Speakers and wire length

Guest
Since the Pioneer SX-434 is working so well, thanks in part to advice
I recieved here, I'm going to hang the speakers in the shop and use
it. One speaker will be about two feet from the SX-434 while the other
will be 30 feet away. Will it make that big of a difference if the
speaker wires are of such different lengths? I have enough speaker
wire laying around that I could make them both the same length. I
could just coil up the wire to near speaker.
Thanks,
Eric
 
If the wires are heavy-gauge, it shouldn't make any difference. If they're
thin, yes, you might hear a difference in bass output (maybe). I would make
them the same length.
 
On 4/1/2014 11:57 AM, etpm@whidbey.com wrote:
Since the Pioneer SX-434 is working so well, thanks in part to advice
I recieved here, I'm going to hang the speakers in the shop and use
it. One speaker will be about two feet from the SX-434 while the other
will be 30 feet away. Will it make that big of a difference if the
speaker wires are of such different lengths? I have enough speaker
wire laying around that I could make them both the same length. I
could just coil up the wire to near speaker.
Thanks,
Eric

How many uh inductor is your speaker wire coil going to be? ;-)

On a note I found interesting;
The local marine tech was here at the marina to
do a welding repair on a boat. For some reason
he could not get the welder to work properly. The
problem turned out to be the excess cable was wound
up and piled on his trailer, making an inductor. The
inductor was enough to limit the High frequency current.
He straightened out the cable and all worked well.
Another thing I found odd, the High frequency was generated
using a spark gap. Seems so old tech, but I guess it is still
common in welders.
Mikek
 
If you really want the scoop on tis, here is the technalia. (sp)

An amplifier has what is called a damping factor. Tube amps are usually low, solid state amps are higher. Just looked and yours is rated at 25. this assumes that means into eight ohms. This is because that is what all the other readings are taken at.

the damping (or dampering n the old days) factor is the ratio of the speaker impedance to the effective impedance of the output of the amp. what is means is the amp is an absolutely perfect voltage source with unlimited current, but there is a resistor in series with it.

Since your damping factor is 25, that means the effective output impedance of the amp is one twenty-fifth of eight ohms. Tht is about 0.32 ohms, which is probably the value of the emitter resistors, but those do NOT determine the output impedance and therefor the damping factor. They are almost always witin the feedback loop so their resistance is nullified, mostly that is..

A higher damping factor means that the voltage output is not affected as much with variations of the load impedance. For example, if you connect another speaker to the same cannel, the original speaker will not decrease in vlume. Instead of the voltage dropping, the amp puts out more current, to maintain the voltage output which is determined by the input and your settings..

I have a Phas Linear 400-2, which has a damping factor of 1,000, actually the highest I have ever seen. It is not just done with feedback, it takes current drive to do it, which is how it does do it. With enough feedback you can enhance the damping factor, but then the amp could become unstable with highly reactive loads. It makes for enough math to give you a headache, better off just to design it to not need too much feedback.

NOW TO THE QUESTIO ! TADA !

Here's a table on wire resistance :

http://www.interfacebus.com/AWG-table-of-different-wire-gauge-resistance.html

Let's just say you are using 16 guage wire which has a resistance of four ohms per thousand foot. Of course there are two wires so that means eight ohms, IF you are using a thousand feet. If the impedance of the speaker is a perfect eight ohms, you will lose exactly half of the power. It is also not in the feedback loop so that means when the speaker impedance is lower, like at lower bass frequencies, there will be more loss which will affect the low end.

But you are not using a thousand feet. Let's take it as 100 feet then, one tenth. That means that if the damping factor of the amp was infinite, after the wire you are left with a damping factor of ten. Is it one tenth of the load (speaker) impedance. At twenty feet, which is one fifth of that, just multiply by five because that's the side of the equation you are on.

The actual damp(er)ing factor as seen by the speaker itself is the vector sum of the wire resistance or other inpedance and the actual source resistance (damping factor/load) of the amplifier in quesrtion.

In other words it realy doesn't matter all that much. Also you could just kick it up to 14 guage, which is only 2.5 ohms per thousand feet.

