Hard Disc Player Sound Quality
 

On Fri, 23 Jan 2004 17:31:31 +0000 (UTC)
(Stewart Pinkerton) wrote:

The local clock will be a VCO, and hence will *not* have as low
jitter as a free-running clock,

There is nothing that says it *cant* though. just because one given
design doesnt achieve the goal doesnt mean the concept is flawed.

Yeah, right, so now you're going to pretend that you were only talking
theoretically? Pathetic.

Demonstrate Im not. Show me one reference I made to a specific DAC and
the way it works.

--
Spyros lair: http://www.mnementh.co.uk/ |||| Maintainer: arm26 linux

Do not meddle in the affairs of Dragons, for you are tasty and good with
ketchup.

Hard Disc Player Sound Quality
 

On Fri, 23 Jan 2004 17:34:00 +0000 (UTC)
(Stewart Pinkerton) wrote:

bull****. A single master clock may give you a nice, precise reference
so that synchronising your sources is easy, but it doesnt (necessarily)
have any impact on jitter.

Of course it does, you cretin!

Your single master clock could have worse jitter than the host clock in a SP/DIF setup.

the fact that you send the same clock signal to all the devices does NOT mean the signal is jitter free. that depends *solely* on the quality of the clock. All you would be doing was making all the devices jitter the same way.

With a single master clock, you don't *need* a PLL,

Duh.

and you *can* totally reclock the signal when it reaches
the DAC.

It wouldnt be 'reclocking' as such, just clocking. you're effectively describing a synchronous audio bus. The computer industry has used similar busses for years. Congratulations.

In practice, the free-running low-noise clock would of course
be placed as close as possible to the DAC chip, as this is the only
place where jitter actually matters.

Heres where I dont have the know how to argue, so I hope someone I trust the opinion of will step in (Jim?)

Can cable length increase jitter? I know it can increase capacitance / inductance etc. but surely that would apply the same characteristics to all pulses on the line, thus not altering the jitter, so long as the pulse edges can be reliably picked out ?

If not, then I doubt it matters where you place the clock.

In fact, for audio use, Im sure it doesnt matter, as Jodrell bank does radio interferometry using much higher frequencies over far greater distances, which also requires a synchronised time source.

--
Spyros lair: http://www.mnementh.co.uk/ |||| Maintainer: arm26 linux

Do not meddle in the affairs of Dragons, for you are tasty and good with ketchup.

Hard Disc Player Sound Quality
 

In message , Jim Lesurf
writes

Is that 25% from one slot to the next? Or an averaged eye-pattern 'blur' on
screen showing a number of slots?


It was using the recovered clock from the CS8414 to sync the scope, and
looking at the cumulative (blur!) transitions on the data edges from the
CD player, for one bit period only.

This is valid, as this is what the chip sees when using the recovered
clock to sample the incoming data. The recovered clock itself has got
jitter on it as it is from a fast PLL. (Very fast in fact. Have a look
at the data for the CS8414 on the Cirrus web site. They provide very
comprehensive data sheets, and give graph of jitter rejection versus
frequency.)

You can then see why a second PLL to clean up the clock going to the DAC
is useful!
--
Chris Morriss

Hard Disc Player Sound Quality
 

In message , Ian Molton
writes

Can cable length increase jitter? I know it can increase capacitance /
inductance etc. but surely that would apply the same characteristics to
all pulses on the line, thus not altering the jitter, so long as the
pulse edges can be reliably picked out ?


Both optical fibre and co-ax are 'dispersive mediums' so they will cause
deterioration of the edges of the data (nothing to do with HF roll-off).
Whether this is relevant over the short lengths is another matter!
--
Chris Morriss

Hard Disc Player Sound Quality
 

On Fri, 23 Jan 2004 18:47:44 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 17:29:47 +0000 (UTC)
(Stewart Pinkerton) wrote:


The difference is that you *were* in error. I am not.

