolence), etc.
There are just so many ways that "value" can, and in my opinion,
SHOULD guide a buyer to what is appropriate for their needs.
Linn's better gear is, in my opinion, very, VERY hard to beat. I will
never forget the first time I heard a a small Linn system with the
now-discontinued (regrettably) Tukan speakers, intended for a small
room.. All I could think was: "So THAT'S what a high-end system
sounds like!" I was bitten by the bug that causes audiophilitis as a
child, but even a modest (by many standards) Linn system redefined in
my mind what a system could sound like.
It still upsets me to hear something better than what I have at a
price I can afford.
It might be better to put the question thusly, and consider Lord
Pinkerton's example, the CD-23: is the Linn Ikemi, at $4,000, more
than twice as good as the CD-23 at $1,800?
My own dealer shares Pinkerton's opinion that the Arcam is one of the
best players made, and without question the finest value to be had in
a source component. There is better gear to be had, for sure, but
nowhere near this price range. What that might be, exactly, depends
on what you want in a source component.
I am saving for a Proceed PMDT myself. Like I've said before:
multichannel is coming - you can't stop it. I believe that the
convergence of very high-quaility traditional (stereo) systems and
home theater is inevitable, maybe just because it's more practical
that way.
- Justin
vince c
although it could use better damping of the chassis. But the
question is what is "state of the art", defined as the most accurate
sound reproduction. It is not the Arcam. Is the Ikemi better? I'm
not sure, but its clearly not "twice as good" in your phrase. The CD
12 is audibly better. Of course some will argue that's not so from
DBT. But lets not rake up that hoary debate once again.
Vince Cannistraro
"Justin Sullivan" <jus...@netapp.com> wrote in message
news:959sv...@news2.newsguy.com...
mcn...@my-deja.com
"vince c" <vinc...@bellatlantic.net> wrote:
> Well we certainly all agree that the CD-23 is excellent value,
> although it could use better damping of the chassis. But the
> question is what is "state of the art", defined as the most accurate
> sound reproduction. It is not the Arcam. Is the Ikemi better? I'm
> not sure, but its clearly not "twice as good" in your phrase. The CD
> 12 is audibly better. Of course some will argue that's not so from
> DBT. But lets not rake up that hoary debate once again.
>
offering something new here. Just how is it that you propose to
determine that some CD player out there is more "sonically accurate"
than the Arcam? Besides appointing yourself arbiter, of course.
bob
Stewart Pinkerton
>Well we certainly all agree that the CD-23 is excellent value,
>although it could use better damping of the chassis. But the
>question is what is "state of the art", defined as the most accurate
>sound reproduction. It is not the Arcam.
Oh, really? Evidence, please?
> Is the Ikemi better? I'm
>not sure, but its clearly not "twice as good" in your phrase. The CD
>12 is audibly better. Of course some will argue that's not so from
>DBT. But lets not rake up that hoary debate once again.
Oh indeed, let's just state that 'the CD 12 is audibly better' and not
even *think* about what might happen if you were to compare it to the
Arcam under level-matched double-blind conditions. Perish the thought
that we would actually want to try anything so dangerous as *trusting
our ears* when it comes to deciding what sounds best. After all, this
*is* the 'high end'................
Steve Emerson
The Alpha 8SE also includes HDCD.
SE.
dcs
Stuart,
My brother has a Meridian 861 and a while ago we were playing around
with the setup of his home theatre system, different modes of
bi-amping his KEF Reference 4's etc.
Out of interest we tried using my Marantz CD-10 as a transport
instead of his Meridian (200 series, I think it's a 206).
The result was the Marantz sounded considerably better, more natural
and transparent.
The Marantz had a Trichord Clock 3 and some power supply upgrades
fitted, which may have helped make the jitter on the digital output
lower than that of his Meridian transport (can't say for sure as we
had no way of measuring it).
If I am right in understanding the previous information, then this
means there must be a difference between the digital data stream
output from the two transports that is down to more than just jitter.
Since as you rightly say the Meridian is re-clocking the data itself
and therefore doesn't rely on the embedded clock in the incoming
datastream.
