cable length needed to create phase shift?

[Patrick]

how much of a difference in cable length would it take to create a
noticeable phase difference between two signals coming from the same
source simultaneously? i.e., wouldn't a 1000' xlr cable arrive at the
pre, recorder, etc. much later than the signal that only had to travel
through a 25' xlr cable?
thanks,
patrick

[ScotFraser]

<< i.e., wouldn't a 1000' xlr cable arrive at the
pre, recorder, etc. much later than the signal that only had to travel
through a 25' xlr cable? >>

The signal travels at the speed of light. Yes, theoretically it will arrive
later through the longer cable. No human will ever detect the arrival
difference, though.


Scott Fraser

[Scott Dorsey]

Patrick > wrote:
>how much of a difference in cable length would it take to create a
>noticeable phase difference between two signals coming from the same
>source simultaneously? i.e., wouldn't a 1000' xlr cable arrive at the
>pre, recorder, etc. much later than the signal that only had to travel
>through a 25' xlr cable?


Okay, let's pick worst case. You have a 20 KHz signal, and you want to
delay it a quarter wave. One cycle is 0.05 milliseconds, so a quarter
wave is 0.0125 milliseconds.

This really isn't much of a difference, but it's enough to be noticeable
if you're picky and sum it with the original signal and have really good
ears and monitors.

Now, C is 186 miles per millisecond. Going through Belden cat-5 cable
gives you a speed of .68C, which is 126.48 miles per millisecond.

So, at that rate, how much cable do you need to go 0.0125 milliseconds?
About a mile and a half. Not a huge amount, but more than you'll encounter
in the studio. A thousand feet is nothing.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Jay Kadis]

In article >
(Patrick) writes:
> how much of a difference in cable length would it take to create a
> noticeable phase difference between two signals coming from the same
> source simultaneously? i.e., wouldn't a 1000' xlr cable arrive at the
> pre, recorder, etc. much later than the signal that only had to travel
> through a 25' xlr cable?
> thanks,
> patrick

At the velocity electrons travel in metal (nearly 186,000 miles/sec or almost a
trillion (10^9) feet/sec), that length difference wouldn't matter.

A 20 kHz sine wave would have a wavelength of about 50,000 feet, or nearly 10
miles. So a 1000 foot difference in path length would produce about a 7 degree
phase shift at 20 kHz.

(Somebody check my math, please: that sounds like more than I expected.)

-Jay
--
x------- Jay Kadis ------- x---- Jay's Attic Studio ----x
x Lecturer, Audio Engineer x Dexter Records x
x CCRMA, Stanford University x http://www.offbeats.com/ x
x-------- http://ccrma-www.stanford.edu/~jay/ ----------x

[S O'Neill]

Light travels almost a billion feet per second in a vacuum. The signal in a
cable moves about 3/4 of that. It takes a little over 1.35 microsecond to
travel that 1000'. If a latency of 1 millisecond is acceptable (in MIDI circles
it seems to be), a cable run of up to about 740,000 feet (140 miles) should be ok.

But you'd have to coil that cable up to sit next to the 25' cable if you wanted
to find out if you could hear the delay.

Your first question, about a noticeable phase difference, varies with frequency
and your definition of "noticeable".


Patrick wrote:

> how much of a difference in cable length would it take to create a
> noticeable phase difference between two signals coming from the same
> source simultaneously? i.e., wouldn't a 1000' xlr cable arrive at the
> pre, recorder, etc. much later than the signal that only had to travel
> through a 25' xlr cable?
> thanks,
> patrick

[Ben Bradley]

In rec.audio.pro, (ScotFraser) wrote:

><< i.e., wouldn't a 1000' xlr cable arrive at the
>pre, recorder, etc. much later than the signal that only had to travel
>through a 25' xlr cable? >>
>
>The signal travels at the speed of light.

This is true for 'practical purposes' related to this question, but
a signal in a cable travels significantly slower than does light in a
vacuum. The figures I (vaguley) recall seeing vary from maybe 60
percent to 85 percent of lightspeed, and of course this varies with
construction and materials.