Yes you are, but you seem unable to understand this fact.

Two other people than myself have said the same as me. One of which
I certainly value the opinion of. (the other I dont know but hes right anyhow)

No, they most certainly have *not* said the same as you. You are so
far in denial that you've left Egypt...................
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality
 

On Fri, 23 Jan 2004 18:48:58 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 17:31:31 +0000 (UTC)
(Stewart Pinkerton) wrote:

The local clock will be a VCO, and hence will *not* have as low
jitter as a free-running clock,

There is nothing that says it *cant* though. just because one given
design doesnt achieve the goal doesnt mean the concept is flawed.

Yeah, right, so now you're going to pretend that you were only talking
theoretically? Pathetic.

Demonstrate Im not. Show me one reference I made to a specific DAC and
the way it works.

Yeah, like I said, pathetic.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality
 

On Fri, 23 Jan 2004 18:07:06 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 17:50:22 +0000 (UTC)
(Stewart Pinkerton) wrote:

I'm talking about the devices actually used in digital audio, rather
than what is theoretically possible.

I was talking about the theoretical, so that might have been your problem.

No, you're just trying to duck out of admitting that you were wrong.
I'm not the one with the problem, here...............
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality
 

On Fri, 23 Jan 2004 18:56:30 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 17:34:00 +0000 (UTC)
(Stewart Pinkerton) wrote:

bull****. A single master clock may give you a nice, precise reference
so that synchronising your sources is easy, but it doesnt (necessarily)
have any impact on jitter.

Of course it does, you cretin!

Your single master clock could have worse jitter than the host clock in a SP/DIF setup.

The moon could be made of green cheese. Grow up.

the fact that you send the same clock signal to all the devices does NOT mean the signal is jitter free. that depends *solely* on the quality of the clock. All you would be doing was making all the devices jitter the same way.

Back here in the real world, it's a simple fact that one-box players
almost always have significantly lower jitter than two-box solutions -
and the exceptions either reclock or use a DAC-based master clock.

With a single master clock, you don't *need* a PLL,

Duh.

and you *can* totally reclock the signal when it reaches
the DAC.

It wouldnt be 'reclocking' as such, just clocking. you're effectively describing a synchronous audio bus. The computer industry has used similar busses for years. Congratulations.

Yes, I knew that, thanks. The point is that the high end audio
industry is chronically incompetent.

In practice, the free-running low-noise clock would of course
be placed as close as possible to the DAC chip, as this is the only
place where jitter actually matters.

Heres where I dont have the know how to argue, so I hope someone I trust the opinion of will step in (Jim?)

Well, maybe you should just stay out of technical discussions, huh?

Can cable length increase jitter? I know it can increase capacitance / inductance etc. but surely that would
apply the same characteristics to all pulses on the line, thus not altering the jitter, so long as the pulse
edges can be reliably picked out ?

A poorly shielded cable can allow jitter to be introduced.

If not, then I doubt it matters where you place the clock.

As you already stated, you don't know enough about the subject.

In fact, for audio use, Im sure it doesnt matter, as Jodrell bank does radio interferometry using much
higher frequencies over far greater distances, which also requires a synchronised time source.

As you already stated, you don't know enough about the subject.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality
 

In article , Ian Molton
wrote:
On Fri, 23 Jan 2004 17:34:00 +0000 (UTC) (Stewart
Pinkerton) wrote:


In practice, the free-running low-noise clock would of course be
placed as close as possible to the DAC chip, as this is the only place
where jitter actually matters.

Heres where I dont have the know how to argue, so I hope someone I trust
the opinion of will step in (Jim?)

OK. Yes and (academically at this point) No. :-)

Yes. In that what really matters is that the data samples are converted
into analogue output at (ideally) perfectly regular intervals. Otherwise
you would be distorting the output by adding a form of phase modulation. If
you can't do this, then all else is a bit futile.

