Phase adjustment on subwoofers

[damonrpayne]

So, I'm reading the 4th edition "Master Handbooke of Acoustics" right now, and I had a question about phase and the Phase

Given that the Phase of a sound wave determines its startposition relative to some other point, another sound wave usually, what is the value of adjusting phase on a subwoofer, such as mine? What is it being put out of phase with if I adjust, is it creating a delay from when the signal is received before its sent to the driver? Is this to compensate for room position or ?....

[CECAA850]

You are adjusting the sub to be in phase with your mains to help eliminate cancellation. Usually at the crossover point.

[damonrpayne]

Why would the sub be out of phase with the mains, its location? The electronics should send everything out at the proper time, correct?

[AVPhile135722]

Yes, your receiver does send everything out at the same time, but it does not compensate for your room acoustics and speaker placement. The placement of your subwoofer in relation to your front speakers may cause a phase issue, as the sound waves from each speaker will interact differently with your room and reach your ears at different times. The best way to adjust your phase will require an audio test tone CD and two people. You will need to wire your main speakers out of phase (i.e. reverse the + and - wires on the back of the speaker), then play the test tone at the crossover frequency (typically 80Hz). While you are seated in your typical listening position, have a buddy slowly turn the phase adjustment knob on the back of the subwoofer until you hear the LEAST amount of bass. At this point, your main speakers are completely (180 degrees) out of phase with your subwoofer. When you have reached this point, you will want to switch the wires back to normal on your main speakers, thus putting your main speakers perfectly in phase (0 degrees) with your subwoofer.

[DrWho]

Just bumping this into my 'my forums' so I can get back to it when I have more time...It's actually a pretty loaded question.

[SilverSport]

...It's actually a pretty loaded question.

Dr. Who...Horn Loaded???...Bill

[damonrpayne]

More than 24 hours later, I got "loaded" joke.

A little slow, maybe I'm loaded.

[DrWho]

Ok, let's see if I can do this without pictures (bare with me because I've had 6 hours of sleep in the last 77 hours - EDIT: make that 6 hours of sleep in the last 78 hours...).

First of all, "time delay" and "phase" are two very different things...though they can sorta be related and that's why it's confusing.

The best way to think about phase is to compare the input and output of a "system" (which can be things like crossovers, amplifiers, speakers, etc etc...). The easiest example is a system that introduces a 180 degree phase shift: when a positive signal arrives on the input, you get a negative signal on the output. The frequency and amplitude remain the same and the signal still arrives at the same time (electricity is pretty much instant). But what happens is the output "decides" to start at a different point on the sine wave relative to the input.

So in the case of a subwoofer, you have the same input signal going to your mains and your subwoofer. Your mains introduce a phase shift of there own, but when playing alone it is completely inaudible and doesn't matter at all. But when you add a subwoofer, it is introducing a different phase shift of its own. And now that there are two sources of sound, the phase affects whether or not the two signals will add or subtract. We want the signals to add up - and in order to accomplish it we need to align the phase of the subwoofer to that of the mains. Turning the phase dial on the sub is merely changing "where on the sine wave" the driver on the subwoofer starts (ie: does it start moving out or in?). There is no delay involved.

Ok, now for time-delay...don't forget that electricity is pretty much instant and for the sake of discussion we will ignore the group delay of the drivers (the time it takes the electrical signal to get the driver started moving). All that to say...the only 'real' source of time-delay is the physical placement of the speakers relative to each other. It's pretty much intuitive...the sound from the closer speaker is going to arrive earlier.

Take one speaker (speaker A), place it 12 feet away from your listening position and measure the phase of a constant 90Hz signal. Now move the speaker to a distance of 18 feet and measure the phase of 90Hz again. You will find that the sound from both situations arrives with the same phase (no matter how far away the driver is). This makes sense because the pressure wave created by the driver is merely gliding through the air (If you could sit on the wave and travel with it you will notice that it doesn't go up and down).

Now get another speaker (speaker identical to speaker A, and place both A and B a distance of 12 feet away from your listening position. Play a constant 90Hz signal and measure the phase...it will be the same. Leaving speaker B alone, move speaker A to a distance of 18 feet away and play a constant 90Hz tone. In the ideal world you won't hear any sound because of "phase cancellation" (which was introduced because of time-delay). The wavelength of 90Hz is 12 feet, so when you shift one of the waves by 6 feet you end up shifting the relative phase by 180 degrees (which we know causes perfect cancellation). The most important thing to realize here is that you are playing a constant 90Hz signal...

