Polarity of speaker drivers


Many speakers do NOT have all the drivers aligned in the same polarity.  This is seen on many of the stereophile speaker measurements.  In certain designs,  I guess this is done for better summing of driver output.  Is the time domain compromise audible?
128x128glai
glai,

Not by 90% of listeners, no. There are some who claim to be able to hear this difference, and I cannot prove if they can or cannot. All I can say is I most definitely cannot.

What all listeners could hear is if the drivers are not matched properly. Take a look at the second graph on this page:

http://speakermakersjourney.blogspot.com/2016/05/lm-1-bookshelf-measurements.html

You'll see a severe dip that happens when the tweeter is inverted from normal. It's quite audible.


Best,

Erik
What is more important is that BOTH speakers are wired with the same polarity.  You would be surprised at how many speakers are shipped with mismatch in polarity.
@glai This is potentially a huge issue. Yes, MANY designs do this and some argue that it's because the result is potentially better measurements which in turn look better in magazine reviews, and result in higher sales. Others would say that the most common configuration, midrange reversed from tweeter and woofer in a 3way, results in increased air & ambience. 
I am in the camp that thinks it's great to test it both ways and trust your ears. More often than not, I prefer the sound with all drivers in polarity. Of course, this is easier to test on some speakers more than others. For example, the Nolas and Alons that I've owned included tri-wiring so it was simple to just swap leads for the reversed driver. Doing so on a pair of 5 ways with internal crossover is a huge PITA! 
A good friend of mine and audio tech carries a "Cricket" polarity tester with test CD everywhere he goes. I've seen him shock & delight over a dozen system owners by getting their drivers in polarity. 
@brf makes a good point too. QA needs to be held to a higher standard. Cheers,
Spencer
Some speaker companies (like Vandersteen and the old Thiel) stake their design philosophy on the premise that these time domain issues are crucial.  Listen to those designs yourself and see what you think.  I'm a Vandersteen guy, so you know where I stand.
Full range headphones, using a single driver without a crossover for all frequencies, can tell us what a loudspeaker should sound like, at least in amplitude and temporal terms. So do some planar speakers, such as the Eminent Technology LFT-8b, which has a single driver reproducing 180Hz to 10kHz, without a crossover in that range. The x/o at 180Hz to a dynamic woofer, and at 10kHz to a ribbon tweeter, are both 1st order-6dB/octave, just as in Vandersteen designs.
@sbank is right.

However, that not every speaker maker has adopted time and phase matched designs should tell you something about how important it seems. If time coherence was that revolutionary no one would be able to sell any other kind of speaker.  However this is not the case.

Make up your own mind, of course and buy what you will enjoy.

I think maybe it's a mistake to learn to tell speaker polarity by ear. :) I mean, let's say you can learn this through experimentation and feedback. You go from 2016 when you can't tell speaker polarity, to sometime in 2017 when you can. Doesn't that just make your comfortable music world more limited?

Best,


Erik
Forgot to mention, you can experiment with full-range drivers, those that have no crossover, relatively cheaply. Madisound is selling a complete kit on sale. Some people get a couple of hits of that and never look back to multi-way speakers again.

glai, your basic assumption of why switching a drivers polarity is correct.
It has everything to do with phasing.... There is no speaker with a crossover that is completely phase coherent.... period. Even the most "Phase Coherent" speakers have phasing issues.
We normally talk about crossovers typically in Electrical slopes... we may call it 6db per octave or 12db per octave, but when you add the natural rolloff of a driver along with phasing, in the end a 6db slope may be 9 or 10 db per octave.... I often might use a 18db per octave electrical slope on a tweeter and a 12db per octave slope on a woofer...On my last pair pair of speakers,  the end result was a 24db per octave roll off and the crossover point being 6db down. This combination rolls phasing around to come back in alignment and in fact, quite often in closer alignment than using what most consider the only truly phase coherent 6db slopes. When you listen to a pair of speakers and the polarity is crossed on a speaker,  most folks do not say that the speakers polarity is wrong,  they say "The Speaker is out of Phase".... When you switch the polarity, phasing comes back in alignment.
Under Electrical crossover conditions, when you cross a tweeter and midrange at 12db per octave slopes, they are 180 degrees out of phase, the idea of switching the polarity of 1 driver is an attempt to improve phasing or bring the drivers back in phase with each other. Anytime the polarity is switched on a driver,  this should be the reasoning behind it. 
This is a very elementary explanation. I hope that it helps.  Tim
I put together a speaker kit, the LM-1, and posted the simulation files so you can play with crossover design yourself. This may help you get a better understanding of polarity and phase relationships.

https://forum.audiogon.com/discussions/try-your-hand-at-crossover-design-free

Best,


Erik
Having jumped into a two channel system head first I took the advice of some very experienced audio experts and went with what I felt were the speakers with the fewest crossover/phase issues that would deliver the audio signal across the entire frequency range. 
As Tim has stated, there are ways to bring drivers back into phase and the really good speaker manufacturers do this with aplomb, I'm sure. But you still have notches in critical areas of the frequency response as well as crossover inefficiency and IM distortion as the crossovers heat up. 
Active crossovers that sit outside the speaker and are placed between the preamp and the amps with the amps wired directly to the drivers had some appeal and serve as answers to the above problems, but all in all the simple full range driver with first order crossovers to the tweeters and built in subs had appeal in many facets. Coherence in a speaker is a pretty strong quality. 
JMO, but any move toward straight wire with gain is a plus. Full range drivers have their issues as well but phase/polarity isn't one of them. Actually the immediacy of the sound is startling at first, becomes endearing with time.
Well, I think that trying to divine what speakers will sound better or worse based on crossover complexity is very difficult indeed.

While I have my own preferences for drivers, and crossover components, I would always let my ears listen before worrying about the technology. Same with DAC chips. Implementation and total system synergy matters mroe to me than statistical evaluation of components.

My LM-1 Kit has a single cap and 1 resistor on the way to the tweeter by the way, with only 9 crossover components in total. The woofer has only 1 coil on series, and has fantastic phase integration between the two drivers.

Best,

Erik