Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[John Roberts {JR}]

I have always regarded the term "absolute polarity" to mean that the sound coming out of the speakers in in the same polarity as the sound when it was recorded.

+1 and carrying that further, the signal recorded should be same absolute polarity as the acoustic waveform coming from say a kick drum. Most sources are not that simple or obvious.

Absolute polarity has merit in recording for archival purposes. The audibility of it for most complex waveforms is weak to un-proveable with any statistical significance.

Ie-not flipping the polarity on the WHOLE signal somewhere in the chain.

The recording industry only accepted this a thing in the '80s.

The people who insist that all the DRIVERS be in polarity, have obviously never looked at what happens when you run that driver through high and low pass filters of different slopes.

Multiple drivers in loudspeakers are a melange of compromises on top of compromises.

Crossover filters must bandpass the signal sent to individual drivers to protect those drivers from out of bandpass signal energy.

Further adjacent driver HP/LP filters must play nice together during the transition region when both drivers are active. When both drivers are making similar amplitude levels, deep notches can occur from phase cancellation.  Of course this is only looking at one dimension of how they combine. Sound fields are 3D so this summation in 3D space is complex.   

It is the FINAL result of how the PHASE of adjacent bands works together, not the POLARITY of the individual drivers that give the CORRECT sound.

Back in the day with 2 pole crossovers it was routine to reverse polarity of midrange drivers to prevent a suck-out from 180' signals combining in the transition region. 3-pole crossovers were an improvement being 90' out wrt each other, for modest bumps, but no deep cancellation. 

But as Ivan shared the speaker design matters. I recall back at Peavey, we made a simple crossover plug-in module for CS power amps, and I asked transducer engineers for proper settings (frequency and polarity) for sundry Peavey speakers. A speaker engineer that I respect declined, explaining that moving the crossover tuning slightly can impact the optimal polarity for adjacent drivers.

These days the crossover technology is superior to what we had to work with last century, but guess what. The classic compromises mostly still remain. One reason I am enthusiastic about Danley's novel tapped horn technology.

JR

[Jay Barracato]

I have always regarded the term "absolute polarity" to mean that the sound coming out of the speakers in in the same polarity as the sound when it was recorded.

Ie-not flipping the polarity on the WHOLE signal somewhere in the chain.

The people who insist that all the DRIVERS be in polarity, have obviously never looked at what happens when you run that driver through high and low pass filters of different slopes.

It is the FINAL result of how the PHASE of adjacent bands works together, not the POLARITY of the individual drivers that give the CORRECT sound.

And I think we would do the general industry a service if people would start referring to "change in phase" rather than the more generic "phase".

When teaching thermodynamics I harp on the difference between absolute energy and relative energy. Even scientists are clumsy about the he terminology and you really have to look at the context to figure out which is meant.

And to add to what Tom was saying about the impedance, I really think of that as a result of the two mechanisms of fluid mechanics at work here, the driver doing work on the air and then the wave propagation. Too often I think techs jump to the wave explanation when it really is a direct vibration and the cohesion of air. Why does that 18 inch driver move air a significant distance behind the driver before the wave even propagates?

Remember the black hole sum syn aud con discussions?

What happens to the impedance of a pair of speakers when one is pushing on a slug of air at the same time the other is pulling?

I don't think it can be explained in terms of waves if no wave is formed.

Sent from my XT1254 using Tapatalk

[Ivan Beaver]

A speaker engineer that I respect declined, explaining that moving the crossover tuning slightly can impact the optimal polarity for adjacent drivers.

And most people don't realize that when you change the xover freq, you also change the alignment delay needed.

And if changed enough-a polarity reversal may be needed and produce better results.

Despite what people want-there is no simple answer.  Only a possible wrong answer that they "think" they understand.

[Jerome Malsack]

I see JBL has a JBL tool app on the app store ?   loading to play with and see what the results will be.

[Mark Wilkinson]

+1 and carrying that further, the signal recorded should be same absolute polarity as the acoustic waveform coming from say a kick drum. Most sources are not that simple or obvious.

Absolute polarity has merit in recording for archival purposes. The audibility of it for most complex waveforms is weak to un-proveable with any statistical significance. The recording industry only accepted this a thing in the '80s. Multiple drivers in loudspeakers are a melange of compromises on top of compromises.

Crossover filters must bandpass the signal sent to individual drivers to protect those drivers from out of bandpass signal energy.

Further adjacent driver HP/LP filters must play nice together during the transition region when both drivers are active. When both drivers are making similar amplitude levels, deep notches can occur from phase cancellation.  Of course this is only looking at one dimension of how they combine. Sound fields are 3D so this summation in 3D space is complex.   Back in the day with 2 pole crossovers it was routine to reverse polarity of midrange drivers to prevent a suck-out from 180' signals combining in the transition region. 3-pole crossovers were an improvement being 90' out wrt each other, for modest bumps, but no deep cancellation. 

But as Ivan shared the speaker design matters. I recall back at Peavey, we made a simple crossover plug-in module for CS power amps, and I asked transducer engineers for proper settings (frequency and polarity) for sundry Peavey speakers. A speaker engineer that I respect declined, explaining that moving the crossover tuning slightly can impact the optimal polarity for adjacent drivers.

These days the crossover technology is superior to what we had to work with last century, but guess what. The classic compromises mostly still remain. One reason I am enthusiastic about Danley's novel tapped horn technology.

JR

Hey JR, Ivan, all

Let's say you have a coaxial compression driver, and raw impulse response shows normal polarity on the HF section and inverted on the VHF. 
Let's say you don't correct the inversion and you use non-inverting filters both sides of x-over, ie LR 24 or 48.
....and you've gotten phase slopes to match, along with timing ...

Then, I'm thinking the polarity inversion should carry through...and give the expected magnitude notch at x-over..

Instead, I just got great level summation with the above...

I had used FIR to flatten phase on each section @ near zero degrees.  I screwed up and flattened the VHF section, before inverting polarity.
Magnitude summation didn't give a shit  about the polarity inversion, after FIR phase correction;
but impulse response, post FIR, still shows inversion..

Intuitively this seems wrong  
Is this possible in the analog world?

[John Roberts {JR}]

I am not a speaker expert, but afaik the phase issues are mostly problems in the transition region where both speakers are making similar amplitude sounds, that can interfere constructively or destructively.

If different frequency components of a complex sound are wholly in two different bandpasses but not the same polarity, I don't expect them to interfere. 

Of course better is always better, but don't blow the transition region combining for something that may not be very audible, if at all.

But I repeat the caveat that I am not the big dog speaker guy here,,, I still prefer powered cabinets where actual speaker engineers make these decisions.   

JR

[Ivan Beaver]

Hey JR, Ivan, all

Let's say you have a coaxial compression driver, and raw impulse response shows normal polarity on the HF section and inverted on the VHF. 
Let's say you don't correct the inversion and you use non-inverting filters both sides of x-over, ie LR 24 or 48.
....and you've gotten phase slopes to match, along with timing ...

Then, I'm thinking the polarity inversion should carry through...and give the expected magnitude notch at x-over..

Instead, I just got great level summation with the above...

I had used FIR to flatten phase on each section @ near zero degrees.  I screwed up and flattened the VHF section, before inverting polarity.
Magnitude summation didn't give a shit  about the polarity inversion, after FIR phase correction;
but impulse response, post FIR, still shows inversion..

Intuitively this seems wrong  
Is this possible in the analog world?

It does not matter what the polarity of the drivers is.

As long as the phase doesn't have any sharp shifts and the magnitude is reasonably flat, then all is good.

[ProSoundWeb Community]