Jon
dcs
Stuart,
My brother has a Meridian 861 and a while ago we were playing around with
the setup of his home theatre system, different modes of bi-amping his KEF
Reference 4's etc.
Out of interest we tried using my Marantz CD-10 as a transport instead of
his Meridian (200 series, think it's a 206?).
The result was the Marantz sounded considerably better, more natural and
transparent.
The Marantz had a Trichord Clock 3 and some power supply upgrades fitted,
which may have helped make the jitter on the digital output lower than that
of Meridian transport (can't say for sure as we had no way of measuring it).
However as the Meridian is re-clocking the data itself, and as you say it
"uses a local free-running clock with no reliance whatever on the embedded
clock in the incoming datastream" then there must be a difference between
the output digital data stream from the two transports that is more than
just the level of jitter?
Jon
vince c
about comparing the Arcam under DBT conditions or are you arguing that one
should trust their ears? Or do you want it both ways? Please clarify.
BTW, you may improve your CD-23 by replacing the power supply caps with
Black Gates. Try it, you may like it- unless of course you are convinced
that your CD-23 is already "state of the art." In that case, this
controversy can never be reasonably resolved.
Vince Cannistraro
"Stewart Pinkerton" <pat...@popmail.dircon.co.uk> wrote in message
news:95emn...@news1.newsguy.com...
vince c
listen carefully to the Arcam and the CD 12, playing on both units a piece
of recorded acoustic music you have listened to live in the concert hall,
and then you'll make up your mind. Of course you may not trust your ears,
in which case you will do a level-matched double blind test under carefully
controlled environmental conditions, and then decide. Whichever method you
choose, and whichever conclusion you reach, will be fine. You see, you are
the arbiter.
Regards,
Vince Cannistraro
<mcn...@my-deja.com> wrote in message
news:95dk5u$s54$1...@bourbaki.localdomain...
Stewart Pinkerton
>My brother has a Meridian 861 and a while ago we were playing around
>with the setup of his home theatre system, different modes of
>bi-amping his KEF Reference 4's etc.
>Out of interest we tried using my Marantz CD-10 as a transport
>instead of his Meridian (200 series, I think it's a 206).
>The result was the Marantz sounded considerably better, more natural
>and transparent.
>If I am right in understanding the previous information, then this
>means there must be a difference between the digital data stream
>output from the two transports that is down to more than just jitter.
There is of course another possibility - the difference does not exist
in the physical world. I recommend further comparisons under
double-blind conditions.
Stewart Pinkerton
>There seems to be some inconsistency in your response. Are you arguing
>about comparing the Arcam under DBT conditions or are you arguing that one
>should trust their ears? Or do you want it both ways? Please clarify.
I thought it *was* clear! The essence of a DBT is that you *must*
'trust your ears', because you don't *know* which player is connected.
I remain fascinated by those 'subjectivists' who insist that we must
'trust our ears', but who only ever indulge in sighted tests, where
they are in reality trusting their eyes.............
>BTW, you may improve your CD-23 by replacing the power supply caps with
>Black Gates. Try it, you may like it- unless of course you are convinced
>that your CD-23 is already "state of the art." In that case, this
>controversy can never be reasonably resolved.
That's a non-sequitor, as anything can be improved - or degraded.
krom...@my-deja.com
Stewart, do you have any more technical information about how the
Meridian does this? For example, what I'm thinking about is how the
system deals with the fact that two fully independant clocks will never
be running at 100% the same speed, one will always be a little faster
than the other.
Does the Meridian toss away or replicate samples if the FIFO becomes
full or empty? (Or perhaps it tries to interpolate samples?)
If it uses some kind of Negative Feedback (ie. a PLL) to try to sync
the clocks, then I imagine some percentage of jitter effect would be
propagated...
Andy K.
>
> --
>
> Stewart Pinkerton | Music is art, audio is engineering
>
>
Sent via Deja.com
http://www.deja.com/
Stewart Pinkerton
>> The Meridian genuinely buffers and totally reclocks the data from a
>> large RAM. This is *not* true of the Levinson, despite what they'd
>> like you to think. I'm not familiar with the Weiss or dB Tech models.