All in all you are not going to hear that much difference. It takes ten times the power to seem twice as loud. A doubling or halving of power results ion a very small difference in percieved volume. It is a bit easier to discern when it comes to frequency response/tonal balance, but not all that much.

A fifty foot run that introduces 0.2 ohms resistance, with a speaker that is eight ohms but drops to two ohms at 25 Hz (if it can even reproduce that), will have a loss at 25 Hz of ten percent of the voltage.

Double or half the voltage is APPROXIMATELY 3 dB. the difference here, I am pretty sure is not even one dB, and even the most golden ears in the world will have a hard time hearing that.

So indeed, all the hype about monster cable and all that shit is just that - hype. On a long ruin you CAN hear the difference but you will have to do a a/b compare to notice. I actually have. I used to have a bench that was set up so I could do that.

Funny though, that some people would probably prefer the "degraded" sound. Really.
 
On 04/01/2014 9:57 AM, etpm@whidbey.com wrote:
Since the Pioneer SX-434 is working so well, thanks in part to advice
I recieved here, I'm going to hang the speakers in the shop and use
it. One speaker will be about two feet from the SX-434 while the other
will be 30 feet away. Will it make that big of a difference if the
speaker wires are of such different lengths? I have enough speaker
wire laying around that I could make them both the same length. I
could just coil up the wire to near speaker.
Thanks,
Eric

If it is just background music, then you can use almost anything on
hand. If you are going for music quality with some volume, then minimum
18 gauge.

Louder? 16 gauge.

Stadium? 00 gauge. (right...)

John :-#)#

--
(Please post followups or tech inquiries to the newsgroup)
John's Jukes Ltd. 2343 Main St., Vancouver, BC, Canada V5T 3C9
(604)872-5757 or Fax 872-2010 (Pinballs, Jukes, Video Games)
www.flippers.com
"Old pinballers never die, they just flip out."
 
<jurb6006@gmail.com> wrote:
If you really want the scoop on tis, here is the technalia. (sp)

An amplifier has what is called a damping factor. Tube amps are usually
low, solid state amps are higher. Just looked and yours is rated at 25.
this assumes that means into eight ohms. This is because that is what all
the other readings are taken at.

the damping (or dampering n the old days) factor is the ratio of the
speaker impedance to the effective impedance of the output of the amp.
what is means is the amp is an absolutely perfect voltage source with
unlimited current, but there is a resistor in series with it.

Since your damping factor is 25, that means the effective output
impedance of the amp is one twenty-fifth of eight ohms. Tht is about 0.32
ohms, which is probably the value of the emitter resistors, but those do
NOT determine the output impedance and therefor the damping factor. They
are almost always witin the feedback loop so their resistance is
nullified, mostly that is.

A higher damping factor means that the voltage output is not affected as
much with variations of the load impedance. For example, if you connect
another speaker to the same cannel, the original speaker will not
decrease in vlume. Instead of the voltage dropping, the amp puts out more
current, to maintain the voltage output which is determined by the input
and your settings.

I have a Phas Linear 400-2, which has a damping factor of 1,000, actually
the highest I have ever seen. It is not just done with feedback, it takes
current drive to do it, which is how it does do it. With enough feedback
you can enhance the damping factor, but then the amp could become
unstable with highly reactive loads. It makes for enough math to give you
a headache, better off just to design it to not need too much feedback.

NOW TO THE QUESTIO ! TADA !

Here's a table on wire resistance :

http://www.interfacebus.com/AWG-table-of-different-wire-gauge-resistance.html

Let's just say you are using 16 guage wire which has a resistance of four
ohms per thousand foot. Of course there are two wires so that means eight
ohms, IF you are using a thousand feet. If the impedance of the speaker
is a perfect eight ohms, you will lose exactly half of the power. It is
also not in the feedback loop so that means when the speaker impedance is
lower, like at lower bass frequencies, there will be more loss which will
affect the low end.

You will loose half of the available power, which will be 1/4 to the
speaker.