No. In that there is a slight risk that jitter might be so bad as to cause
a bit to be lost (or duplicated) on the way to the DAC somewhere. Leading
to the system getting into a muddle. However this is pretty unlikely if the
system is a reasonable one, but may require some extra work at an earlier
stage.

In practice I assume that decent players/DACs employ suitable buffering,
etc, to be able to ensure that the data is read from the disc reliably.
Once this is done and the data is in memory, the primary requirement is
that the samples be converted in a regular stream. This is where a really
smooth, low noise, clock is desirable. It also means that any other
systematic errors that affect timing should be dealt with. Cables between
boxes may not be the only way that 'data induced jitter' can arise.

Can cable length increase jitter? I know it can increase capacitance /
inductance etc. but surely that would apply the same characteristics to
all pulses on the line, thus not altering the jitter, so long as the
pulse edges can be reliably picked out ?

What really matters here is the non-dispersive bandwidth of the cables. The
(theoretical) problem is that bandwidth limited or dispersive cables rounds
over the edges of the waveforms and does not allow them to reach as close
to the nominal '0' and 1' levels after short level assertions as after
longer ones.

The S/PDIF waveform mixes together both the data pattern and a clock
signal. *IF* the receiver uses the timing of the 'edges' of the waveform as
it clock indication the result is that the apparent clock will then jitter
with the data pattern *even if the source had a perfect clock*.

In the event described above, the receiver might then produce irregular
output samples, and the sound becomes modulated by the effect of the finite
bandwidth or dispersion of the cables used, via the daft use of the signal
edges as the clock.

in effect, the data pattern becomes a form of sideband phase-noise so far
as the received clock info is concerned. The finite bandwidth and
dispersion of the cables acts as a sort of AM-to-FM convertor (not the
correct discription, but an analogy) to do this and confuse the receiver.

The purpose of the receiver loops (or whatever) would be to overcome this
kind of problem.)

In most cases in practice, though, I doubt this should matter much. :-)

If not, then I doubt it matters where you place the clock.

In fact, for audio use, Im sure it doesnt matter, as Jodrell bank does
radio interferometry using much higher frequencies over far greater
distances, which also requires a synchronised time source.

That is different for various reasons. The key one is that the clock is
kept *seneparate* to any data. The reference clock is sent between antenna
stations in a way that is not 'corrupted' by data transfer. The S/PDIF
system (and various comms systems) puts the clock and data together to use
a single channel. This is cheap and works neatly, but has potential
limitations.

In theory, it does not matter where, physically, you locate the master
clock. However this is 'provided you can pipe it around with no problems
due to noise, etc'. Hence it would not matter if a master clock was in a
player, a DAC, or what, but this is complicated in domestic cases by the
combined clock/data method of S/PDIF.

FWIW the longest baselines I know of were back in the Pioneer/Voyager days.
The spacecraft transmitter was phase locked to the signals it received from
Earth (I think the ratio was 222/221 or some-such.) Thus here the phase
lock was acheived over a loop distance greater than the diameter of the
major planets of the solar system! :-) This was useful for control
purposes, and also for getting a 'doppler' indicator of the spacecraft
velocity in the line of sight.

Conventional long-baseline interometry will use a master clock in one
country and then send it to the others. This may be via microwave or otehr
links (which has problems) or via carrying 'atomic' clocks about (which has
other problems). There the baselines are typically a few thousand km.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Hard Disc Player Sound Quality
 

On Sat, 24 Jan 2004 07:48:53 +0000 (UTC)
(Stewart Pinkerton) wrote:

Demonstrate Im not. Show me one reference I made to a specific DAC and
the way it works.

Yeah, like I said, pathetic.

I note theres no reference cited there...


--
Spyros lair: http://www.mnementh.co.uk/ |||| Maintainer: arm26 linux

Do not meddle in the affairs of Dragons, for you are tasty and good with ketchup.