Phase lesson (ChrisA? since I am quoting you)

[Coytee]

Chris, in the LaScala thread you said:

 

"Also note that any passive or active crossover will insert from 90-360 degrees of phase lag of the lower frequency driver relative to the higher frequency driver (90 degrees per "order" of the filter used--for a forth order filter, that's 360 degrees lag of the LF driver relative to the HF driver). "

 

If you or anyone can explain to my 8 year old mentality....  what does it mean to be 90 degrees out of phase?

 

What would happen to the sound? 

 

I can understand what a plate rotated 90 degrees might look like but when you put it into the world of phase I'm lost.

 

If I have some sound at phase "0" and the next incremental phase is turned 90 degrees, what happens?

 

Would 180 phase cancel the first '0' phase?

 

[Travis In Austin]

360 would cancel the 0 phase.

If you want to hear 180 out of phase, reverse the + and - wires on the back of one of your speakers in a two speaker system.

90 out of phase creates time alignment issues.

http://www.prosoundweb.com/article/polarity_and_phase_explained/

[Travis In Austin]

So the 90 degree shift per order is true with alo passive filters?

How is this accounted for/ corrected with a properly designed speaker?

[CECAA850]

How is this accounted for/ corrected with a properly designed speaker?

 

Just spewing thoughts, not facts here. 

 

Couldn't you  compensate for that by moving the drivers forward or rearward in relation to each other in the cabinet?

[Marvel]

Yes, but it's a bit more difficult than that. A steeper slope would help in that regard, i believe.

That's always been a problem with passives. Phase and polarity aren't the same,

[Travis In Austin]

I think it is a great question posed by Richard.

So keeping it simple, two way speaker. You could go 4th order. 360 phase, so no phase problem? Could you go 2nd order, 180 out of phase, but reverse polarity on either the woofer or tweeter to get them back in phase?

[Coytee]

If you go 2-way and they are 180 out of phase, does that mean the initial transient of say the woofer bin is on the outward pressure stroke and the tweeter would be on the inward 'negative' stroke?

 

(if true then...)

 

So if you have a bass kick drum you would hear/feel the transient impulse from it....  and if on the tweeter, you had a cymbol playing, you would hear it as well

 

BUT

 

If you had a higher frequency instrument that crossed the threshold of the crossover point, then you'd have part of the sound pushing and the other half pulling?

 

(if true then...)

 

So, if you have something 90 degrees out of phase, you'd have the (same as above) woofer pushing the sound and where it crossed the threshold of the crossover, it would have already begun the pull phase and it would pull only have say, 50% of the signal to pull?

 

See what I mean when I say you're talking to an 8 year old!?!?!  (wife would only be willing to give me credit to six) 

[Travis In Austin]

If you go 2-way and they are 180 out of phase, does that mean the initial transient of say the woofer bin is on the outward pressure stroke and the tweeter would be on the inward 'negative' stroke?

(if true then...)

So if you have a bass kick drum you would hear/feel the transient impulse from it.... and if on the tweeter, you had a cymbol playing, you would hear it as well

BUT

If you had a higher frequency instrument that crossed the threshold of the crossover point, then you'd have part of the sound pushing and the other half pulling?

(if true then...)

So, if you have something 90 degrees out of phase, you'd have the (same as above) woofer pushing the sound and where it crossed the threshold of the crossover, it would have already begun the pull phase and it would pull only have say, 50% of the signal to pull?

See what I mean when I say you're talking to an 8 year old!?!?! (wife would only be willing to give me credit to six)

No you would have mud.

That your 2 way jubes, reverse just one of the tweeter connections, listen, switch, listen. The sound stage will snap into focus.

Do the same with the woofer, you should notice a bigger difference, depending on the song.

What active crossover are you running these days? Some will allow you to change phase in 90 degree increments.

Roy can tell if the left channel is in phase with the right just by steping back and listening. I believe he does that by noticing a lack of bass response. If left is out of phase with right, they are canceling each other out to a degree, which is perceived as either impaired bass extension, or lack of center sound stage with stereo.

Of course, I have just the most basic understanding of this stuff from playing around mixing boards where you can shift phase

[Ski Bum]

So keeping it simple, two way speaker. You could go 4th order. 360 phase, so no phase problem? Could you go 2nd order, 180 out of phase, but reverse polarity on either the woofer or tweeter to get them back in phase?