>Yes, theoretically it will arrive
>later through the longer cable. No human will ever detect the arrival
>difference, though.

Lightspeed is 186,000 miles per second, cable speed might be
150,000 miles per second. 1000 feet is about 1/5 of a mile. The time
to go that distance would be 1/5 * 1/150000 = 1/(5*150000) = 1/750000
= 1.3e-6 seconds, or 1.3 microseconds.
Her are couple of things for comparison:
The sample rate for CD's is 44100 per second or 1 sample every 22.7
microseconds.
The distance you would have to move a microphone away from a sound
source to give the same delay is ... sound goes at 1100 feet per
second, so in 1.3 microseconds sound would travel 1100*1.3e-6 =
0.00143 feet, or 0.017 inches, or almost two hundredths of an inch.

What could be a real problem with 1000 feet of cable is high
frequency rolloff due to the capacitance of the cable.

>Scott Fraser

[Chris Hornbeck]

On Thu, 30 Oct 2003 22:09:17 +0000 (UTC), (Jay
Kadis) wrote:

>At the velocity electrons travel in metal (nearly 186,000 miles/sec or almost a
>trillion (10^9) feet/sec), that length difference wouldn't matter.

Strangely enough, electrons themselves only travel about as fast
as a person can walk. A signal / impulse travels essentially in the
insulators of a cable, typically at about .7C.
At least, this is the number that several knowledgable folk on
this board accept for typical cables. In a recent thread, I
dredged a number of .1C out of ancient memory, and was soon
disabused of the idea.
There's also something called drift velocity, but I've not
seen it translated down to my level of understanding.
Curiouser and curiouser...


Chris Hornbeck
new email address

[Ben Bradley]

In rec.audio.pro, Chris Hornbeck >
wrote:

>On Thu, 30 Oct 2003 22:09:17 +0000 (UTC), (Jay
>Kadis) wrote:
>
>>At the velocity electrons travel in metal (nearly 186,000 miles/sec or almost a
>>trillion (10^9) feet/sec), that length difference wouldn't matter.

This looks to be the first error in the thread so far... the power
of 10 is correct, but it's called (at least in the USA) a billion - a
trillion would be 10^12. Or is a trillion actually 10^9 "over there?"
I know many of these words change meaning by three orders of magnitude
when they cross The Pond. I recall a "billiard" but I forget where it
goes in the sequence.

>Strangely enough, electrons themselves only travel about as fast
>as a person can walk.

Does anyone know how this was determined? I've always been
fascinated at how these things are discovered, like how the electron
(or 'quantum of') charge was determined with tiny charged oil drops
under a microscope.

>A signal / impulse travels essentially in the
>insulators of a cable, typically at about .7C.
>At least, this is the number that several knowledgable folk on
>this board accept for typical cables. In a recent thread, I
>dredged a number of .1C out of ancient memory, and was soon
>disabused of the idea.
>There's also something called drift velocity, but I've not
>seen it translated down to my level of understanding.

In this case the 'water pipe' analogy of electricity flow works
well. When you turn the valve to make water flow (into, let's say, a
hose already full of water but at low pressure), the pressure increase
travels at the speed of sound through water (which is I forget, but a
few times faster than it is in air), but the actual water only travels
a few feet per second.

>Curiouser and curiouser...
>
>
>Chris Hornbeck
>new email address

[Mike Rivers]

In article > writes:

> how much of a difference in cable length would it take to create a
> noticeable phase difference between two signals coming from the same
> source simultaneously? i.e., wouldn't a 1000' xlr cable arrive at the
> pre, recorder, etc. much later than the signal that only had to travel
> through a 25' xlr cable?

Well, the phase shift for a given delay depends on the frequency,
however allowing for velocity factor, the signals travel through a mic
cable at about 600 million feet per second. That's about 1.6
microseconds per 1000 feet.