But you are not using a thousand feet. Let's take it as 100 feet then,
one tenth. That means that if the damping factor of the amp was infinite,
after the wire you are left with a damping factor of ten. Is it one tenth
of the load (speaker) impedance. At twenty feet, which is one fifth of
that, just multiply by five because that's the side of the equation you are on.

The actual damp(er)ing factor as seen by the speaker itself is the vector
sum of the wire resistance or other inpedance and the actual source
resistance (damping factor/load) of the amplifier in quesrtion.

In other words it realy doesn't matter all that much. Also you could just
kick it up to 14 guage, which is only 2.5 ohms per thousand feet.

All in all you are not going to hear that much difference. It takes ten
times the power to seem twice as loud. A doubling or halving of power
results ion a very small difference in percieved volume. It is a bit
easier to discern when it comes to frequency response/tonal balance, but
not all that much.

A fifty foot run that introduces 0.2 ohms resistance, with a speaker that
is eight ohms but drops to two ohms at 25 Hz (if it can even reproduce
that), will have a loss at 25 Hz of ten percent of the voltage.

Double or half the voltage is APPROXIMATELY 3 dB. the difference here, I
am pretty sure is not even one dB, and even the most golden ears in the
world will have a hard time hearing that.

6 dB. Very noticeable.

Greg


So indeed, all the hype about monster cable and all that shit is just
that - hype. On a long ruin you CAN hear the difference but you will have
to do a a/b compare to notice. I actually have. I used to have a bench
that was set up so I could do that.

Funny though, that some people would probably prefer the "degraded" sound. Really.
 
wrote in message news:6294b77d-1990-42e7-bd22-bdee135eac6e@googlegroups.com...

> Double or half the voltage is APPROXIMATELY 3 dB.

6 dB

By the way, the term has always been "damping factor". Not dampering (there's
no such word) or dampening.
 
wrote in message
news:6294b77d-1990-42e7-bd22-bdee135eac6e@googlegroups.com...

If you really want the scoop on tis, here is the technalia. (sp)

An amplifier has what is called a damping factor. Tube amps are usually low,
solid state amps are higher. Just looked and yours is rated at 25. this
assumes that means into eight ohms. This is because that is what all the
other readings are taken at.

the damping (or dampering n the old days) factor is the ratio of the
speaker impedance to the effective impedance of the output of the amp. what
is means is the amp is an absolutely perfect voltage source with unlimited
current, but there is a resistor in series with it.

Since your damping factor is 25, that means the effective output impedance
of the amp is one twenty-fifth of eight ohms. Tht is about 0.32 ohms, which
is probably the value of the emitter resistors, but those do NOT determine
the output impedance and therefor the damping factor. They are almost always
witin the feedback loop so their resistance is nullified, mostly that is.

A higher damping factor means that the voltage output is not affected as
much with variations of the load impedance. For example, if you connect
another speaker to the same cannel, the original speaker will not decrease
in vlume. Instead of the voltage dropping, the amp puts out more current, to
maintain the voltage output which is determined by the input and your
settings.

I have a Phas Linear 400-2, which has a damping factor of 1,000, actually
the highest I have ever seen. It is not just done with feedback, it takes
current drive to do it, which is how it does do it. With enough feedback you
can enhance the damping factor, but then the amp could become unstable with
highly reactive loads. It makes for enough math to give you a headache,
better off just to design it to not need too much feedback.

NOW TO THE QUESTIO ! TADA !

Here's a table on wire resistance :

http://www.interfacebus.com/AWG-table-of-different-wire-gauge-resistance.html

Let's just say you are using 16 guage wire which has a resistance of four
ohms per thousand foot. Of course there are two wires so that means eight
ohms, IF you are using a thousand feet. If the impedance of the speaker is a
perfect eight ohms, you will lose exactly half of the power. It is also not
in the feedback loop so that means when the speaker impedance is lower, like
at lower bass frequencies, there will be more loss which will affect the low
end.

But you are not using a thousand feet. Let's take it as 100 feet then, one
tenth. That means that if the damping factor of the amp was infinite, after
the wire you are left with a damping factor of ten. Is it one tenth of the
load (speaker) impedance. At twenty feet, which is one fifth of that, just
multiply by five because that's the side of the equation you are on.

