A history of Time Aligned Loudspeakers

[gberchin]

But the time alignment is measurable.

Exactly my point.

Assuming fidelity is how well the reproduced signal matches the input signal, then there may be some hifi definition there.

Also my point. The general assumption is that anything that is not measurable is not audible, and anything that is audible is measurable. Now we have the opposite case; something that definitely is measurable, but not easily audible. In my opinion, audible or not, it is still worth getting right. YMMV, as they say.

 

[norman bates]

I've always wondered if "time aligned/phase aligned" with acoustic centers lined up are "better" or "more accurate", why aren't they predominantly used at recording studios ?

You can address diffraction through roundovers / felt / etc......


And I agree, the farfield can audibly swamp the benefits of time alignment.........
I thought a Vandersteen 1c (about a year ago) sounded awful, but I enjoyed Thiel model 04 and cs2.

 

[pablolie]

How many loudspeaker companies built these types of loudspeakers? Post or quote your favorite pics or example.

I recall the first ones coming out in the late 70’s here is one by Leak, that used a stepped baffle. Its was all the rage in the 70’s

View attachment 533709

question is how much should you read into an ad at the time.

 

[pablolie]

Exactly my point.

Also my point. The general assumption is that anything that is not measurable is not audible

Prove the contrary. Audio honestly is low lying "science"... 192kHz? 24 bits? wake me up if that's ever posed a remote challenge to our infrastructure in decades ...

And. come. on. Nyquist is mathematically proven real. only the proverbial village idiot would doubt it now.

 

[Waveform Fidelity]

question is how much should you read into an ad at the time.

In my opinion, it depends on the content of the ad - wayback then, it sparked my interest as a young engineer interested in loudspeakers as a hobby.

If ads are backed up by real measurements that explains the design principles, then perhaps read further. If it's all "wishy-washy" with subjective descriptors... move on

I think the KEF company have got it "right" over the years - they have the general marketing approach and if your want to dig deeper you get the white papers, which share enough engineering detail for the Hi Fi enthusiast.

 

[Waveform Fidelity]

Exactly my point.

Also my point. The general assumption is that anything that is not measurable is not audible, and anything that is audible is measurable. Now we have the opposite case; something that definitely is measurable, but not easily audible. In my opinion, audible or not, it is still worth getting right. YMMV, as they say.

Yes getting it right is worth it, at least on an intellectual level as an Engineer. Continuous improvement and doing the best that can be done is worth it ( I know... I am just about to annoy the "just good enough" crowd).

Personally, I don't agree that the human is "good enough" at comparing sounds that they just heard and memorized with the ones they are now hearing. Sure if the sounds are radically different then yes, but we are only talking about the fine detail. If in doubt, and its cost effective then make it perfect, just in case......

 

[Waveform Fidelity]

Prove the contrary. Audio honestly is low lying "science"... 192kHz? 24 bits? wake me up if that's ever posed a remote challenge to our infrastructure in decades ...

And. come. on. Nyquist is mathematically proven real. only the proverbial village idiot would doubt it now.

Some idiots live in towns too!

 

[Holmz]

Total BS. Cable differences are easily measurable, e.g. at RF frequencies. Can any of these audiophools claims that these differences are audible but unmeasurable withstand properly conducted double blind tests?

Sometimes those RF can IMD their way and appear somewhere like a wraith at a seance.
But that is also dependant upon amplifier stability.

 

[Trifonov Audio]

But the time alignment is measurable.
Assuming fidelity is how well the reproduced signal matches the input signal, then there may be some hifi definition there.

Absolutely true! But there are "audiophiles" - who say: "This is not necessary" or"This is impossible with passive filters higher than first order"!.......

 

[Suono]

At most, it's known that larger offsets cause interference and change the directional characteristics. But those effects are often much less audible than internet discussions would suggest. Especially in a normal listening room, with reflections and the listener's head movement.

So in my opinion, the original poster did a perfectly sensible practical experiment, and his conclusions are not 'wrong' at all. They just didn't confirm a theory that someone really wants to believe in.

