The greatest speakers ever?

[thewas]

I really appreciate the deep dive into the blades. I considered getting them at one point but had a feeling they won’t be a a good fit in my rather reflective room with walls relatively near to speaker position. Now I better understand why it would have been a bad idea .

I would rather recommend you listening to them yourself in a similar setup at home or at your local dealer as there are better documented opinions of people owning them or having thoroughly tested them like Kal (who is even an active member here) and Erin (who used to be one) who are quite contradictory to such. In my opinion directivity steps so low are not an issue as there humans don't really discriminate direct and reflected sound so well (which is the reason equalisation/DRC works well there) and many if not most here well renown loudspeakers have rather increasing directivity vs constant one.

KEF Blade Two Meta loudspeaker | Stereophile.com

It seems as if I have been waiting for these all my life. Not in any existential sense, but in a literal, practical way: The arrival of the Blade Two Meta is the culmination of a lifelong fascination with KEF. As a teenager, I was introduced to founder Raymond Cooke and his innovative...

www.stereophile.com

KEF Blade 2 Meta Speaker Review

KEF Blade 2 Meta Speaker

www.erinsaudiocorner.com

In the end you must enjoy them and my or other opinions while helpful cannot replace an own judgement.

 

[Blockader]

I would rather recommend you listening to them yourself in a similar setup at home or at your local dealer as there are better documented opinions of people owning them or having thoroughly tested them like Kal (who is even an active member here) and Erin (who used to be one) who are quite contradictory to such. In my opinion directivity steps so low are not an issue as there humans don't really discriminate direct and reflected sound so well (which is the reason equalisation/DRC works well there) and many if not most here well renown loudspeakers have rather increasing directivity vs constant one.

KEF Blade Two Meta loudspeaker | Stereophile.com

It seems as if I have been waiting for these all my life. Not in any existential sense, but in a literal, practical way: The arrival of the Blade Two Meta is the culmination of a lifelong fascination with KEF. As a teenager, I was introduced to founder Raymond Cooke and his innovative...

www.stereophile.com

KEF Blade 2 Meta Speaker Review

KEF Blade 2 Meta Speaker

www.erinsaudiocorner.com

In the end you must enjoy them and my or other opinions while helpful cannot replace an own judgement.

To be fair, kimmosto has a point.

There are 2 flaws to KEF Blade's design,

Side woofers make the dispersion shift from omnidirectional to directional without the same DI consistency and slope the speakers maintain above the low end side woofers range.
The side woofers are too high.

However, all designs have flaws. A flush mounted, human sized speaker with the tweeter at ear level and a 21" woofer at the bottom would, I believe, be better than KEF Blades, though such speakers would also cost more.

 

[thewas]

However, all designs have flaws.

I would also rather say compromises. The side woofers of the Blade and narrow baffle make it loose directivity at 300 Hz and lower (which in my experience isn't by far as critical as directivity steps above) but on the other side make it image better than most loudspeakers in that class and also look more like a piece of art compared to a classic big box, but that is also personal taste and priority dependent. Also any limited number of woofers will have nulls, the question is where they are located and how wide they are, that's why when it comes to highest bass quality solutions like the W371 are needed which are very expensive and also often not living room compatible.

 

[Blockader]

I would also rather say compromises. The side woofers of the Blade and narrow baffle make it loose directivity at 300 Hz and lower (which in my experience isn't by far as critical as directivity steps above) but on the other side make it image better than most loudspeakers in that class and also look more like a piece of art compared to a classic big box, but that is also personal taste and priority dependent. Also any limited number of woofers will have nulls, the question is where they are located and how wide they are, that's why when it comes to highest bass quality solutions like the W371 are needed which are very expensive and also often not living room compatible.

Having side woofers allows them to make the speakers very narrow and curvy which allows the directivity above 300hz smooth like we have never seen before. I believe that's the trade-off. Yes.

 

[thewas]

Having side woofers allows them to make the speakers very narrow and curvy which allows the directivity above 300hz smooth like we have never seen before. I believe that's the trade-off. Yes.

Yes, plus design/looks and maximum reduction of secondary sound diffraction.

 

[Arindal]

In my opinion directivity steps so low are not an issue as there humans don't really discriminate direct and reflected sound so well (which is the reason equalisation/DRC works well there)

I agree to that in general, below 500Hz humans are not capable of differentiating between direct sound and reflections just by the delay between them.