The actual damp(er)ing factor as seen by the speaker itself is the vector
sum of the wire resistance or other inpedance and the actual source
resistance (damping factor/load) of the amplifier in quesrtion.

In other words it realy doesn't matter all that much. Also you could just
kick it up to 14 guage, which is only 2.5 ohms per thousand feet.

All in all you are not going to hear that much difference. It takes ten
times the power to seem twice as loud. A doubling or halving of power
results ion a very small difference in percieved volume. It is a bit easier
to discern when it comes to frequency response/tonal balance, but not all
that much.

A fifty foot run that introduces 0.2 ohms resistance, with a speaker that is
eight ohms but drops to two ohms at 25 Hz (if it can even reproduce that),
will have a loss at 25 Hz of ten percent of the voltage.

Double or half the voltage is APPROXIMATELY 3 dB. the difference here, I am
pretty sure is not even one dB, and even the most golden ears in the world
will have a hard time hearing that.

So indeed, all the hype about monster cable and all that shit is just that -
hype. On a long ruin you CAN hear the difference but you will have to do a
a/b compare to notice. I actually have. I used to have a bench that was set
up so I could do that.

Funny though, that some people would probably prefer the "degraded" sound.
Really.







On the other hand, that receiver has a control labelled "Balance". If you
find the speakers unbalanced, you might want to tweak that instead.



Gareth.
 
On Tue, 1 Apr 2014 15:02:56 -0700 (PDT), jurb6006@gmail.com wrote:

If you really want the scoop on tis, here is the technalia. (sp)

An amplifier has what is called a damping factor. Tube amps are usually low, solid state amps are higher. Just looked and yours is rated at 25. this assumes that means into eight ohms. This is because that is what all the other readings are taken at.

the damping (or dampering n the old days) factor is the ratio of the speaker impedance to the effective impedance of the output of the amp. what is means is the amp is an absolutely perfect voltage source with unlimited current, but there is a resistor in series with it.

Since your damping factor is 25, that means the effective output impedance of the amp is one twenty-fifth of eight ohms. Tht is about 0.32 ohms, which is probably the value of the emitter resistors, but those do NOT determine the output impedance and therefor the damping factor. They are almost always witin the feedback loop so their resistance is nullified, mostly that is.

A higher damping factor means that the voltage output is not affected as much with variations of the load impedance. For example, if you connect another speaker to the same cannel, the original speaker will not decrease in vlume. Instead of the voltage dropping, the amp puts out more current, to maintain the voltage output which is determined by the input and your settings.

I have a Phas Linear 400-2, which has a damping factor of 1,000, actually the highest I have ever seen. It is not just done with feedback, it takes current drive to do it, which is how it does do it. With enough feedback you can enhance the damping factor, but then the amp could become unstable with highly reactive loads. It makes for enough math to give you a headache, better off just to design it to not need too much feedback.

NOW TO THE QUESTIO ! TADA !

Here's a table on wire resistance :

http://www.interfacebus.com/AWG-table-of-different-wire-gauge-resistance.html

Let's just say you are using 16 guage wire which has a resistance of four ohms per thousand foot. Of course there are two wires so that means eight ohms, IF you are using a thousand feet. If the impedance of the speaker is a perfect eight ohms, you will lose exactly half of the power. It is also not in the feedback loop so that means when the speaker impedance is lower, like at lower bass frequencies, there will be more loss which will affect the low end.

But you are not using a thousand feet. Let's take it as 100 feet then, one tenth. That means that if the damping factor of the amp was infinite, after the wire you are left with a damping factor of ten. Is it one tenth of the load (speaker) impedance. At twenty feet, which is one fifth of that, just multiply by five because that's the side of the equation you are on.

The actual damp(er)ing factor as seen by the speaker itself is the vector sum of the wire resistance or other inpedance and the actual source resistance (damping factor/load) of the amplifier in quesrtion.

