Re: OT question about sound cards/chip-sets and high-end music systems

On 20091016_151335, ghe wrote:

On Oct 16, 2009, at 1:50 PM, Paul E Condon wrote:

Analog signals degrade on long cable runs, particularly
the high freq. part of the signal.

Not if it's low impedance balanced, it doesn't. Not at 100' anyway.

Impedance and balance are two different things. Impedance only becomes
an issue when the wave length of the signal on the cable becomes
comparable to the length of the cable run.

That's interesting -- I never heard that. Maybe I just never dealt with
cables long enough. I'd like to learn more about it (off list would be
more appropriate, I suspect).

Actually my experience was with cable about 100m long, but with
signals that were ~10ns pulses. We wanted to measure time of arrival
to somewhat better than 1ns. The speed of light is about 30cm/ns.
Because of the dielectric constant of the insulator in coax cable, the
speed of signals in cable is close to 20cm/ns. We used RG58U cable
which is designed to have a characteristic impedance of 50 ohms.
Without termination, pulses would reflect off the ends of the cable
and 'echo' back and forth in the cable. Echos of earlier pulses would
corrupt the wave form of following pulses, and really mess things up.

If anyone claims that this sort of thing corrupts analog audio signals
in a significant way, I would mark it as marketing obfuscation, not
proper electrical engineering. But I'm really not sure. I don't
believe proper double blind listening tests have ever been conducted.
Such tests are expensive, and the question really doesn't have the
social importance of drug safety and efficacy.

I do know that I can tell the difference between hifi system sound and
desktop PC sound and I prefer the former. And I have a lot of
theoretical knowledge that inclines me to be skeptical of explanations
offered by marketing people.

OTOH, almost anything learned during the early days telephony should be
taken into account when doing the analog part of a modern sound system.
All sorts of unwanted sounds and distortions of wanted sounds can happen.
Much of that has dropped form view in the mad dash to computerize the
world.


Balance OTOH has only to do
with rejection of common mode environmental noise, e.g. hum pickup, not
with loss of signal amplitude.

Yup. But hums and pops and buzzes and stuff definitely count as
degradation.

Well yes, of course. It was being suggested that I just buy an adapter
cable and try it. But I know already that an adapter cable alone will
give performance that I find unsatisfactory. I gave one good reason,
you give three more. Any one of them, alone, would invalidate the
test.


But this is theoretical knowledge. It precludes me from believing much
of the marketing pitch of consumer grade electronics. I'm hoping to
find some practical information that is in better conformance the
established theory.

To an old audio engineer, that's gratifying to hear: today's hifi
marketing is astounding BS. And if you want some really practical info
about bandwidth and noise, get an oscillator and a meter and measure it.

I think you'll find that, coming from a modern solid state amp
(vanishingly low output impedance), you'll be hard pressed to find
significant high frequency loss on just about any 100' long piece of
cable. Your definition of "significant" is allowed to differ from mine,
however.

I'm older now than when I bought the hifi. Hearing declines with
age. But I can still tell the difference between the sound from my
computer and from my hifi.

I'd first suspect the DAC/ADCs and the analog circuitry in the computers
-- in yours and the one that digitized in the first place. Or maybe the
digital sound's sample rate. Or, of course, the file could be (badly done)
mp3 or one of its buds...

It may be that the age of real hifi has
passed, just as the age of the vacuum tube has passed, but I'm hoping
not (for real hifi. I don't mind the new dominance of transistors.)

I was around at the transition. And the early solid state amps were,
indeed, pretty nasty. Then somebody discovered how easy and cheap op-amps
were :-(

Op-amps were not easy until Nyquist published his famous criterion. So by the
time transistors came on the scene that had been done, but most hifi firms
didn't want to pay their engineers to read his papers.


But they've learned how to work with silicon, and things are much better
now. These days, a properly designed solid state amplifier is at least as
good as could be done with vacuum tubes -- and a lot quieter and more
reliable.

Thanks for reading to the end of this rant.

You're welcome :-)

--
Glenn English
ghe@xxxxxxxxxxx




--
To UNSUBSCRIBE, email to debian-user-REQUEST@xxxxxxxxxxxxxxxxxxxx a
subject of "unsubscribe". Trouble? Contact listmaster@xxxxxxxxxxxxxxxx


--
Paul E Condon
pecondon@xxxxxxxxxxxxxxxx


--
To UNSUBSCRIBE, email to debian-user-REQUEST@xxxxxxxxxxxxxxxx
with a subject of "unsubscribe". Trouble? Contact listmaster@xxxxxxxxxxxxxxxx

Relevant Pages

  • Re: RS485 is bidirectional does it mean it is fullduplex?
    ... >>>Now consider the case where the left hand ground moves up to 80V away from ... >>>better include some impedance in the connection that limits dangerous ... >>>frequency signals. ... subject than comm cables. ...
    (comp.arch.embedded)
  • How does co-ax cable actually work?!?
    ... >>>the same impedance. ... It has characteristic impedence which is ... This is regardless of any load on the 'far end' since the signals haven;t ... > One thing that many people overlook is that cables with the same ...
    (uk.tech.digital-tv)
  • Re: Composite vs. Component Cables
    ... since the characteristic impedance of the RCA connector creates a huge ... Each consist of 3 shielded rca cables. ... expect good results with high-bandwidth signals. ... RCA connectors don't have a characteristic impedance. ...
    (alt.tv.tech.hdtv)
  • Re: Recording digital data to analog tape... revisited
    ... The signals were transformer coupled for impedance ... > signal is zero over the differential line pair. ... I'll look into that idea if I work with long cables again and have ...
    (comp.arch.embedded)
  • Re: Recording digital data to analog tape... revisited
    ... The signals were transformer coupled for impedance ... > signal is zero over the differential line pair. ... I'll look into that idea if I work with long cables again and have ...
    (sci.electronics.design)