>> Lots of companies make vague reference to jitter reduction and
>> re-clocking (Linn even make the ludicrous claim that the CD12 has *no*
>> clock!), but AFAIK the Meridian 861 is the only model which is a
>> *true* reclocker, using a local free-running clock with no reliance
>> whatever on the embedded clock in the incoming datastream.
>
>Stewart, do you have any more technical information about how the
>Meridian does this? For example, what I'm thinking about is how the
>system deals with the fact that two fully independant clocks will never
>be running at 100% the same speed, one will always be a little faster
>than the other.
They use a FIFO RAM buffer. I don't know the exact details
>Does the Meridian toss away or replicate samples if the FIFO becomes
>full or empty? (Or perhaps it tries to interpolate samples?)
Use a decent size buffer, you have no problem. A couple of seconds
worth of buffering is adequate for any input within class 1 limits for
clock accuracy.
>If it uses some kind of Negative Feedback (ie. a PLL) to try to sync
>the clocks, then I imagine some percentage of jitter effect would be
>propagated...
It doesn't, although it does of course use different clocks for
different input sampling rates. As an aside, I've just been informed
of another true reclocker, the Bel Canto. This uses the alternative
system of also using a free-running clock, but employing an
asynchronous sample rate converter to avoid worries about buffering.
What effect this has on the sound, I don't know.
mcn...@my-deja.com
"vince c" <vinc...@bellatlantic.net> wrote:
> Let us resolve this simply, Bob. You can be the arbiter. I assume you'll
> listen carefully to the Arcam and the CD 12, playing on both units a piece
> of recorded acoustic music you have listened to live in the concert hall,
In other words, your definition of "sonic accuracy" is that it
reproduces a recording of a piece played in one hall by a particular
orchestra in such a way that it reminds you of the same piece you
once heard being played by another orchestra in a different hall. Do
I have that right?
> and then you'll make up your mind. Of course you may not trust your ears,
Actually, Vince, it's your eyes I do not trust.
> in which case you will do a level-matched double blind test under carefully
> controlled environmental conditions, and then decide. Whichever method you
> choose, and whichever conclusion you reach, will be fine. You see, you are
> the arbiter.
And when I do, I will free to make broad, definitive statements about
their relative sonic merits. But not until.
vince c
> "vince c" <vinc...@bellatlantic.net> writes:
>
> >There seems to be some inconsistency in your response. Are you arguing
> >about comparing the Arcam under DBT conditions or are you arguing that
one
> >should trust their ears? Or do you want it both ways? Please clarify.
>
> I thought it *was* clear! The essence of a DBT is that you *must*
> 'trust your ears', because you don't *know* which player is connected.
> I remain fascinated by those 'subjectivists' who insist that we must
> 'trust our ears', but who only ever indulge in sighted tests, where
> they are in reality trusting their eyes.............
Then we have no quarrel on this point. But the conditions have to be set up
correctly. I can recall when AR did an infamous test as part of an
advertising campaign. They had a tape recorder behind one curtain playing
music through an AR speaker and a live musician behind another, playing the
same piece of music. The audience was then invited to guess which was live.
And, you might suspect, many in the audience guessed the tape recorder. Of
course, you and I would both argue that the environment was not correctly
set up.
> >BTW, you may improve your CD-23 by replacing the power supply caps with
> >Black Gates. Try it, you may like it- unless of course you are convinced
> >that your CD-23 is already "state of the art." In that case, this
> >controversy can never be reasonably resolved.
>
> That's a non-sequitor, as anything can be improved - or degraded.
No, its not a non-sequitor. You argued that the Arcam was state of the art,
and cited at least on one post the fact that it had a DAC more up to date
than the CD-12. My argument is that excellent engineering with less than
state of the art technology is often better than mediocre engineering with
the latest "gee whiz" devices. The fact that the CD-23 can be improved by
better engineering of its power supply caps is relevant.
Vince
vince c
accuracy.