A violin, a guitar, a double bass, a flute, play in your listening room. Is the sound in phase and temporally coherent? YES.
Why shouldn't the speakers be? I don't understand.

 

[Trifonov Audio]

A violin, a guitar, a double bass, a flute, play in your listening room. Is the sound in phase and temporally coherent? YES.
Why shouldn't the speakers be? I don't understand.

When we listen to live musical instruments without an amplification system - there are no phase-time problems.
The problem is caused by crossovers higher than the first order in the speakers including mounting all drivers on one common
vertical plane!

 

[robh36062]

A violin, a guitar, a double bass, a flute, play in your listening room. Is the sound in phase and temporally coherent? YES.
Why shouldn't the speakers be? I don't understand.

Hello

There is a window where we can't hear group delay. Most speakers are not time or phase coherent. A well designed speaker system will fall into this window so we can't hear the time issues that they have. We can measure using Excess Group Delay and clearly see issues in the Step Response but they are not audible because they are under the threshold of audibility.


Rob

 

[LSPhil]

When we listen to live musical instruments without an amplification system - there are no phase-time problems.
The problem is caused by crossovers higher than the first order in the speakers including mounting all drivers on one common
vertical plane!

A loudspeaker driver itself already behaves as both a high-pass and low-pass acoustic filter. Once you add even a single capacitor (a 1st-order electrical filter), the resulting acoustic slope is, in the best-case scenario, effectively 2nd order.

So the idea of a “true 1st-order acoustic high-pass loudspeaker” is largely theoretical and practically nonexistent in real-world driver behavior.

That’s why I think your conclusion is based on information that sounds well-read, but is actually quite questionable from a physical and acoustical standpoint.

Also, phase and time behavior are not problems that magically appear only because of higher-order crossovers. Every driver has its own phase rotation, bandwidth limitations, acoustic center offset, breakup behavior, and natural roll-off characteristics. The crossover simply interacts with those realities.

And regarding the “live instruments are perfectly phase coherent” argument — real instruments are spatially separated sources radiating into a room with reflections, propagation delays, and complex phase relationships all the time. Absolute phase coherence at the listening position does not exist there either.

For reference, here’s a measurement I made today of my Scan-Speak 18M.

One more thing.

My daughter plays flute, classical guitar, and piano. During the lockdowns I had to record her performances with two microphones and send them to her teacher regularly.

Honestly, there was no dramatic “this is obviously fake reproduction” moment for me when comparing the sound of my speakers to hearing her play live in the same room. Of course they are not literally identical, I’m not claiming that, but the gap was surprisingly small.

Maybe this sounds a bit arrogant, but my speakers are apparently not as terrible as some people here would like to believe.

 

[robh36062]

And regarding the “live instruments are perfectly phase coherent” argument — real instruments are spatially separated sources radiating into a room with reflections, propagation delays, and complex phase relationships all the time. Absolute phase coherence at the listening position does not exist there either.

Hello

If you look at the directivity of musical instruments it's complex for each individual instrument none are point sources and tone changes with the physical orientation to the listener.

Rob

 

[kimmosto]

A well designed speaker system will fall into this window so we can't hear the time issues that they have.

Are Genelec 8260A, 8351A, Revel Ultima Salon 2, KEF Reference 207/2, Linkwitz Orion, Linkwitz 521 etc. and all mid-sized and small speakers requiring sub-woofer but otherwise okay, well designed speakers?
Where are the studies indicating that timing below audibility limit cannot be perceivable in soundstage with multi-way speakers with separate drivers + possible diffraction, or as tactile sensation? Available phase/GD studies within last 40 years have been terribly poor so why should we believe that audibility limit is adequate target, and very close zero excess GD would be just exaggeration?

 

[robh36062]

Available phase/GD studies within last 40 years have been terribly poor so why should we believe that audibility limit is adequate target, and very close zero excess GD would be just exaggeration?

There have been more modern studies than the 1978 BL curve posted in this thread. I will post a summary of 3 different studies one from 2018.

Where are the studies indicating that timing below audibility limit cannot be perceivable in soundstage with multi-way speakers with separate drivers + possible diffraction, or as tactile sensation?

