Audio, hi-fi and car audio - Hard Disc Player Sound Quality

Audio Banter (https://www.audiobanter.co.uk/forum.php)

- uk.rec.audio (General Audio and Hi-Fi) (https://www.audiobanter.co.uk/uk-rec-audio-general-audio/)

- - Hard Disc Player Sound Quality (https://www.audiobanter.co.uk/uk-rec-audio-general-audio/1521-hard-disc-player-sound-quality.html)

Hard Disc Player Sound Quality

In article , Chris Morriss
wrote:

Provided that the jitter is less than (say) 25% of the data bit period,
then a two-loop PLL with a fast loop feeding a clock to the data
receiver, and a slow PLL then feeding a clean clock to the DAC, will
eliminate any problem.

The 'two loops' method is, I think, the one that Meridian have often used.
In my experience even with only the 'fast' loop working their outboard DACs
seem pretty effective.

Of course, if the short term jitter is greater than 50% of the bit
period then there's no chance!

Depends upon how long 'short term' might be. :-) If it is withing a
single slot then yer stuffed. However if it is over a few slots or more you
can deal with it by suitable means. But it would be much better if the
source were decent in the first place. Hopefully, no normal sources are
anything like that bad ... although given some of what I've seen sold at
times I should perhaps not be over-confident of this. ;- (See below.)

I use the CS8414 data receiver and CS4396 DAC at work at the moment, and
have no jitter problem. The SPDIF signal from the CD sources at work,
and my own CD63 and CD52SE have jitter less than 25% of the bit period.
(That's checked by looking at the eye diagram of the data stream.)

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

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

In article , Stewart Pinkerton
wrote:

No, you can *not* achieve as low phase-noise from a PLL as you can from
a free-runnijng clock. That is why one-box CD players almost always have
significantly lower jitter than outboard DACs.

I think this is something of an 'it depends' matter. It depends upon the
timescales you are concerned with for the jitter or phase noise.

When I used to work on 100 - 300 GHz oscillator systems for astronomy it
was routine to use quite fancy PLL systems to reduce the 'long' and 'medium
term' phase noise (and hence jitter) of sources that were fairly stable.
This did reduce the total integrated rms phase noise levels up to, say, a
few hundred kHz modulation frequency. However if you pushed this, you got
extra noise at high modulation frequencies - i.e. made the short term
jitter worse.

This was reducing the jitter by controlling the oscillator, though. So this
becomes a question of what you're doing, and why. In effect, with a good
loop you end up largely imposing the phase noise of one oscillator on
another within the loop bandwidth, and leaving it unchanged outwith the
loop bandwidth. How you use this, and what effect it has, depends upon how
you are applying the loop to the problem.

With the kinds of recovery loops I think people are talking about here the
effect of the loop is to give you a 'new' clock for the data that does not
have any 'fast' modulation that may have been present in the original, so
it should suppress short-term jitter. However it won't remove any slower
jitter or drift as the loop will then prompt the receiver LO to track any
errors in the source. Also, this assumes the receiver LO is better than
the source in terms of the short term phase noise. It may not be in
all cases.

For the 'fast' (short timescale) jitter you then may have an actual
reduction in the magnitude of the jitter if the receiver clock is cleaner
than the sideband noise level on the received signal. But note the 'if'.

In the context of music the slower variations in sample/bit rate frequency
probably don't matter if small in size as people aren't likely to notice.
The short timescale jitter would arguably be more audible, so is worth
removing if you can.

Hence I suspect people may arguing at cross purposes at little in this
thread as you need to specify timescales, etc, to discuss what may or may
not be occuring.

What any specific audio DAC does, though, is less clear... :-)

Also debatable how often this actually matters... :-)

And if you do have a low-noise clock, then as Stewart has pointed out, you
might as well use that for everything and avoid any need for reclocking if
you can. ;-

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 Fri, 23 Jan 2004 15:21:14 +0000 (GMT)
Jim Lesurf wrote:

And if you do have a low-noise clock, then as Stewart has pointed out,
you might as well use that for everything and avoid any need for
reclocking if you can. ;-

Just to be clear, I have nothing against that idea *at all* except that
its somewhat impossible on SP/DIF ;-) (without modifying it in evil
ways, of course ;-)

--
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 09:26:41 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 08:29:52 +0000 (UTC)
(Stewart Pinkerton) wrote:

I guess mine isnt the only one...

The difference is that I both understood my error, and was happy to
admit it.

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

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

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality

On Fri, 23 Jan 2004 09:28:56 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 08:29:52 +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.

plus it's a plain fact that far too many
outboard DACs have poorly-designed input receivers and PLLs which
allow jitter to be transmitted from the transport with little
attenuation.

Im sure thats the case. nontheless it has no bearing on my specific
point.