One is free to make whatever statements, definitive or otherwise, on the
merits of two pieces of gear. It is always, of course, better to make those
statements on the basis of empirical evidence.
Vince
<mcn...@my-deja.com> wrote in message news:95mo...@news2.newsguy.com...
krom...@my-deja.com
pat...@popmail.dircon.co.uk wrote:
> krom...@my-deja.com writes:
>
> >> The Meridian genuinely buffers and totally reclocks the data from a
> >> large RAM. This is *not* true of the Levinson, despite what they'd
> >> like you to think. I'm not familiar with the Weiss or dB Tech models.
> >> Lots of companies make vague reference to jitter reduction and
> >> re-clocking (Linn even make the ludicrous claim that the CD12 has *no*
> >> clock!), but AFAIK the Meridian 861 is the only model which is a
> >> *true* reclocker, using a local free-running clock with no reliance
> >> whatever on the embedded clock in the incoming datastream.
> >
> >Stewart, do you have any more technical information about how the
> >Meridian does this? For example, what I'm thinking about is how the
> >system deals with the fact that two fully independant clocks will never
> >be running at 100% the same speed, one will always be a little faster
> >than the other.
>
> They use a FIFO RAM buffer. I don't know the exact details
>
> >Does the Meridian toss away or replicate samples if the FIFO becomes
> >full or empty? (Or perhaps it tries to interpolate samples?)
>
> Use a decent size buffer, you have no problem. A couple of seconds
> worth of buffering is adequate for any input within class 1 limits for
> clock accuracy.
Well, not really. What if the Meridian clock is .00001% faster than
the input coming from the source device? It may take minutes or
hours, but eventually the FIFO will run dry.
Of course, if it takes days or weeks, then perhaps for all intents
and purposes the issue is irrelevant.
(The chances are pretty good that some cheap equipment out there is
running a not-insignificant amount slower than S/PDIF spec, but the
chances are slim that you own any of them if you also own a Meridian
pre/pro :-) )
Andy K.
Sverre Brubaek
Pinkerton) wrote:
[snip]
>They use a FIFO RAM buffer. I don't know the exact details
>
>>Does the Meridian toss away or replicate samples if the FIFO becomes
>>full or empty? (Or perhaps it tries to interpolate samples?)
>
>Use a decent size buffer, you have no problem. A couple of seconds
>worth of buffering is adequate for any input within class 1 limits for
>clock accuracy.
That assumes a finite playing time, which is OK for cd's i guess, but
will get into problems with other types of sources. For instance
would it handle a day of input from my mp3-pc (assuming it never
sends digital mute)?
>
>>If it uses some kind of Negative Feedback (ie. a PLL) to try to sync
>>the clocks, then I imagine some percentage of jitter effect would be
>>propagated...
>
>It doesn't, although it does of course use different clocks for
Quite frankly this seems like a hazardous design, are you absolutely
positive that the two clocks are entirely independent or might just
Meridian have snuck a fast one by you?
>different input sampling rates. As an aside, I've just been informed
>of another true reclocker, the Bel Canto. This uses the alternative
>system of also using a free-running clock, but employing an
>asynchronous sample rate converter to avoid worries about buffering.
That is a genuinely safe method. Although i'm not certan they use
sufficiently long windows to avoid audibility (i'm guessing they do
however)
>What effect this has on the sound, I don't know.
--
They both savoured the strange warm glow of being much more ignorant
than ordinary people, who were only ignorant of ordinary things.
-- Discworld scientists at work (Terry Pratchett, Equal Rites)
Ed Seedhouse
>> The essence of a DBT is that you *must*
>> 'trust your ears', because you don't *know* which player is connected.
>Then we have no quarrel on this point. But the conditions have to be set up
>correctly. I can recall when AR did an infamous test as part of an
>advertising campaign.
I don't think "infamous" is a fair word for this. "Famous" fits the bill
better, but there was *no* claim that this was a "single blind" experiment
and no attempt to derive statistics for analysis from it. It was just what
it was advertized to be, a single blind demonstration (not a test).
Any blind or double blind study would of course have to be much more
rigorously designed than that, but since it didn't clam to be a study
[Moderator's Note: Not enough mussel? RD]
what's the problem?