 

Dunning Kruger Alert!  I may not know what I'm talking about here, so fair warning.  

 

I would answer, kinda, sorta, yes.  There are additional factors to consider, such as using a -6db crossover point with 12db/oct slopes, and the drivers must be wired out of phase.  Compare to a -3db point for a fourth order, 24db/oct slope style crossover, where you must have the drivers wired in phase.  Both methods sum flat in amplitude response.   As long as you choose slopes with phase shifts with factors of 180, as occurs using the 12/12 slope crossover, then yes, switching driver polarity can ensure flat response at the crossover frequency, and phase of 0 or 360.

[Coytee]

That your 2 way jubes, reverse just one of the tweeter connections, listen, switch, listen. The sound stage will snap into focus.

 

For simplicity, let's keep this at a single speaker.

 

My contention would be that if the two halves are out of phase, you would hear less of an issue in mono than stereo.

 

Remember....  8 year old.  What happens with a single speaker, then I'll try to wrap my pea brain with stereo.

 

I'm using the DX38.  I don't know that it changes phase, but I do think it changes polarity (unless that is the same thing)

[Travis In Austin]

I forgot to add one thing. all of this assumes that you are listening to a recording that was recorded in phase, many are not.

Some preamps have phase shifting for that very reason (Nagra for example).

So buy a pink nouse generator to test some of this stuff out.

[WMcD]

I'm waiting for Chris to come in and make everything clear. We need lots of diagrams.

But I'll start.

1) Phase is just about always a relative thing. It is a relation between two signals. But which two?

2) If we look at a filter, we generally look at the relation between the input and the output. We start at zero Hertz or a very low frequency and compare input and output. They may be out of phase. But then we increase frequency and see what happens to the output. Therefore we're looking at how the output changes in phase, or zero crossings of the sine.

The greatest change of phase relation occurs where the magnitude of the output is changing. Therefore, the phase difference is intimately related to the slope or roll up or roll down. The more steep the filter, the greater the change of phase.

3) PWK pointed out that we can't hear phase of a SINGLE source. He pointed out that as we move our ear a foot or two, the phase (because of the different delay relative to whatever is coming out the speaker) shifts very much.

4) This means that if the magnitude of the source is rolling up or down, we just hear that change in magnitude or intensity, not the phase change inherent in the filter.

5) So what's the problem?

6) The problem comes in when we are "mixing" two sources of the sine wave at, the same frequency. We know that even if the same overall shape (frequency) there is an oddball adding. If one is going up while the other is going down they are 180 degrees out of phase. If this happens as they reach our ear, we hear nothing (subtraction). Of course if they are both in phase, the add.

7) How do we describe this? One thing is to examine each output of the two relative to its individual phase relative to each's very low frequency input. That is the benchmark of each's output. But by these benchmarks, we can calculate their relative phase to each other. And whether they add or subtract or something in between.

8) Hopefully, just looking at filters, we can find a situation where the filter rolling off the woofer and the one rolling up the tweeter all play nice so that the magnitudes and the phase add up to something smooth, so they blend. This is a surprising difficult situation to achieve.

9) There is a big joker 1 in this pack. It is that the mechanics in the driver will create frequency roll up and roll down. Remember that the roll up and roll down of magnitude cause phase changes. Therefore, each speaker without an external electrical filter is its own filter and phase changer. If you look at the frequency response of horns, they don't "roll off," they drop like a rock and there are radical phase changes even before filters and time delay get involved.

10) The big joker 2 in the pack has to do with actual time delay. If you have two speakers which are identical but offset in distance, laterally, the phase of the sound reaching your ear are shifted in phase unless you are exactly on the triangle where they are equidistant.

11) With horns of different lengths (time delay), the problem in 10) is even more cranky.

Has this confused things?

WMcD

WMcD

[Chris A]

I'm waiting for Chris to come in and make everything clear. We need lots of diagrams.

Thanks for the buildup, Gil. 

 

 

 

...what does it mean to be 90 degrees out of phase? What would happen to the sound? I can understand what a plate rotated 90 degrees might look like but when you put it into the world of phase I'm lost. If I have some sound at phase "0" and the next incremental phase is turned 90 degrees, what happens? Would 180 phase cancel the first '0' phase?