Are you doing these studies in a real listening room or an anechoic chamber? In a real room any reasonable GD issues, will be further masked by the reverberant field in the room. Making even a borderline system inaudible.

Are Genelec 8260A, 8351A, Revel Ultima Salon 2, KEF Reference 207/2, Linkwitz Orion, Linkwitz 521 etc. and all mid-sized and small speakers requiring sub-woofer but otherwise okay, well designed speakers?

All? How would anyone know? What does requiring a sub have to do with it? We are talking complete speaker systems individually looking at GD for that system. You are worried about sub integration in the modal region of a typical room where the room dominates?

Where are the studies indicating that timing below audibility limit cannot be perceivable in soundstage with multi-way speakers with separate drivers + possible diffraction, or as tactile sensation?

Simply not concerned.

Rob

 

[olieb]

Available phase/GD studies within last 40 years have been terribly poor so why should we believe that audibility limit is adequate target, and very close zero excess GD would be just exaggeration?

Well, you may believe whatever you want.
What exactly is so poor about all those studies besides the result, that you seem not to like?And if all those studies were so poor, it should be no problem to do it better and show evidence for the overwhelming importance of flat phase.
But so far the evidence is what it is.
And then there is the evidence of “the proof is in the pudding“. Your list of speakers above isn’t going down too badly with the enthusiasts, so I wonder if the big issue with phase is actually not that big after all.

Where are the studies indicating that timing below audibility limit cannot be perceivable in soundstage with multi-way speakers with separate drivers + possible diffraction, or as tactile sensation?

You can always invent a (new) case that has not been considered and thereby open a new backdoor for your statements. The story of that teapot of Bertrand.

Edit: I posted this before, but FWIW a paper for Liski (2020) that gives a short overview of earlier studies, too.

https://acris.aalto.fi/ws/portalfiles/portal/66449704/Audibility_of_Group_Delay_Equalization.pdf

 

[Trifonov Audio]

A loudspeaker driver itself already behaves as both a high-pass and low-pass acoustic filter. Once you add even a single capacitor (a 1st-order electrical filter), the resulting acoustic slope is, in the best-case scenario, effectively 2nd order.

So the idea of a “true 1st-order acoustic high-pass loudspeaker” is largely theoretical and practically nonexistent in real-world driver behavior.

That’s why I think your conclusion is based on information that sounds well-read, but is actually quite questionable from a physical and acoustical standpoint.

Also, phase and time behavior are not problems that magically appear only because of higher-order crossovers. Every driver has its own phase rotation, bandwidth limitations, acoustic center offset, breakup behavior, and natural roll-off characteristics. The crossover simply interacts with those realities.

And regarding the “live instruments are perfectly phase coherent” argument — real instruments are spatially separated sources radiating into a room with reflections, propagation delays, and complex phase relationships all the time. Absolute phase coherence at the listening position does not exist there either.

For reference, here’s a measurement I made today of my Scan-Speak 18M.
View attachment 534953
One more thing.

My daughter plays flute, classical guitar, and piano. During the lockdowns I had to record her performances with two microphones and send them to her teacher regularly.

Honestly, there was no dramatic “this is obviously fake reproduction” moment for me when comparing the sound of my speakers to hearing her play live in the same room. Of course they are not literally identical, I’m not claiming that, but the gap was surprisingly small.

Maybe this sounds a bit arrogant, but my speakers are apparently not as terrible as some people here would like to believe.

Okay, okay... There is some "reasonableness" in the arguments presented. Do you have a suggestion for compensating for the phase-time problems in a loudspeaker in multi-way loudspeakers?
Of course - there are speakers without perfect Phase and Time (good Step Response) - that sound acceptably good.
In some rare cases, the presence of an interesting, effect and attractively pleasant sound can "replace" the need for
Transient Perfect Sound......
A specific (exact) model and manufacturer should always be commented on.

 

[hex168]

Any idea what percentage of source material accurately records undistorted time domain information? To what extent does that factor make this discussion moot?

 

[kimmosto]

To what extent does that factor make this discussion moot?

The funniest excuse to design bad speakers would be to trust that recordings have some delay errors and manipulations which hide problems of own product.