Ed Seedhouse
Victoria, B.C.
Stewart Pinkerton
It's only relevant if the CD-12 cannot be improved in the same manner
- and who says that Black Gates would produce a better engineered
power supply in any case? Evidence, please? You seem fond of staing as
'fact' what is mere speculation.
State Of The Art is simply that, it's not perfection. Your basic point
is of course correct, a poor implementation of the latest technology
is not necessarily an improvement in the state of the art. As noted
before however, you can't just state 'the CD-12 is audibly better'
without offering some proof other than its price tag.
Stewart Pinkerton
>In article <95mo...@news2.newsguy.com>,
> pat...@popmail.dircon.co.uk wrote:
>> krom...@my-deja.com writes:
>> >Does the Meridian toss away or replicate samples if the FIFO becomes
>> >full or empty? (Or perhaps it tries to interpolate samples?)
>>
>> Use a decent size buffer, you have no problem. A couple of seconds
>> worth of buffering is adequate for any input within class 1 limits for
>> clock accuracy.
>
>Well, not really. What if the Meridian clock is .00001% faster than
>the input coming from the source device? It may take minutes or
>hours, but eventually the FIFO will run dry.
Do the math. If the Meridian clock is 0.00001% faster, that's less
than 1 second in three months!
If it's 0.01% faster, a more reasonable but still pretty sloppy
situation, that's 1 second in three hours.
>Of course, if it takes days or weeks, then perhaps for all intents
>and purposes the issue is irrelevant.
>
>(The chances are pretty good that some cheap equipment out there is
>running a not-insignificant amount slower than S/PDIF spec, but the
>chances are slim that you own any of them if you also own a Meridian
>pre/pro :-) )
Quite so, and the class 1 limit I quoted for clock accuracy is 50ppm,
which is 1 second in 5.5 hours.
Ed Seedhouse
>You are right. As I recall, the advertisement didn't claim to be a study.
>It was provided to allegedly show that AR speakers sounded like the real
>thing, i.e. live music. You ask what the problem was. It misleadingly
>demonstrated that live music was reproduced so accurately by a speaker that
>one could not tell the difference between the reproduced sound and the
>original music.
It wasn't particularly misleading in this claim so far as I can see,
or at least no more misleading than the vast run of advertisements.
In fact a lot of listeners couldn't tell the difference. Sure that
was under special conditions, but it did at least show that under
some conditions AR speakers could sound enough like a live string
quartet to fool a good many people.
You notice you didn't see an avalanche of other speaker makers doing
the same demo. If it was easy to do maybe they might have hopped on
the bandwagon.
Sure there is an implication that this is because the AR speakers of
the day were so good, and no mention in the ad, for example, of the
equalization applied to the high frequencies of the AR speakers used
for the test . Heck it was an *advertisement*, whose purpose is
always to make the advertisers product look good. If you are going
to get a head of steam up over adds like this then you're going to
spend your entire life in a stew, since they are all around you every
day, most of them far worse than that ancient AR ad.
At least AR had to give paying jobs to some string quartet players,
and that can't be such a bad thing, it might even justify the add all
by itself. :-)
I hardly think that makes it "infamous" in any normal usage of that
word. Since it was effective and a lot of people still remember it,
that makes it "famous" pretty much by definition.
Ed Seedhouse
Victoria, B.C.
Stewart Pinkerton
>The evidence lies in improving the power supply in one unit and
>leaving another unit unchanged. You can do this if you are a dealer
>in the units, as a close friend of mine is. Then you run the
>testing. Regarding the CD-12, I don't think I ever used the "price
>tag" as evidence one was audibly superior.
You offered no other 'evidence', nor have you offered any for your
unsupported claim that adding Black Gate electrolytics will somehow
magically 'improve' the power supply in the Arcam. Now you expect us
to believe that an Arcam dealer will happily void the warranty on a
brand new CD-23 by changing the power supply capacitors.
Indeed, you have offered zero evidence in support of *any* of your
claims in this thread, merely vague statements apparently based only
on the fact that the CD-12 is a very expensive player made by Linn!