If you look at the animated picture, above, you will see two signals on the top, the second of which is continuously shifting in phase relative to the first signal, but the frequency of the two signals is the same (same "pitch" of sound, for instance). 

 

The third trace is the important one - its the sum of the first two traces.  You can think of this like taking a tweeter and moving it slowly toward your ears, while the midrange is stationary, all while playing a sine wave on both drivers at a frequency which is at the center of the crossover frequencies between to the two drivers. 

 

Let's say a KHorn crosses at 5 kHz between the tweeter and midrange, and we're playing a 5 kHz tone (a pure tone or sine wave) and listening to the two drivers at our listening position.  If we slowly start moving the tweeter toward or away from our ears, what we hear will be found in the third signal trace, above--the interference sum of the two tones: one from the tweeter, one from the midrange driver.  It will sound loud, then quiet, then loud, etc., as the tweeter is moved continuously toward or away from our listening position, due to additions and cancellations of the two signals at our ears.  If we stop the movement of the tweeter, it's just like stopping the animation.

 

You know that "360 degrees" is equal to one full wavelength or cycle before everything repeats.  Ninety degrees is 1/4 of the cycle, 180 degrees is half of the cycle (which is full cancellation of the sine wave at our ears if we are sitting in an anechoic chamber, to answer your question), 270 degrees is 3/4 of the cycle, and at 360 degrees, everything repeats again in phase.  Every place where we move the tweeter which is 1/4 or 3/4 of the wavelength (5 kHz in our example), the sound pressure level (SPL) will be attenuated by half at our listening position (-6 dB in relative amplitude level).

 

Chris

Edited November 3, 2015 by Chris A

[Schu]

So there would be a fluttering sound on a pure signal similar to when you play two signals and slowly begin to change the pitch of one of them?

The pitches begin to flutter or worble the futher you get away from that original pitch...

Edited November 3, 2015 by Schu

[Chris A]

"Flutter" implies something that is happening relatively fast, such as moving the tweeter at a higher speed toward or away from our listening position.

 

What I described in my example, above, was a slowly moving tweeter.  The sound for a slowly moving tweeter would sound just like the picture shows in the third signal: slow ramp-ups and downs of the SPL at 5 kHz. 

 

It actually has nothing to do with distance other than one sound source will be closer to your ears than the other, so the cancellation effect at 180 degrees relative phase won't exactly cancel the 5kHz tone, in the example above.  If you move the tweeter fast enough, you will start hearing a Doppler shift - like a train blowing a whistle coming or going.  This is due to the compression or expansion of the sound waves from the coming or going sound source and will introduce "beats" between the two signals (midrange, tweeter) which are now at slightly different frequencies at our listening position.  This is also known as frequency modulation (FM).

 

Chris

Edited November 3, 2015 by Chris A

[JL Sargent]

Ironically, shifting in and out of phase is also how Marshall's hypnotic frog puts you under.

[Ski Bum]

Chris, this phase relation between two drivers oriented vertically in a single speaker as coytee is asking about would result in different vertical lobing patterns in the speakers power response, depending on the slopes involved and physical distance between the drivers, correct?  

[Chris A]

Yes.  There is a Wikipedia page on this subject. 

 

There is a good article that Rane has on the web about the general subject of time/phase alignment and crossover filters. Also, Wayne Parham at Pi loudspeakers has this .pdf file (see pg.13),

 

 

 

Chris

Edited November 3, 2015 by Chris A

[Chris A]

Couldn't you compensate for that by moving the drivers forward or rearward in relation to each other in the cabinet?

 

Yes.

[Ski Bum]

10) The big joker 2 in the pack has to do with actual time delay. If you have two speakers which are identical but offset in distance, laterally, the phase of the sound reaching your ear are shifted in phase unless you are exactly on the triangle where they are equidistant.

 

 

That can be compensated for, at least to some extent.  This is more room/placement and psychoacoustics here, so pardon the digression.  Taking advantage of time/intensity trading via heavy toe in can help expand the sound stage outside of positions equidistant from the L and R speaker.  It doesn't fix the phase/delay, but given how our perceptual filters weave together the illusion of stereophonic sound, it seems to work surprisingly well.