In other words it realy doesn't matter all that much. Also you could just kick it up to 14 guage, which is only 2.5 ohms per thousand feet.

All in all you are not going to hear that much difference. It takes ten times the power to seem twice as loud. A doubling or halving of power results ion a very small difference in percieved volume. It is a bit easier to discern when it comes to frequency response/tonal balance, but not all that much.

A fifty foot run that introduces 0.2 ohms resistance, with a speaker that is eight ohms but drops to two ohms at 25 Hz (if it can even reproduce that), will have a loss at 25 Hz of ten percent of the voltage.

Double or half the voltage is APPROXIMATELY 3 dB. the difference here, I am pretty sure is not even one dB, and even the most golden ears in the world will have a hard time hearing that.

So indeed, all the hype about monster cable and all that shit is just that - hype. On a long ruin you CAN hear the difference but you will have to do a a/b compare to notice. I actually have. I used to have a bench that was set up so I could do that.

Funny though, that some people would probably prefer the "degraded" sound. Really.
Thanks for the long winded reply. It was very interesting. What I
ended up doing is running 30 feet of 16 gauge to one speaker and 10
feet of 18 gauge to the other. That's because one speaker only had
about 6 inches of 18 gauge coming out of the box and the other one had
10 feet or so. The 16 gauge wire I had just sitting around on a spool
so it got connected to the speaker with the short wires. The speakers
are 10 feet up on a wall and 30 feet apart. They sound great but I
think would sound a little better if they were closer together.
Eric
 
"amdx"

How many uh inductor is your speaker wire coil going to be? ;-)

** Coiling a twin lead cable creates no additional inductance.

Since there are equal and opposite currents flowing in adjacent wires it is
a "non inductive" wind.



..... Phil
 
In article <lheu3l$979$1@dont-email.me>, nojunk@knology.net says...
On 4/1/2014 11:57 AM, etpm@whidbey.com wrote:
Since the Pioneer SX-434 is working so well, thanks in part to advice
I recieved here, I'm going to hang the speakers in the shop and use
it. One speaker will be about two feet from the SX-434 while the other
will be 30 feet away. Will it make that big of a difference if the
speaker wires are of such different lengths? I have enough speaker
wire laying around that I could make them both the same length. I
could just coil up the wire to near speaker.
Thanks,
Eric


How many uh inductor is your speaker wire coil going to be? ;-)

On a note I found interesting;
The local marine tech was here at the marina to
do a welding repair on a boat. For some reason
he could not get the welder to work properly. The
problem turned out to be the excess cable was wound
up and piled on his trailer, making an inductor. The
inductor was enough to limit the High frequency current.
He straightened out the cable and all worked well.
Another thing I found odd, the High frequency was generated
using a spark gap. Seems so old tech, but I guess it is still
common in welders.
Mikek

What you saw most likely was the arc starter at the tip, which means it
must of been a tig welder of sorts.

The arc is generated at the tip down on to the work piece where it can
create the plasma needed to create a conductive path to the work piece
for the main current, which in most cases is high frequency from an
inverter.

Jamie
 
On Tue, 1 Apr 2014 10:25:23 -0700, William Sommerwerck wrote:
If the wires are heavy-gauge, it shouldn't make any difference. If they're
thin, yes, you might hear a difference in bass output (maybe). I would make
them the same length.

It's more likely there will be a difference noted due
to the distance involved in the physical separation.
IF any difference is noted at all.
I believe the human ear is more sensitive to time
differences than intensity differences.
 
On 4/1/2014 5:02 PM, jurb6006@gmail.com wrote:
> It takes ten times the power to seem twice as loud.

Ten times the power is 10 dB, that is 10 times as loud. Twice
as loud would be 3 dB.

Bill
 
"Allodoxaphobia" wrote in message
news:slrnljo2mb.r8e.knock_yourself_out@vps.jonz.net...
On Tue, 1 Apr 2014 10:25:23 -0700, William Sommerwerck wrote:

If the wires are heavy-gauge, it shouldn't make any difference. If they're
thin, yes, you might hear a difference in bass output (maybe). I would make
them the same length.