If you have any *real* evidence that the CD-23 can be improved, or
that the Linn CD-12 even sounds *different* from the Arcam, then let's
hear it. Otherwise...............
krom...@my-deja.com
pat...@popmail.dircon.co.uk wrote:
> krom...@my-deja.com writes:
>
> >In article <95ocf0$do1$1...@bourbaki.localdomain>,
> > pat...@popmail.dircon.co.uk wrote:
> >> >Of course, if it takes days or weeks, then perhaps for all intents
> >> >and purposes the issue is irrelevant.
> >> >
> >> >(The chances are pretty good that some cheap equipment out there is
> >> >running a not-insignificant amount slower than S/PDIF spec, but the
> >> >chances are slim that you own any of them if you also own a Meridian
> >> >pre/pro :-) )
> >>
> >> Quite so, and the class 1 limit I quoted for clock accuracy is 50ppm,
> >> which is 1 second in 5.5 hours.
> >
> >equipment, you need to be careful that any digital devices you plug
> >into it are also 'Class 1' rated. Interesting knowledge.
>
> Not really - check those numbers again.
>
> >I wonder what would happen if both devices were not out of spec
> >individually, but collectively they were to the point where you may get
> >a 1-sec skip every couple of hours. Surely this would be a customer
> >support nightmare?
>
> If you could find a CD which lasts 2 hours, this would certainly be
> true.................
What about a 3-hour movie? What about listening to digital radio for
an evening?
Yes, the actual number of glitches is low, but its still interesting
that Meridian thought customers wouldnt mind it.
You may be on to something with some kind of a 'backup PLL'. Either
that, or it could have 8 or 10 different clock frequencies, some
slower than spec and some faster than spec, that it switches to to
compensate?
Andy K.
[quoted text deleted -- deb]
dcs
news:95p5n8$n2g$1...@bourbaki.localdomain...> news:95k2j...@news1.newsguy.com...
>
> We effectively did the test as a double blind, so let me explain more
fully.
>
> We had both players connected at the same time, with the same digital
cables
> to 2 digital inputs on the 861, both players playing the same CD, we tried
3
> CD's in total. We started both players in sync, so bar a small pause while
> the Meridian switched inputs there was no discontinuity in the music, and
of
> course the same volume level.
>
> With my brother in control of the remote switching between the two I chose
> the Marantz every time as the better sounding player, the difference was
> obvious.
> I did the same thing for my brother and he also picked the Marantz as the
> better player.
>
> We did some further testing with his Pioneer DVD player as a CD transport,
> and that sounded worse than the Meridian, we also tried the digital output
> of a Pioneer CD-ROM in his PC, this sounded very close to the Meridian.
>
> If we assume the Meridian 861 is doing it's D to A conversion the same
> regardless of the amount of jitter in the incoming digital signal, then
the
> issue of transports sounding different must be down to more than jitter
> levels on the output of the transport.
> Either that or the Meridian actually is affected by jitter.
> I will try to get hold of their technical people this afternoon and ask,
> they were very helpful in the past with technical questions about my 557
> power amp.
>
I spoke to technical people at Meridian today. They said although the 861
does a much better job of reclocking the signal than virtually all D-A
converters it can not be said to be totally immune to jitter in the incoming
digital signal, as he put it "It can't work miracles".
Jon
Stewart Pinkerton
>In article <95q2og$3g7$1...@bourbaki.localdomain>,
> pat...@popmail.dircon.co.uk wrote:
>> krom...@my-deja.com writes:
>> >I wonder what would happen if both devices were not out of spec
>> >individually, but collectively they were to the point where you may get
>> >a 1-sec skip every couple of hours. Surely this would be a customer
>> >support nightmare?
>>
>> If you could find a CD which lasts 2 hours, this would certainly be
>> true.................
>
>What about a 3-hour movie? What about listening to digital radio for
>an evening?