At 1 kHz, that would be about 0.6 degrees of phase shift. I'd say that
wouldn't be noticable.




--
I'm really Mike Rivers - )

[Michael R. Kesti]

Ben Bradley wrote:

<snip>

> What could be a real problem with 1000 feet of cable is high
>frequency rolloff due to the capacitance of the cable.

Hmmm... Everybody's been discussing the propogation delay aspect of
the question, but the phase shift introduced by the RC filter that
results in that HF rolloff is far more significant. At frequencies
less than 1/10th that filter's corner frequency, the frequency at
which 6 dB of loss occurs, the phase shift is negligible. The shift
is 45 degrees at the corner frequency and grows to nearly 90 degrees
at frequencies about 10 times the corner frequency.

Depending on the signal's source impedance and the cable's capacitance,
then, relatively short cables can result in appreciable phase shift.

--
================================================== ======================
Michael Kesti | "And like, one and one don't make
| two, one and one make one."
| - The Who, Bargain

[Chris Hornbeck]

On Fri, 31 Oct 2003 01:33:35 GMT,
(Ben Bradley) wrote:

> I recall a "billiard" but I forget where it
>goes in the sequence.

It's equal to 10 ^ 3 snookers.

Chris Hornbeck
new email address

[Patrick]

thanks for all the replies, definitely a good double-check to my math!

[Pooh Bear]

Patrick wrote:

> how much of a difference in cable length would it take to create a
> noticeable phase difference between two signals coming from the same
> source simultaneously? i.e., wouldn't a 1000' xlr cable arrive at the
> pre, recorder, etc. much later than the signal that only had to travel
> through a 25' xlr cable?

A few picoseconds / nanoseconds later, as a guess, without getting the
calculator out.

In days of old, TV studios with multiple cameras used drums of cable to
delay the signal from some cameras, so that all of them were precisely
synched.

If your worried about time delays ( not actually phase shift ) , worry
about the different processing time in digital desks !

E.g. - if you patch in a compressor algorithm, the extra instructions
required will delay the signal by 'X' number of CPU cycles.


Graham

[William Sommerwerck]

The signal travels at close to the speed of light. Assume it's really slow (say,
half) and figure it out yourself.

Don't the schools teach math any more?

> how much of a difference in cable length would it take to create a
> noticeable phase difference between two signals coming from the
> same source simultaneously? ie, wouldn't a 1000' xlr cable arrive
> at the pre, recorder, etc. much later than the signal that only had to
> travel through a 25' xlr cable?

[William Sommerwerck]

> At the velocity electrons travel in metal (nearly 186,000 miles/sec or
> almost a trillion (10^9) feet/sec), that length difference wouldn't matter.

Actually, electrons travel very slowly in metal -- a fraction of an inch per
second, IIRC.

[Mike Turk]

"William Sommerwerck" > wrote in message
...
> > At the velocity electrons travel in metal (nearly 186,000 miles/sec or
> > almost a trillion (10^9) feet/sec), that length difference wouldn't
matter.
>
> Actually, electrons travel very slowly in metal -- a fraction of an inch
per
> second, IIRC.
>

Hmmnn ... that might be interesting .... a CD player might be done playing
the CD before you heard any music coming out of the speakers! <g>

-mike

[Arny Krueger]

"Michael R. Kesti" > wrote in message

> Ben Bradley wrote:
>
> <snip>
>
>> What could be a real problem with 1000 feet of cable is high
>> frequency rolloff due to the capacitance of the cable.
>
> Hmmm... Everybody's been discussing the propogation delay aspect of
> the question, but the phase shift introduced by the RC filter that
> results in that HF rolloff is far more significant. At frequencies
> less than 1/10th that filter's corner frequency, the frequency at
> which 6 dB of loss occurs, the phase shift is negligible. The shift
> is 45 degrees at the corner frequency and grows to nearly 90 degrees
> at frequencies about 10 times the corner frequency.
>
> Depending on the signal's source impedance and the cable's
> capacitance, then, relatively short cables can result in appreciable
> phase shift.