It's more likely there will be a difference noted due
to the distance involved in the physical separation.
IF any difference is noted at all.
I believe the human ear is more sensitive to time
differences than intensity differences.

Think about how long it takes a signal to propagate 30' of wire, then decide
whether that would be audible.
 
"Bill Gill"
It takes ten times the power to seem twice as loud.

Ten times the power is 10 dB, that is 10 times as loud.

** Bollocks.

Most would say it was 2 or maybe 3 times as loud.


Twice as loud would be 3 dB.

** Bollocks again.

+3dB is only slightly louder.

+1dB is almost unnoticeable.


..... Phil
 
On Wed, 2 Apr 2014 06:28:37 -0700 "William Sommerwerck"
<grizzledgeezer@comcast.net> wrote in Message id:
<lhh3ab$rmv$1@dont-email.me>:

"Allodoxaphobia" wrote in message
news:slrnljo2mb.r8e.knock_yourself_out@vps.jonz.net...
On Tue, 1 Apr 2014 10:25:23 -0700, William Sommerwerck wrote:

If the wires are heavy-gauge, it shouldn't make any difference. If they're
thin, yes, you might hear a difference in bass output (maybe). I would make
them the same length.

It's more likely there will be a difference noted due
to the distance involved in the physical separation.
IF any difference is noted at all.
I believe the human ear is more sensitive to time
differences than intensity differences.

Think about how long it takes a signal to propagate 30' of wire, then decide
whether that would be audible.

I think he meant through the air, not the wire. At least I would hope so!
 
It's more likely there will be a difference noted due
to the distance involved in the physical separation.
IF any difference is noted at all.
I believe the human ear is more sensitive to time
differences than intensity differences.

Think about how long it takes a signal to propagate 30' of wire, then decide
whether that would be audible.

The ear/brain system can be sensitive to inter-aural timing
differences down to a few microseconds... it shifts the left/right
positioning of the apparent sound source. For timing shifts of this
magnitude to actually be audible, the rest of the timing needs to be
stable and fixed - e.g. you need to be listening to the two signals
via headphones. I used this trick to do a simple "one-channel sampled
sound, to two-channel pseudo-3D-sound" DSP system for a videogame
system design back in the 1990s.

30' of wire would work out to under 50 nanoseconds of added
delay... this is well below any timing shift that I was ever able to
show resulted in an audible shift of apparent position.

In addition, if you're listening over loudspeakers, even a tiny change
in the location of your ears (with respect to the speakers) will
introduce more more timing shift than this. The speed of sound in air
is roughly a million times slower than the speed of electricity in a
wire.

One effect of a "too long, too thin" speaker wire that I don't think
has been mentioned in this thread is that of a shift in the speaker's
effective frequency response. Speaker systems almost always do not
have anything approaching a constant impedance... it's quite high
(tens of ohms) at the resonance frequencies of the drivers, and low
(roughly the "nominal" impedance) at intermediate frequencies.

Throw an ohm or two of speaker-wire resistance into the mix, and the
speaker's output may be attenuated significantly (a dB or so?) at some
frequencies where the speaker's impedance is at a minimum, and
attenuated very little at the driver resonance frequencies where the
impedance is high. This might, possibly, make a perceptible change in
the speaker's tonal balance.

For most speaker systems with wire runs of reasonable length, using
any halfway-decent (i.e. inexpensive) 12- or 14-gauge stranded speaker
cable should be all you need. Any tonal shifts or other effects due
to its resistance should be well below audibility, and a lot less than
(e.g.) changes in the driver stiffness due to humidity changes, or
changes in your own hearing due to exposure to noise during the day,
caffeine intake, etc.
 
On Wed, 02 Apr 2014 09:38:07 -0400, JW wrote:
On Wed, 2 Apr 2014 06:28:37 -0700 "William Sommerwerck"
grizzledgeezer@comcast.net> wrote in Message id:
lhh3ab$rmv$1@dont-email.me>:

"Allodoxaphobia" wrote in message
news:slrnljo2mb.r8e.knock_yourself_out@vps.jonz.net...
On Tue, 1 Apr 2014 10:25:23 -0700, William Sommerwerck wrote:

If the wires are heavy-gauge, it shouldn't make any difference. If they're
thin, yes, you might hear a difference in bass output (maybe). I would make
them the same length.