Directivity is not completely irrelevant in the lower regions, though, as omnidirectional (or in case of the KEF negative D.i.) behavior can lead to the room being excited in a different way, with the extra resonances and longer decay potentially audible. And that is exactly the thing you cannot EQ satisfactory.

The second issue is that directivity steps are one thing, but also a continuously increasing directivity index can lead to tonal imbalance of the indirect sound which I would consider a major flaw.

many if not most here well renown loudspeakers have rather increasing directivity vs constant one.

That does not surprise me, as proper constant directivity speakers are much more complicated to design, and necessarily more expensive with similar SPL capabilities. Increasing directivity, on the other hand, seems to be accepted by a number of listeners and reviewers, as their primary goal is the on-axis FR, so they just accept imbalanced reverb as they got used to it (or circumvent it in case of studio/nearfield/mono listeners).

 

[thewas]

Directivity is not completely irrelevant in the lower regions, though, as omnidirectional (or in case of the KEF negative D.i.) behavior can lead to the room being excited in a different way, with the extra resonances and longer decay potentially audible. And that is exactly the thing you cannot EQ satisfactory.

Even typical hifi loudspeakers (non-cardioid) with a bit wider baffle just take a the directivity a bit lower, also in that reasons most problematic phenomena like peaks are mainly minimal-phase so EQ actually reduces decay and resonances well.

The second issue is that directivity steps are one thing, but also a continuously increasing directivity index can lead to tonal imbalance of the indirect sound which I would consider a major flaw.

I think every frequent reader knows that by now as well as our disagreement and arguments so let's not start it all over again here.

 

[Arindal]

in that reasons most problematic phenomena like peaks are mainly minimal-phase so EQ actually reduces decay and resonances well.

When we are talking about room modes, resonators and decay, almost nothing is truly minimum phase over the whole system. You can EQ the amplitude response for one point, but almost certainly nulls at other places, group delay and decay persist somewhere.

as well as our disagreement and arguments so let's not start it all over again here.

We don´t have to, but maybe some people are encouraged to do their own listening comparisons.

 

[thewas]

When we are talking about room modes, resonators and decay, almost nothing is truly minimum phase over the whole system. You can EQ the amplitude response for one point, but almost certainly nulls at other places, group delay and decay persist somewhere.

Nulls cannot be equalized but room modes (which are usually also the biggest audible problems) and those are mainly minimum phase.

We don´t have to, but maybe some people are encouraged to do their own listening comparisons.

Which is what I had recommended above.

 

[Arindal]

Nulls cannot be equalized but room modes (which are usually also the biggest audible problems) and those are mainly minimum phase.

Nulls are in most cases also an outcome of room modes, so how do you expect their peaks to be fully correctible in terms of decay?

A single resonator might be described as a minimum phase system, but the moment you have several of them, overlaying each other, plus introducing group delay distortion and all-pass filter behavior, well the whole system is not minimum phase anymore and not fully EQ-able, particularly not if there is audibly long decay or nulls as a part of the whole thing.

 

[thewas]

Nulls are in most cases also an outcome of room modes, so how do you expect their peaks to be fully correctible in terms of decay?

Resonances are typically minimum-phase phenomena, because the peak (magnitude response) is tightly linked to the corresponding phase shift. That means: if you know the amplitude response, the phase response can be derived via the Hilbert transform. The system is causal and stable, and its transfer function has poles inside the unit circle (in discrete-time terms), which gives the minimum-phase property.

A room null is a destructive interference effect: at a certain point in space, the direct sound and reflected sound(s) arrive out of phase and cancel each other. Unlike resonances, these are caused by superposition and cancellation, not by poles in the transfer function. This introduces non-minimum phase behavior because the frequency response zeros may lie outside the unit circle. In simple terms: a dip (null) in magnitude response does not correspond to the "expected" minimum-phase phase shift. These cancellations can come from multiple path delays — e.g. direct + reflected path = comb filtering — which is a mixed-phase situation.

A single resonator might be described as a minimum phase system, but the moment you have several of them, overlaying each other, plus introducing group delay distortion and all-pass filter behavior, well the whole system is not minimum phase anymore and not fully EQ-able, particularly not if there is audibly long decay or nulls as a part of the whole thing.

