Tweeter's inverted phase.

[Soandso]

I'm discussing speaker technology, crossovers in particular. We see a contradiction between direct tests of audibility of group delay (aka phase, inverted) and your speculations derived from quite specialized research that was done with a completely different focus in basic neurophysiology. And that research was done with decidedly inverted phase on the two ears, not the same as it would be w/ a stereo speaker setup. Your switch to ad hominem accusations was expected.

You can discuss whatever you want. Just do not insist on my validating you or, for that matter, accuse me of what you've been trying to slip in among walls of type.

 

[skankhobag]

Buy new speakers. According to your description, what you have sucks.

 

[Suono]

It doesn't matter if the recordings are manipulated. I might make my own recording and want everything to be in phase. Conceptually, a speaker should have all the transducers in phase. There are speakers that are all in phase

 

[Arindal]

Conceptually, a speaker should have all the transducers in phase.

Why, if humans cannot perceive the difference of absolute polarity, except from lower frequencies ones and experimental signals? Speaker design is always a bunch of compromises, and why should one pursue a goal which does not mean a thing.

There are speakers that are all in phase

Polarity and phase are different things! All drivers might be wired in the same polarity, that does not mean that acoustic phase is constant over the whole frequency range. In contrary, a majority of crossover filters are actually shifting the phase, with the popular 2nd order filters in theory having the exact outcome of 180deg acoustic phase shift in theory. That is exactly the reason why some manufacturers flip polarity of the tweeter or midrange to be in phase again.

 

[MIsha]

Dear friends, I enjoyed reading this wonderful discussion. Thank you, it was interesting.
Yes, the electrical phase can indeed change in the crossover. Yes, polarity changes and spatial shifts of the drivers are indeed compensation tools.
What happened to me is somewhat paradoxical. And I am well aware of the formal "incorrectness" of my finding. I am interested in a simple question: why does this sound so much better?
I admit there are nuances unknown to me that led to this paradox. In fact, this is the practical part of my question: what could I have done wrong? What can I do to restore the polarity and achieve even better sound?

 

[Suono]

Polarity and phase are different things! All drivers might be wired in the same polarity, that does not mean that acoustic phase is constant over the whole frequency range. In contrary, a majority of crossover filters are actually shifting the phase, with the popular 2nd order filters in theory having the exact outcome of 180deg acoustic phase shift in theory. That is exactly the reason why some manufacturers flip polarity of the tweeter or midrange to be in phase again.

Thanks for the explanation. I'm not a technician. The important thing is that the midrange and tweak transducers are in phase. Music doesn't have notes and harmonics in antiphase.

 

[Heinrich]

... What can I do to restore the polarity and achieve even better sound?

I appreciate how cool you are, rephrasing the original question.

As mentioned before, especially because the model of your speakers isn't known, a systematic approach starting from scratch is my recommendation.
- measure the thing at different vertical (!) and horizontal angles
- is a similar subjective impression achieved w/ gradually equalising a correctly connected tweeter to that hole in ampl/ freqency response of inverted version
- what if the hole is gradually e/q to flat
- how does another speaker pair w/ similar x/over frequency behave
- comparison near/farfield
- e/q unsuspected frequency ranges e/g around 1k (seems the correct one has a slight but broad droop there), is there any e/q that would bring the correct to what you want

It is a lot to do, and documentation (for yourself) is an additional task, but what else would you do?

Hint: to invert the tweeter changes the vertical disperion pattern a lot; all the 'energy' lost on-axis is radiated off-axis, bouncing off floor and ceiling, reaching your ears later (hence: phase distorted ;-) than w/ correct polarity. And so forth, too much to be explained from remote.

 

[Arindal]

The important thing is that the midrange and tweak transducers are in phase.

That is technically impossible, if there is a conventional crossover in the signal path. Even low-order filters or existing fullrange drivers will in practice shift the acoustical phase, not necessarily 180deg or more, but they cause shift. Please don't mix polarity and phase, they are different things!

The only practical measure to avoid this, is time/phase corrected digital filters, like FIR-based crossover networks. In theory, a very broadband fullrange driver for mids and highs would do, with phase shift solely well outside its audible range, but I am not aware of such driver. Maybe Manger comes close to that.

 

[Suono]

Inter-driver phase deviation: 3 degrees within all cross bands

By the way, what does this mean? Can you explain it to me please? Are these XO circuits unknown?

 

[Suono]

I guess the above creates this
Group delay distortion: <3 milliseconds 100 Hz to 16 Hz. < 1 millisecond 100 Hz to 20 Khz

 

[Heinrich]

I guess the above creates this
Group delay distortion: <3 milliseconds 100 Hz to 16 Hz. < 1 millisecond 100 Hz to 20 Khz

In this thread, there were several hints suggesting that inverting the tweeter's connection is sometimed necessary to compensate for a previous inversion in certain crossover topologies. These hints were not picked up eagerly. A DIYer is very familiar with the what and why, though. Comes second after understanding Thiele/Small parameters.

Why should a consumer who has no persistet interest in the details care about it? It takes quite a bit of foundational knowledge to really understand the connections. Why waste time on that? What really matters is what modern technology provides.

And yes--around 1 ms of group delay or less is sufficient to maintain musical enjoyment; it still sounds simply "original", no differences heard even in quite focussed experiments. "Trust the science" one may say, direct evidence, there is no problem, isn't that a good thing?

You know, randomly altering a loudspeaker's design in a way that creates a deep notch in the frequency response--that's not a subtle phase effect. (That's what the o/p was about.)
How can anyone simply ignore such a dip and then go on to claim the tiniest effects elsewhere? I just don't get it ...