Yes, but there are frequency response losses associated with delay obtained
this way. Most people who say they are looking for phase shift are looking
for phase shift without frequency response losses. If moderate amounts of
delay are desired, all-pass networks can be a solution for this requirement.
If more delay is desired, then digital delays can be good and priced
reasonably. However, all such delay must be causal - IOW the delay has a
positive amount of time associated with it.

In a DAW environment practically arbitrary amounts of what, from the
viewpoint of the recording amounts to be causal or acausal delay are
basically *free*.

[Rob Adelman]

William Sommerwerck wrote:
>>At the velocity electrons travel in metal (nearly 186,000 miles/sec or
>>almost a trillion (10^9) feet/sec), that length difference wouldn't matter.
>
>
> Actually, electrons travel very slowly in metal -- a fraction of an inch per
> second, IIRC.

But doesn't each electron just bump the next one into submission? Kind
of like a relay race?

[Michael R. Kesti]

Rob Adelman wrote:

>William Sommerwerck wrote:
>>>At the velocity electrons travel in metal (nearly 186,000 miles/sec or
>>>almost a trillion (10^9) feet/sec), that length difference wouldn't matter.
>>
>>
>> Actually, electrons travel very slowly in metal -- a fraction of an inch per
>> second, IIRC.
>
>But doesn't each electron just bump the next one into submission? Kind
>of like a relay race?

Yes. Think of a hose filled with marbles. Push one more marble in at one
end and a marble falls out the other end almost instantaneously. The net
velocity of the individual marbles, however, is relatively low.

--
================================================== ======================
Michael Kesti | "And like, one and one don't make
| two, one and one make one."
| - The Who, Bargain

[Scott Dorsey]

In article . rogers.com>,
Mike Turk > wrote:
>
>"William Sommerwerck" > wrote in message
...
>> > At the velocity electrons travel in metal (nearly 186,000 miles/sec or
>> > almost a trillion (10^9) feet/sec), that length difference wouldn't
>matter.
>>
>> Actually, electrons travel very slowly in metal -- a fraction of an inch
>per
>> second, IIRC.
>
>Hmmnn ... that might be interesting .... a CD player might be done playing
>the CD before you heard any music coming out of the speakers! <g>

No, the wave travels at a good fraction of the speed of light, but the
electrons themselves travel much more slowly.

Imagine a row of ping pong balls. You tap one, and it taps the next one,
and the ball at the end of the row wiggles very quickly after the taps. But
it takes an awful lot of tapping before the next-to-last ball is moved into
the last ball position. The electrons move slowly, the wave travels quickly.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Jay Kadis]

In article > "William Sommerwerck"
> writes:
> > At the velocity electrons travel in metal (nearly 186,000 miles/sec or
> > almost a trillion (10^9) feet/sec), that length difference wouldn't matter.
>
> Actually, electrons travel very slowly in metal -- a fraction of an inch per
> second, IIRC.

As individual electrons, yes, but they force their neighbors to move and the
net effect is a flow at a decent fraction of the velocity of light.

-Jay
--
x------- Jay Kadis ------- x---- Jay's Attic Studio ----x
x Lecturer, Audio Engineer x Dexter Records x
x CCRMA, Stanford University x http://www.offbeats.com/ x
x-------- http://ccrma-www.stanford.edu/~jay/ ----------x

[Mike Rivers]

In article > writes:

> Strangely enough, electrons themselves only travel about as fast
> as a person can walk.

> Curiouser and curiouser...

There's molecular motion, and there's the time it takes for
electricity to flow through a wire. The clock doesn't stop when the
atom you charge at the one end of the wire arrives at the other end of
the wire (that might take quite a while), it stops when the effect of
putting charge on one end appears on the other end. That's the speed
of electromagnetic radiation (C) reduced 20 to 40 percent by practical
things.

What's curious is that people worry about phase shift caused by cables
in a studio.