Several peaks overlaying see superposition, the rest is not not what makes the problems of the modal peaks which can be well equalised and many room measurements have also confirmed that.

 

[Arindal]

if you know the amplitude response, the phase response can be derived via the Hilbert transform.

I am aware of the theory. But the moment you have several pole frequencies of different resonators/filters close to each other, this is not applicable anymore. The easiest example is a simple crossover consisting of lowpass and highpass filter, 2nd order or higher. The resulting frequency response is perfectly flat, which would mean zero group delay distortion, if the transformation phase to amplitude would be transparent. But a simple simulation or measurement would show that the resulting overall system induces group delay (primarily to the lowpass-filtered signals), behaves partly like a bandwidth-limited allpass filter which is by definition not minimum phase.

his introduces non-minimum phase behavior because the frequency response zeros may lie outside the unit circle. In simple terms: a dip (null) in magnitude response does not correspond to the "expected" minimum-phase phase shift. These cancellations can come from multiple path delays — e.g. direct + reflected path = comb filtering — which is a mixed-phase situation.

Correct, but the same is true to most of peaks, if several resonators with different pole frequencies are contributing to the sum. They are also non-minimum-phase, which makes the idea of getting all decay and group delay distortion back to zero somehow hopeless.

The genie is out of the bottle (ou: le bouchon est sorti de la bouteille de champagne).

the rest is not not what makes the problems of the modal peaks which can be well equalised and many room measurements have also confirmed that.

Which waterfall plot or group delay measurement has shown that you can fully eradicate the consequences of room modes?

 

[thewas]

I am aware of the theory. But the moment you have several pole frequencies of different resonators/filters close to each other, this is not applicable anymore. The easiest example is a simple crossover consisting of lowpass and highpass filter, 2nd order or higher. The resulting frequency response is perfectly flat, which would mean zero group delay distortion, if the transformation phase to amplitude would be transparent. But a simple simulation or measurement would show that the resulting overall system induces group delay (primarily to the lowpass-filtered signals), behaves partly like a bandwidth-limited allpass filter which is by definition not minimum phase.


Correct, but the same is true to most of peaks, if several resonators with different pole frequencies are contributing to the sum. They are also non-minimum-phase, which makes the idea of getting all decay and group delay distortion back to zero somehow hopeless.

The genie is out of the bottle (ou: le bouchon est sorti de la bouteille de champagne).

Not at all, rather a showcase of a poor analogy used to win an argument. Minimum-phase doesn’t mean "linear phase" or "no group delay", a system can be minimum-phase and still have frequency-dependent group delay. The crossover’s lowpass + highpass system is indeed minimum-phase (its poles and zeros are in the stable region). The non-linear phase/group delay is completely consistent with minimum-phase behavior. So the "perfectly flat magnitude" but "non-zero delay" isn’t a contradiction — it’s exactly what you’d expect from a minimum-phase all-pass-like system. Room modes though are not independent filters stitched together. In a real acoustic space, the standing-wave resonances are coupled modes of the wave equation. Each mode contributes poles (resonant peaks) to the transfer function. Poles inside the stable region always yield a minimum-phase contribution to the system. The crossover example does not demonstrate that resonances stop being minimum-phase when they overlap. It only demonstrates that minimum-phase does not imply constant group delay or linear phase. On the other hand the acoustic transfer function of a room (without considering interference nulls) is a causal, stable system. Its resonant modes arise from poles in the transfer function. Poles inside the stable region have by definition minimum-phase contribution. Even with many overlapping modes, the total system remains minimum-phase, because the pole-zero pattern still satisfies the conditions (all poles and zeros inside the unit circle, for the relevant paths). The only time you break minimum-phase in a room is with cancellations from multiple path interference (nulls), not from piling up resonances. The crossover analogy confuses "minimum-phase vs. non-minimum-phase" with "linear-phase vs. non-linear-phase. Room modes being minimum-phase doesn’t mean they have flat group delay. It just means that the phase is completely determined by the magnitude response. Thus the "crossover delay" example is irrelevant, the phase distortion is still consistent with minimum-phase behavior.

Which waterfall plot or group delay measurement has shown that you can fully eradicate the consequences of room modes?