Just for curiosity, but what are you after? Want to get into designing speakers, as an engineer?

 

[Suono]

I try to understand how to correlate perceived music reproduction quality to the speaker specifications and design philosophies.

 

[Arindal]

I try to understand how to correlate perceived music reproduction quality to the speaker specifications and design philosophies.

My advice: Forget phase shift, group delay <1ms and polarity for all frequencies >200Hz, as the measurements (including step response) for a layman don't correlate with anything. Not saying that they are irrelevant, but one IMHO needs to be a speaker designer or experienced lab engineer to extract useful information from them.

 

[dualazmak]

My advice: Forget phase shift, group delay <1ms and polarity for all frequencies >200Hz, as the measurements (including step response) for a layman don't correlate with anything. Not saying that they are irrelevant, but one IMHO needs to be a speaker designer or experienced lab engineer to extract useful information from them.

Nevertheless,,,
A new thread has been (again) already started "Can anyone help with Alignment?"
where I just briefly participated by my post #53...

Spoiler

There in my post #53, I wrote;

As for time alignment among multiple SP-drivers in DSP based multichannel multi-SP-driver multi-amplifier active audio system...

In my case, I evaluated, validated (important!), measured, and applied;

1. Primitive but reliable reproducible "Time-shifted multiple-Fq-tone-bust method";
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-1_ Precision pulse wave matching method: #493

2. "Multiple-Fq-simultaneous-tone-burst energy-peak matching method";
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-2_ Energy peak matching method: #494

3. "Single sine-wave matching method";
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-3_ Precision single sine wave matching method in 0.1 msec accuracy: #504, #507

In case if you would be seriously interested in applying these methods in your setup, I would be more than happy sharing the test-tone signals I prepared and applied; please simply PM contact me writing your wish.

 

[Heinrich]

I try to understand how to correlate perceived music reproduction quality to the speaker specifications and design philosophies.

O/k ... the o/p might have his answers, the topic got a bit engaged at times, so a personal anecdote is in order.

I once ordered a set of speakers after decades of DIY. Those were deemed nearly perfect, price was a no brainer. I added a set of humble subs, no prob.
Thing is, now on top of the hill as good as it gets I care about room integration in bass, and then after, all is there. My hearing is the limit, but more so my mind. Do I like the music? Rarified my listening sessions, don't like to listen to music that often anymore. Bad, good, who knows. Is it the stereo's flaws - yes and no. I still like to treat a real piano myself, talking to musicians, going to new style concerts (I dismiss the classics). The stereo, after objective perfection, reveals all the limitations that recorded music inherently bears: arbitrary repeatability for one.

That's the effect of a perfect stereo. Once it's done, the curtain is withdrawn, there you are with it.

 

[MIsha]

Dear Heinrich.
Thanks for the recommendations. I’ll try to EQ the dip, but I suspect it’s a “black hole.”

And finally, a huge and sincere thank-you for the “personal anecdote” about your perfect audio system.

 

[Suono]

O/k ... the o/p might have his answers, the topic got a bit engaged at times, so a personal anecdote is in order.

I once ordered a set of speakers after decades of DIY. Those were deemed nearly perfect, price was a no brainer. I added a set of humble subs, no prob.
Thing is, now on top of the hill as good as it gets I care about room integration in bass, and then after, all is there. My hearing is the limit, but more so my mind. Do I like the music? Rarified my listening sessions, don't like to listen to music that often anymore. Bad, good, who knows. Is it the stereo's flaws - yes and no. I still like to treat a real piano myself, talking to musicians, going to new style concerts (I dismiss the classics). The stereo, after objective perfection, reveals all the limitations that recorded music inherently bears: arbitrary repeatability for one.

That's the effect of a perfect stereo. Once it's done, the curtain is withdrawn, there you are with it.

So you've made it to the end? I'm still on the road, even though I've been enjoying it for forty years! That said, the greatest pleasure in listening to high fidelity is found with classical music and jazz. Excellent recordings and defining musical passages, appreciated only with top-notch systems/rooms. The emotion and authenticity transport you to the place where the event was recorded.

 

[Digital_Thor]

Sorry, but you leave us with nothing but guessing, which is rather pointless regarding audio. We already have thousands of places to go, if we just want storytelling and unnecessary ramblings about going nowhere, we can read hifi magazines.
Show us the speaker, name the drivers and some physical measurements of the front baffle - or else we can't do anything but return words without a goal or any true meaning.
Sorry to be a bit harsh, but what did you expect us to do in a rational scientific forum, when we are fed almost no data to work with?
If you want to play, then play

 

[dualazmak]

Sorry, but you leave us with nothing but guessing, which is rather pointless regarding audio. We already have thousands of places to go, if we just want storytelling and unnecessary ramblings about going nowhere, we can read hifi magazines.
Show us the speaker, name the drivers and some physical measurements of the front baffle - or else we can't do anything but return words without a goal or any true meaning.
Sorry to be a bit harsh, but what did you expect us to do in a rational scientific forum, when we are fed almost no data to work with?
If you want to play, then play

I essentially agree!
I would like to invite OP @MIsha to my thread "Let's share diagrams (and photos) of our total physical audio system and the whole signal path, with a few words and/or links" ; @MIsha, you need to share your diagram(s) and photo(s).

 

[MIsha]

Dear friends, I have stated my reasons quite clearly, and I still stand by my position. I do not intend to disclose the name of the loudspeaker until an explanation is found.
I can assure you that my loudspeaker is not fundamentally different from the vast majority of similar designs. It is a floorstanding speaker, 110 cm tall and 22 cm wide. I hope this information helps to explain my problem. Thank you.