Sure it does, because it bears on the reality of digital audio as
implemented, about which you are woefully uninformed.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality

On Fri, 23 Jan 2004 09:31:27 +0000, Ian Molton wrote:

On Fri, 23 Jan 2004 08:29:53 +0000 (UTC)
(Stewart Pinkerton) wrote:

Of course not, since these systems, as I previously stated, merely
*attenuate* jitter, they do *not* eliminate it. For that, you need a
single master clock controlling the entire process.

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! With a single master clock, you don't
*need* a PLL, and you *can* totally reclock the signal when it reaches
the DAC. 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.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality

On Fri, 23 Jan 2004 15:03:25 +0000 (GMT), Jim Lesurf
wrote:

In article , Chris Morriss
wrote:

Provided that the jitter is less than (say) 25% of the data bit period,
then a two-loop PLL with a fast loop feeding a clock to the data
receiver, and a slow PLL then feeding a clean clock to the DAC, will
eliminate any problem.

The 'two loops' method is, I think, the one that Meridian have often used.
In my experience even with only the 'fast' loop working their outboard DACs
seem pretty effective.

Yes, indeed.

Of course, if the short term jitter is greater than 50% of the bit
period then there's no chance!

Depends upon how long 'short term' might be. :-) If it is withing a
single slot then yer stuffed. However if it is over a few slots or more you
can deal with it by suitable means. But it would be much better if the
source were decent in the first place. Hopefully, no normal sources are
anything like that bad ... although given some of what I've seen sold at
times I should perhaps not be over-confident of this. ;- (See below.)

Indeed so, and it often seems that the more you pay, the worse is the
performance!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality

On Fri, 23 Jan 2004 15:21:14 +0000 (GMT), Jim Lesurf
wrote:

In article , Stewart Pinkerton
wrote:

No, you can *not* achieve as low phase-noise from a PLL as you can from
a free-runnijng clock. That is why one-box CD players almost always have
significantly lower jitter than outboard DACs.

I think this is something of an 'it depends' matter. It depends upon the
timescales you are concerned with for the jitter or phase noise.

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

When I used to work on 100 - 300 GHz oscillator systems for astronomy it
was routine to use quite fancy PLL systems to reduce the 'long' and 'medium
term' phase noise (and hence jitter) of sources that were fairly stable.
This did reduce the total integrated rms phase noise levels up to, say, a
few hundred kHz modulation frequency. However if you pushed this, you got
extra noise at high modulation frequencies - i.e. made the short term
jitter worse.

This was reducing the jitter by controlling the oscillator, though. So this
becomes a question of what you're doing, and why. In effect, with a good
loop you end up largely imposing the phase noise of one oscillator on
another within the loop bandwidth, and leaving it unchanged outwith the
loop bandwidth. How you use this, and what effect it has, depends upon how
you are applying the loop to the problem.

Quite so, and there's the rub with audio PLLs, as the available VCOs
are simply not so stable/quiet as the available free-running quartz
oscillators that you find in even quite cheap one-box players.

With the kinds of recovery loops I think people are talking about here the
effect of the loop is to give you a 'new' clock for the data that does not
have any 'fast' modulation that may have been present in the original, so
it should suppress short-term jitter. However it won't remove any slower
jitter or drift as the loop will then prompt the receiver LO to track any
errors in the source. Also, this assumes the receiver LO is better than
the source in terms of the short term phase noise. It may not be in
all cases.

Very true, and that's the problem. If Sony/Philips had ever had any
idea that people would be crazy enough to use separate transports and
DACs for CD, I'm sure that a better data link would have been
standardised from the outset, allowing a nice stable low-noise DAC
clock to control the transport, whose internal jitter performance is
essentially irrelevant.

For the 'fast' (short timescale) jitter you then may have an actual
reduction in the magnitude of the jitter if the receiver clock is cleaner
than the sideband noise level on the received signal. But note the 'if'.

In the context of music the slower variations in sample/bit rate frequency
probably don't matter if small in size as people aren't likely to notice.
The short timescale jitter would arguably be more audible, so is worth
removing if you can.

Yes, the narrow bandwidth loop of the Meridian DACs seems to be very
effective, and they always did provide one of the lowest jitters
available from a two-box player. Of course, their one-box players are
even better, as you'd expect!

Hence I suspect people may arguing at cross purposes at little in this
thread as you need to specify timescales, etc, to discuss what may or may
not be occuring.

What any specific audio DAC does, though, is less clear... :-)

Also debatable how often this actually matters... :-)

And if you do have a low-noise clock, then as Stewart has pointed out, you
might as well use that for everything and avoid any need for reclocking if
you can. ;-

Indeed, although regrettably, only a *very* few companies like the
late lamented TAG-McLaren actually do this in the domestic field. Pro
gear of course does this as standard, with the SDIF-2 link.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Hard Disc Player Sound Quality