Sound takes about 1ms to travel 1 foot...so by moving speaker A back 6 feet, we are essentially delaying it by 6ms. Let's do the same experiment (speaker B at 12 feet and speaker A at 18 feet), but this time play a 6ms burst at 90Hz. The sound from speaker B arrives first and plays for 6ms. Right when the sound from speaker B stops, the sound from speaker A arrives and lasts for 6ms. There is no phase cancellation because there is only one sound source arriving at a time. So as long as the listener isn't hearing 90Hz from both A and B at the same time, the listener doesn't hear any phase cancellation. In the case of the constant 90Hz tone....there is 6ms where it is heard plainly from B and then there is sound from A and B at the same time (cancelling out) until eventually the signal is turned off....and then there will be 6ms of 90Hz coming from A (because the speaker is further away).

Now let's say we shift the phase of A by 180 degrees, place it at 18 feet and place B at 12 feet (same scenario, but A has a phase shift). When we play a cosntant 90Hz tone, we end up getting phase addition (+6dB). But again, think of things in the time domain....for the first 6ms the signal is at +0dB (no addition or subtraction)...then while both A and B are playing 90Hz there is +6dB attenuation and then when the signal stops it will drop to +0dB again before it turns completely off. So if both A and B are playing 90Hz at 90dB for 12ms, you get 90dB for the first 6ms, 96dB for the next 6ms, and then 90dB for the last 6ms.

So just to stress the point....speaker placement has no affect on phase. The reason there are phase shifts in speakers is because a single size driver is playing multiple frequencies (for there to be no phase shift, the driver would have to change size as it played different sounds....not exactly a feasible approach). Speaker placement however does affect time-delay and hopefully I have painted a picture of how the time domain can affect how the waves interact with each other. In order to wrap your head around the concept you have to stop thinking about constant tones in the frequency domain and start thinking about tone burts in the time domain.

Well I hope all this made sense...though it seems like way too many words for such a non-complex concept. I might have to end up making some pictures...or ideally some animations, but them things are hard to make.

[Buckeye_Nut]

Roughly speaking as a rule of thumb........

Providing the speaker distances from your primary seating area are input correctly into your

receiver, your default phase setting will be zero. Your receiver

will automatically put your speakers into phase which is partly why

inputting correct 'speaker distance' is so important.

[CAS]

Ok, let's see if I can do this without pictures (bare with me because I've had 6 hours of sleep in the last 77 hours - EDIT: make that 6 hours of sleep in the last 78 hours...).

First of all, "time delay" and "phase" are two very different things...though they can sorta be related and that's why it's confusing.

The best way to think about phase is to compare the input and output of a "system" (which can be things like crossovers, amplifiers, speakers, etc etc...). The easiest example is a system that introduces a 180 degree phase shift: when a positive signal arrives on the input, you get a negative signal on the output. The frequency and amplitude remain the same and the signal still arrives at the same time (electricity is pretty much instant). But what happens is the output "decides" to start at a different point on the sine wave relative to the input.

So in the case of a subwoofer, you have the same input signal going to your mains and your subwoofer. Your mains introduce a phase shift of there own, but when playing alone it is completely inaudible and doesn't matter at all. But when you add a subwoofer, it is introducing a different phase shift of its own. And now that there are two sources of sound, the phase affects whether or not the two signals will add or subtract. We want the signals to add up - and in order to accomplish it we need to align the phase of the subwoofer to that of the mains. Turning the phase dial on the sub is merely changing "where on the sine wave" the driver on the subwoofer starts (ie: does it start moving out or in?). There is no delay involved.

Ok, now for time-delay...don't forget that electricity is pretty much instant and for the sake of discussion we will ignore the group delay of the drivers (the time it takes the electrical signal to get the driver started moving). All that to say...the only 'real' source of time-delay is the physical placement of the speakers relative to each other. It's pretty much intuitive...the sound from the closer speaker is going to arrive earlier.

Take one speaker (speaker A), place it 12 feet away from your listening position and measure the phase of a constant 90Hz signal. Now move the speaker to a distance of 18 feet and measure the phase of 90Hz again. You will find that the sound from both situations arrives with the same phase (no matter how far away the driver is). This makes sense because the pressure wave created by the driver is merely gliding through the air (If you could sit on the wave and travel with it you will notice that it doesn't go up and down).

Now get another speaker (speaker identical to speaker A, and place both A and B a distance of 12 feet away from your listening position. Play a constant 90Hz signal and measure the phase...it will be the same. Leaving speaker B alone, move speaker A to a distance of 18 feet away and play a constant 90Hz tone. In the ideal world you won't hear any sound because of "phase cancellation" (which was introduced because of time-delay). The wavelength of 90Hz is 12 feet, so when you shift one of the waves by 6 feet you end up shifting the relative phase by 180 degrees (which we know causes perfect cancellation). The most important thing to realize here is that you are playing a constant 90Hz signal...