--
I'm really Mike Rivers - )

[Jay Kadis]

In article <znr1067602252k@trad> (Mike Rivers) writes:
>
> In article >
writes:
>
> > Strangely enough, electrons themselves only travel about as fast
> > as a person can walk.
>
> > Curiouser and curiouser...
>
> There's molecular motion, and there's the time it takes for
> electricity to flow through a wire. The clock doesn't stop when the
> atom you charge at the one end of the wire arrives at the other end of
> the wire (that might take quite a while), it stops when the effect of
> putting charge on one end appears on the other end. That's the speed
> of electromagnetic radiation (C) reduced 20 to 40 percent by practical
> things.
>
> What's curious is that people worry about phase shift caused by cables
> in a studio.
>
>
> --
> I'm really Mike Rivers - )

I think it was more a case of wonder about than worry about. In analog, there
appears to be little reason to worry. The issue might be different if we're
talking about high SR word clock signals and very unequal-length cable runs.

-Jay
--
x------- Jay Kadis ------- x---- Jay's Attic Studio ----x
x Lecturer, Audio Engineer x Dexter Records x
x CCRMA, Stanford University x http://www.offbeats.com/ x
x-------- http://ccrma-www.stanford.edu/~jay/ ----------x

[Roger W. Norman]

How come everyone griped with Stephan Paul brought this up? <g>

--


Roger W. Norman
SirMusic Studio
Purchase your copy of the Fifth of RAP CD set at www.recaudiopro.net.
See how far $20 really goes.




"Scott Dorsey" > wrote in message
...
> In article . rogers.com>,
> Mike Turk > wrote:
> >
> >"William Sommerwerck" > wrote in message
> ...
> >> > At the velocity electrons travel in metal (nearly 186,000 miles/sec
or
> >> > almost a trillion (10^9) feet/sec), that length difference wouldn't
> >matter.
> >>
> >> Actually, electrons travel very slowly in metal -- a fraction of an
inch
> >per
> >> second, IIRC.
> >
> >Hmmnn ... that might be interesting .... a CD player might be done
playing
> >the CD before you heard any music coming out of the speakers! <g>
>
> No, the wave travels at a good fraction of the speed of light, but the
> electrons themselves travel much more slowly.
>
> Imagine a row of ping pong balls. You tap one, and it taps the next one,
> and the ball at the end of the row wiggles very quickly after the taps.
But
> it takes an awful lot of tapping before the next-to-last ball is moved
into
> the last ball position. The electrons move slowly, the wave travels
quickly.
> --scott
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

[Kurt Albershardt]

Mike Rivers wrote:
>
> What's curious is that people worry about phase shift caused by cables
> in a studio.

It sure made a difference in analog TV...

[S O'Neill]

Kurt Albershardt wrote:

> Mike Rivers wrote:
>
>>
>> What's curious is that people worry about phase shift caused by cables
>> in a studio.
>
>
> It sure made a difference in analog TV...

....because the frequencies are about 1000 times as high, so the
distances are about 1000x less.

[Michael R. Kesti]

Kurt Albershardt wrote:

>Mike Rivers wrote:
>>
>> What's curious is that people worry about phase shift caused by cables
>> in a studio.
>
>It sure made a difference in analog TV...

I suppose that this may explain why some people concern themselves about it
in audio, but it's very much an apples and oranges comparison.

First, the frequencies involved in analog video are much higher and it
therefore takes much less cable length to generate specific amounts of shift.

Second, the phase shifts that result from propogation delay in analog video
cables result in color changes to which human eyes are very sensitive. A
shift of just one degree at color shubcarrier (3.58 MHz in NTSC) can be
easily detected. The human ear, on the other hand, is all but deaf to audio
phase shifts. For most people, such shifts become detectable only when the
unshifted and shifted signals are mixed and then only when shifts of tens
of degrees occur.

--
================================================== ======================
Michael Kesti | "And like, one and one don't make
| two, one and one make one."
| - The Who, Bargain