I never claimed that equalisation can "fully eradicate the consequences of room modes" but that it can mitigate their most problematic impacts.
A good AES literature example is https://assets.ctfassets.net/4zjnzn...qualization_Of_Loudspeaker-Room_Responses.pdf

 

[Robin L]

That subwoofers is 6 to 7 feet at best diameter. What is this crap about 20 feet?

Bad AI in a deliberate joke.

 

[Goldhamster916]

I agree to that in general, below 500Hz humans are not capable of differentiating between direct sound and reflections just by the delay between them.

Directivity is not completely irrelevant in the lower regions, though, as omnidirectional (or in case of the KEF negative D.i.) behavior can lead to the room being excited in a different way, with the extra resonances and longer decay potentially audible. And that is exactly the thing you cannot EQ satisfactory.

The second issue is that directivity steps are one thing, but also a continuously increasing directivity index can lead to tonal imbalance of the indirect sound which I would consider a major flaw.



That does not surprise me, as proper constant directivity speakers are much more complicated to design, and necessarily more expensive with similar SPL capabilities. Increasing directivity, on the other hand, seems to be accepted by a number of listeners and reviewers, as their primary goal is the on-axis FR, so they just accept imbalanced reverb as they got used to it (or circumvent it in case of studio/nearfield/mono listeners).

kef actually seems to think directivity tapering is beneficial, this here is from their white paper. they highlight a change in cross over design to change the directivity at the mf/lf crossover. first one is the first blade generation second the meta version, also read the text:

 

[mhardy6647]

The finest loudspeakers I've heard with my own personal ears are these (the two outboard ones, that is -- I've yet to hear the mono loudspeaker shown between them in Herb's photo).

The quantitative data for these RCA field coil-based three-way loudspeakers in situ is also, actually, very good -- not that I have the data (but I have seen it).

 

[Arindal]

a system can be minimum-phase and still have frequency-dependent group delay.

Yes, that was my point. If you have a system with inherent all-pass behavior (including amplitude-independent phase shift/group delay distortion), you can equalize the amplitude to perfection, but the phaseshift or group delay distortion remains. That is the definition of a non-minimum-phase system.

I never claimed that equalisation can "fully eradicate the consequences of room modes" but that it can mitigate their most problematic impacts.

That might be the case for a single position in the room and in case room modes are not causing decay or group delay problems which are audible in the time domain alone. That might be doable in a studio or heavily treated room, but it is rare under home-listening conditions. So the answer is not, DSP cannot mitigate most problematic outcome of room modes, although it might be possible to reduce their dominance.

kef actually seems to think directivity tapering is beneficial, this here is from their white paper.

I don´t really take from that they say it is always beneficial. It seems to be more plausible that they want to avoid suddens steps (which is a good idea per se) and due to their driver/baffle geometry end up with a compromise preferring continuously increasing D.I. over suddenly changing one or combfiltering effects. It is a general problem with almost all slim, rounded baffle speakers and conventional coaxial design with comparably steep midrange cone design. The higher the frequency, the more efforts you need to take to avoid diffraction, cancellation and unwanted directivity changes, so as a tendency you'd better narrow down directivity towards higher frequencies.

first one is the first blade generation second the meta version,

Regarding even behavior over the listening window, the old version is definitely better, while the Meta seems to avoid horizontal lobing and sudden steps in directivity. What this means to sound quality in a room, only a listening test can tell.

 

[kimmosto]

...highest bass quality solutions like the W371 ...

Some say it's bad and useless product which ruins the sound of e.g. 8361. Probably not in all cases, but the design does not look very "idealistic" imo.

 

[Purité Audio]

Sour grapes?
Keith

 

[Mr. Widget]

The finest loudspeakers I've heard with my own personal ears are these (the two outboard ones, that is...

I have heard many uber high end speakers over the years. The original WAMM, the original Infinity IRS, the TAD Model 1 and the Reference 1 several times each, a couple of different iterations and installations of the Magico Ultimates, Steinway Lyngdorf Model D and LS Concerts... every one of these extremely large and extremely costly systems were amazing at times and yet never perfect.

I can not say that any one of these was the finest or the GREATEST. Further, at this stage in my life, I don't want to own any speakers that are as large as any of these!