Sound takes about 1ms to travel 1 foot...so by moving speaker A back 6 feet, we are essentially delaying it by 6ms. Let's do the same experiment (speaker B at 12 feet and speaker A at 18 feet), but this time play a 6ms burst at 90Hz. The sound from speaker B arrives first and plays for 6ms. Right when the sound from speaker B stops, the sound from speaker A arrives and lasts for 6ms. There is no phase cancellation because there is only one sound source arriving at a time. So as long as the listener isn't hearing 90Hz from both A and B at the same time, the listener doesn't hear any phase cancellation. In the case of the constant 90Hz tone....there is 6ms where it is heard plainly from B and then there is sound from A and B at the same time (cancelling out) until eventually the signal is turned off....and then there will be 6ms of 90Hz coming from A (because the speaker is further away).

Now let's say we shift the phase of A by 180 degrees, place it at 18 feet and place B at 12 feet (same scenario, but A has a phase shift). When we play a cosntant 90Hz tone, we end up getting phase addition (+6dB). But again, think of things in the time domain....for the first 6ms the signal is at +0dB (no addition or subtraction)...then while both A and B are playing 90Hz there is +6dB attenuation and then when the signal stops it will drop to +0dB again before it turns completely off. So if both A and B are playing 90Hz at 90dB for 12ms, you get 90dB for the first 6ms, 96dB for the next 6ms, and then 90dB for the last 6ms.

So just to stress the point....speaker placement has no affect on phase. The reason there are phase shifts in speakers is because a single size driver is playing multiple frequencies (for there to be no phase shift, the driver would have to change size as it played different sounds....not exactly a feasible approach). Speaker placement however does affect time-delay and hopefully I have painted a picture of how the time domain can affect how the waves interact with each other. In order to wrap your head around the concept you have to stop thinking about constant tones in the frequency domain and start thinking about tone burts in the time domain.

Well I hope all this made sense...though it seems like way too many words for such a non-complex concept. I might have to end up making some pictures...or ideally some animations, but them things are hard to make.

What he said

[DrWho]

Roughly speaking as a rule of thumb........

Providing the speaker distances from your primary seating area are input correctly into your

receiver, your default phase setting will be zero. Your receiver

will automatically put your speakers into phase which is partly why

inputting correct 'speaker distance' is so important.

Except for the fact that all speakers start to shift in phase in the

lower ranges of their passband...You'll find that most large

floorstanding speakers are about 90 degrees out of phase in the 60-80Hz

region...and it's a smooth climb to 180 degrees at the tuning point.

Subwoofers on the other hand are usually 0 degrees out of phase at the

top of their passband (which is the 80-200Hz region). But just like all

other speakers, they slowly climb to a 180 degree shift at the tuning

point.

In sealed cabinets you are talking about 2/3 of the phase shift when

compared to that of ported cabinets....so mains are like 60Hz out of

phase in the 80Hz region and then 120degress out of phase at what would

be the tuning point.

Anyways, all that to say two things:

First, with a 90 degree phase shift on the mains - a 180 degree phase

switch isn't going to do much of anything (as a 90 degree phase shift

either direction is essentially identical). I'm actually quite

surprised that there isn't mention more often of these switches not

having any audible affect. Gotta love the placebo effect.

Second, adjusting the distance settings in the reciever is still very

important and in a flat phase system it would result in a perfect phase

response. But to achieve a flat phase system would require the diameter

of the drivers to change based on the frequency they're playing...so

when a driver plays multiple frequencies at once it somehow needs to be

two different sizes at the same time []

And in all honesty - it's a bit of a waste of time to worry about the

phase so much. Just shift your listening position by 3 feet and it will

require huge changes in the setup. Sure, you can dial it in for one

position but most movies are watched with multiple people and therefore

a 180 degree switch is usually more than enough (it's there to account

for driver polarity differences).

[Buckeye_Nut]

Interestingly enough, I got that little 'rule of thumb' straight from

the Avia head honcho in another forum when I was googling on the

subject. You were talking about something entirely different. I was

merely offering a basic rule of thumb while not attempting to explain

the finer points of flattening frequency response. That

discussion belongs in a thread discussing accoustical room treatments

and the finer points of eQ'ing with Avia[]

However, I do agree with what you were saying.

[sivadselim]

Buckeye_nut, weren't you involved in a thread either here or on another forum that discussed why the subwoofer distance setting is often set to a seemingly "incorrect" value by auto-calibrating receivers?

[Deang]

http://www.libinst.com/tpfd.htm