BZZZZZT! DVDs are packet-switched, and are by nature reclocked. Do
your homework! Radio signals tend to have quite immaculate frequency
accuracy, else they lose their licenses.............
vince c
this subject as long as it disagrees with your views. But your
presumption that no Arcam dealer would void the warranty to improve
the CD-23 is just that- a presumption. And it is in error. The
dealer has done exactly that, for a customer who understands the
warranty is voided by any change in the stock unit. But none of this
is obviously going to convince you since you persist in the claim
that this is already a state of the art unit.
Vince
"Stewart Pinkerton" <pat...@popmail.dircon.co.uk> wrote in message
news:95s17...@news2.newsguy.com...
Sverre Brubaek
Pinkerton) wrote:
[snip]
>>>Use a decent size buffer, you have no problem. A couple of seconds
>>>worth of buffering is adequate for any input within class 1 limits for
>>>clock accuracy.
>>
>>That assumes a finite playing time, which is OK for cd's i guess, but
>>will get into problems with other types of sources. For instance
>>would it handle a day of input from my mp3-pc (assuming it never
>>sends digital mute)?
>
>One would hope that you wouldn't be using MP3 with a Meridian 861, but
If I were to buy an expensive device like that, i'd bloody well play
what i bloody well like! ;-)
Allthough I have only done blindtesting of MP3's with cbr at 160 and
below which failed at 160, but the test could be made more sensitive
than what I did. I think I would have problems detecting the
differences with my collection of vbr encoded files (average at
180-200kbps).
>I don't know the answer to that one. It may be that the Meridian has a
>secondary PLL clock which can kick in if the RAM is in danger of under
>or overflowing.
That would be a possibility, it would be hell to ensure a safe
handoff from a crystal generated clock to a pll generated one but it
should be doable.
Maybe somone would be so kind as to figure out the size of the buffer
then we can figure somthing about the rejection. (Which now basically
has been reduced to the same as for a pll with a really long
integration time)
Stewart Pinkerton
>"Stewart Pinkerton" <pat...@popmail.dircon.co.uk> wrote in message
>news:95s17...@news2.newsguy.com...
>
>> "vince c" <vinc...@bellatlantic.net> writes:
>>
>> >The evidence lies in improving the power supply in one unit and
>> >leaving another unit unchanged. You can do this if you are a dealer
>> >in the units, as a close friend of mine is. Then you run the
>> >testing. Regarding the CD-12, I don't think I ever used the "price
>> >tag" as evidence one was audibly superior.
>
>> You offered no other 'evidence', nor have you offered any for your
>> unsupported claim that adding Black Gate electrolytics will somehow
>> magically 'improve' the power supply in the Arcam. Now you expect us
>> to believe that an Arcam dealer will happily void the warranty on a
>> brand new CD-23 by changing the power supply capacitors.
>>
>> Indeed, you have offered zero evidence in support of *any* of your
>> claims in this thread, merely vague statements apparently based only
>> on the fact that the CD-12 is a very expensive player made by Linn!
>>
>> If you have any *real* evidence that the CD-23 can be improved, or
>> that the Linn CD-12 even sounds *different* from the Arcam, then let's
>> hear it. Otherwise...............
>I'm not sure that anyone could ever offer you persuasive evidence on
>this subject as long as it disagrees with your views.
You miss the point, which is that *you* have offered *no* evidence to
persude *anyone*, only your own baseless views.
> But your
>presumption that no Arcam dealer would void the warranty to improve
>the CD-23 is just that- a presumption. And it is in error. The
>dealer has done exactly that, for a customer who understands the
>warranty is voided by any change in the stock unit.
Fine, since the dealer took no financial risk. So, where is the
description of 'the testing', and its results?
> But none of this
>is obviously going to convince you since you persist in the claim
>that this is already a state of the art unit.
Once again, you misrepresent what I said. I *specifically* stated that
a 'state of the art' unit was certainly capable of improvement - or
degradation - by modification. I note that you *still* offer
absolutely *zero* evidence that changing to Black Gate capacitors in
the power supply made any audible difference - in either direction.
Nor do you offer *any* evidence of the supposed superiority of the
Linn CD-12. It's all just baseless claims with you, never *any*
evidence of actual comparisons.