It's more likely there will be a difference noted due
to the distance involved in the physical separation.
IF any difference is noted at all.
I believe the human ear is more sensitive to time
differences than intensity differences.

Think about how long it takes a signal to propagate 30' of wire, then decide
whether that would be audible.

I think he meant through the air, not the wire. At least I would hope so!

Yeppers.
 
"David Platt" wrote in message news:bdvt0b-oo.ln1@coop.radagast.org...

The ear/brain system can be sensitive to inter-aural timing
differences down to a few microseconds...

What is your evidence for this? Above about 1kHz, the brain doesn't recognize
phase or timing differences. Audible differences are likely in the range of
tens to hundreds of microseconds.

As for the question of the speakers being at different distances... If you're
trying to get a stereo image with one speaker ten yards farther away than the
other... who cares what kind of wire you're using!
 
"David Platt" wrote in message news:bdvt0b-oo.ln1@coop.radagast.org...

The ear/brain system can be sensitive to inter-aural timing
differences down to a few microseconds...

What is your evidence for this? Above about 1kHz, the brain doesn't recognize
phase or timing differences. Audible differences are likely in the range of
tens to hundreds of microseconds.

"Ten or less" is the figure I see published in a number of sources.
That's the range I was thinking of when I said "a few". I didn't mean
to imply "single digits". This was the sort of number I found in the
references I first looked up when I worked on the "pseudo 3D sound"
idea back about 20 years ago, and my experiments at that time seemed
to agree.

In the case of frequencies below 1 kHz or so, this timing difference
has the effect of shifting the phases of the signals arriving at the
two ears. The ear/brain system is definitely capable of detecting
small phase differences in frequencies in this range... phase shifts
corresponding to 10-15 microseconds are quite audible. The brain is
capable of detecting such phase differences in medium-frequency
sounds, even when a sine wave having this time as its frequency would
be too shrill to be audible at all. There's apparently a "delay line
and coincidence detection" architecture in the nucleus laminaris that
makes this possible. (Wikipedia article on "sound localization" has
some details and references to this).

In the system I put together, a single sampled-sound recording was
being fetched and played back by two independent data-grabbers
(pulling samples out of a circular memory buffer at different rates
and offsets). It used a simple linear-interpolation system to
synthesize sample values between the ones in the buffer... so, in
effect, it was able to resample the input data on the fly and
introduce variable amounts of phase shifts between the two playback
channels. Interaural delay, and Doppler shifting could be
simulated. (Yeah, linear interpolation isn't a hi-fi way of doing
this, but it was cheap to implement in a little flea-bite DSP that we
could tuck into the corner of the ASIC).

If I recall properly, the variable-rate sampler was able to sub-sample
down to 1/16 steps between the original samples. We normally used
CD-quality samples (44 ksamples/second). The time offset between
original samples was thus around 22 microseconds, and we could
fine-tune the offset between the two playback channels (ears) to
within 2 microseconds.

A full one-sample offset between the ears was quite audible over
headphones on a lot of program material. A single 1/16 sub-sample
offset (under 2 microseconds) was not. The point at which you'd begin
to detect a shift in the sound's apparent left/right position, as the
offset was increased, depended a lot on the program material. It's
been a long time since I ran the experiments, but I believe I remember
that the threshold was usually somewhere in the 5-8 microsecond
material (for my ears, on my test setup, with test audio I was very
familar with).

The fact that we were subsampling allowed us to "move" a sound from
left to right smoothly (or Doppler-shift it, or both), without causing
audible ticking and popping as the result of sudden full-sample jumps
in timing offset.

Blauert's classic text "Spatial Hearing" is a good reference for this
sort of thing. You can find a bunch of others cited in U.S. Patent
5,337,363 which is the one which came out the project I worked on.
 

Welcome to EDABoard.com

Sponsor

Back
Top