Ernst Raedecker
>Stuart,
>My brother has a Meridian 861 and a while ago we were playing around
>with the setup of his home theatre system, different modes of
>bi-amping his KEF Reference 4's etc.
>Out of interest we tried using my Marantz CD-10 as a transport
>instead of his Meridian (200 series, I think it's a 206).
>The result was the Marantz sounded considerably better, more natural
>and transparent.
>The Marantz had a Trichord Clock 3 and some power supply upgrades
>fitted, which may have helped make the jitter on the digital output
>lower than that of his Meridian transport (can't say for sure as we
>had no way of measuring it).
Oh yes,
It's the Trichord you are hearing. Or better, the good effects of the
Trichord + power supply.
You know, reducing jitter is *really* important, because it is mostly
the jitter that makes for the notoriously glassy, grey, sharp,
shallow and grainy sound of most cd players, not one DAC or the
other, provided the DAC is of reasonable quality.
Many people believe "that all modern players have good clocks", "that
all clocks are the same", "that the factories would fit better clocks
if it really did matter" etc etc etc.
All not true.
No, all modern players do not have good clock circuits. Even the
players with special circuits are not really good.
Yes, clocks do differ in precision and jitter, and also the
stabilizing circuits do differ.
No, factories do *not* know what a Trichord or a SuperClock or an
LClockXO plus separate power supply do to your system. They have
never heard. They do not believe, hence they feel no need to hear, to
test, to try. I have experienced this lack of interest from the
factories myself.
You know, those companies believe that they have done a good job when
they reduce jitter from say 250 ps to 150 ps or even a bit lower.
Not enough. Even the old Trichords, reducing jitter to say 12 or 15
ps are not enough.
The new clocks (Trichord III, LClock XO) have a precision of about
5ppm instead of the regular 100ppm and reduce jitter to 2 pikosecs.
THAT is the difference. THAT is what you pay for. THAT is what you
hear.
It's the clock that gives you the extra detail, the microinformation
etc., and it's the separate supply that takes away the sharpness, and
adds extra detail. Unfortunately a Trichord or other clock is
expensive.
There is a lot of good material on the internet on jitter, jitter
reduction and why upsampling does/does not work. I'm not a
technician, but as it seems, all circuits on the board generate
jitter, not only the clock, and because all circuits are linked, if
only via the power supply, you must give the Trichord a supply of its
own.
Of course a good DAC is important too, but, well, modern players of
quality all have the Burr Brown's or the Philips etc. I don't think
Arcam's RingDAC is so very important for sound, jitter reduction is
way more important. I tell you this from deep, personal experience
with the Arcam 9.
And there is a lot, an awful lot to be gained by substituting the
cheap opamps in the after DAC filtering circuit for better ones, if
you have active filtering. Again, not cheap. You don't believe the
staggering difference if you haven't heard. Of course for the details
to come out of the filtering you must put them in first, so start
with your Trichord.
Ernest
Norbert Hahn
>>My brother has a Meridian 861 and a while ago we were playing around
>>with the setup of his home theatre system, different modes of
>>bi-amping his KEF Reference 4's etc.
I downloaded the user's manual for the Meridian 861 and could
not find any hints on jitter rejection. The word is mentioned
on those 120+ pages but not details. Some other Meridian D/A
converters were famous for their jitter rejection, though.
Does some kind soul have any numbers?
[snip]
>
>Of course a good DAC is important too, but, well, modern players of
>quality all have the Burr Brown's or the Philips etc. I don't think
>Arcam's RingDAC is so very important for sound, jitter reduction is
>way more important. I tell you this from deep, personal experience
>with the Arcam 9.
Again, depending on the architecture of the DAC, both the chips
used and their environment, you get more or less jitter rejection.
And that contains the question:
How does the RingDAC translate jitter into noise/distortion
at the output?
That value may differ from that of Burr Brown, Philips, Crystal
Semi, ...
Norbert
Ate Siemensma
"Norbert Hahn" <ha...@hrz.tu-darmstadt.de> wrote in message
news:96k0s...@news2.newsguy.com...