KEF Q350 Speaker Review

[civi]

I think KEF should have performed a formal double blind listening test to show the merits of coaxial drivers given how much they rely on it.

Great vertical directivity should be enough to seriously consider coaxial over i.e. elac dbr for bizarre speaker placement scenarios (like one near the ceiling and one placed in reasonably standard position).

 

[Lbstyling]

Amir knows best of course. I'm partially projecting though because that largely reflects my feelings. The difference is when I measure/listen to narrower directivity speakers I will almost universally toe them out or use extreme toe-in to increase the loudness of sidewall reflections. Amir can't really do that with mono listening.

Would be nice to see more research on this. There's some conflicting stuff in Toole's book about it. On one hand, the preference score papers suggest directivity was not a strong factor in preference other than the best speakers had smooth directivity. On the other hand, directivity-specific studies suggest people will pick wider speakers for recreational listening.

Also, we know that the same speakers tend to win overall in both mono and stereo tests, but I wonder whether we can extrapolate that to a single listener who seems (again, this is me guessing and I could be totally wrong) to have particular directivity preferences listening to dozens or (eventually) hundreds of speakers.

Having discussed and read this to somewhat OTT extremes on the JTR thread, Toole states in 2008? paper that wide AND smooth directivity is preferable to the general public Professionals were less polarised towards wide directivity, but still preferred it.

This supersedes his earlier work stating that the width of directivity was a matter of taste.

The problem is that if you are not careful, you can end up concluding that a pair of large Amazon echo speakers are the pinnacle of audio.

 

[napilopez]

Having discussed and read this to somewhat OTT extremes on the JTR thread, Toole states in 2008? paper that wide AND smooth directivity is preferable to the general public Professionals were less polarised towards wide directivity, but still preferred it.

This supersedes his earlier work stating that the width of directivity was a matter of taste.

The problem is that if you are not careful, you can end up concluding that a pair of large Amazon echo speakers are the pinnacle of audio.

I'm familiar with that research, but it doesn't negate it being a matter of taste. Wide is preferred by the majority (including me!), but it doesn't mean it's preferred by all. In the oft-cited blind comparison of the Revel Salon2 and JBL M2, which the Salon2 won by a substantial margin, three participants out of 11 consistently chose the M2 over the Salon2.

 

[Victor Affonso Damas]

Q350 is really good, I used to own a pair. Come on amirm, you did enjoy the bloated midbass of the M16, why did you mistreat the Q350?

 

[Sancus]

Yes, and notice the last updates of Musings by Olive - date and topics http://seanolive.blogspot.com/

About directivity and preference. Seems to me that USA people prefer wide and smooth directivity and Europeans narrow and smooth. Asians are not systematic (in my view at least...). This might have something to do with typical room size and construction (reflectiveness)

It would make a lot of sense if the "directivity width preference" is actually a preference for ratio of direct:reflected sound.

 

[Lbstyling]

I'm familiar with that research, but it doesn't negate it being a matter of taste. Wide is preferred by the majority (including me!), but it doesn't mean it's preferred by all. In the oft-cited blind comparison of the Revel Salon2 and JBL M2, which the Salon2 won by a substantial margin, three participants out of 11 consistently chose the M2 over the Salon2.

Its a weak correlation to be sure. I run a JMLC horn that holds CD to 15 deg, so your preaching to the choir on this one

Toole for whatever reason goes to rather great lengths (much writings, many effort!) to shake loose the earlier conclusions in his work of 'matter of taste' when you read it in his latest book, including citing research showing narrower as better sounding to professionals being a poor non real world test etc. Is it 'reaching', yes maybe, but he makes the statement none the less.

Bear in mind that he covers open baffle designs in his discussions at length, so we really are talking wide when he says this too.

 

[SynthesisCinema]

I shipped these back already so will have to test this in a future review.

Shame you send them back already. I assume you listened the one speaker with port open? There is probably two subjective reviews that claims the bass control and quality is much better when one uses the port bungs with Q350! I got a feeling that the bungs are must with these speakers as the bass is otherwise too strong and loose. Well maybe you get the Q150 in future and hopefully will listen the pair next time aswell. But doing that and listening others with just one you ain`t making new friends.

People are quite pissed for you in different forums how you do the subjective listening part with just one speaker, some in pairs and for the receivers/processors you don´t listen them at all which they can´t understand as a review without listening is meaningless for them. "No one buys product for how it measures, but how it sounds is what matters most". I try to give ASR lot of visibility by sharing links, but people don´t understand measurements they scroll down to listening part ONLY and as you won´t give your recommendation for most products people are not happy for the few words you often write. Last one was at AVSF someone didn´t believe your measurements. They said to trust NRC more.

 

[Sancus]

People are quite pissed for you in different forums how you do the subjective listening part with just one speaker, some in pairs and for the receivers/processors you don´t listen them at all which they can´t understand as a review without listening is meaningless for them. "No one buys product for how it measures, but how it sounds is what matters most". I try to give ASR lot of visibility by sharing links, but people don´t understand measurements they scroll down to listening part ONLY and as you won´t give your recommendation for most products people are not happy for the few words you often write. Last one was at AVSF someone didn´t believe your measurements. They said to trust NRC more.

The problem here is those people though, not the reviews here. Anyone who reads reviews "for the listening part" is wasting their time on ASR. Especially when it comes to electronics.

 

[andreasmaaan]

It's been a while since I read Toole, but I would characterise his comments on wide vs narrow directivity more as well-reasoned hypothesis than as actual science.

The most rigorous actual scientific data we have on the question to my knowledge comes from Olive's multiple regression analysis studies.

I recently posted the following in another thread, so sorry for the re-post here. But below are the target slope values (i.e. slopes most closely correlating with those speakers in the studies that were most preferred) calculated from Olive's studies:

I've plotted the target slopes for the larger, wider-directivity sample set using Olive's formula:

Here are those plots:

As I noted in the other thread, the target slopes of the wider-directivity sample set actually match those of KEF speakers more closely than they match those of Revel.

And yet, it seems that on this forum KEF speakers are frequently cited as being too narrow in directivity to be preferred by most listeners.

Moreover, these target slopes do not represent what I think many people have in mind when they use the term "wide-directivity", although ofc they are not "narrow directivity" either.

As I also mentioned in the other thread, it might be interesting to speculate that wider directivity was the reason 8 out of 11 listeners preferred the Salon2 over the M2 in the oft-cited shootout, but the sample is too small and the variables too numerous to call such speculation science (by a long stretch IMHO).

 

[MZKM]

It's been a while since I read Toole, but I would characterise his comments on wide vs narrow directivity more as well-reasoned hypothesis than as actual science.

The most rigorous actual scientific data we have on the question to my knowledge comes from Olive's multiple regression analysis studies.

I recently posted the following in another thread, so sorry for the re-post here. But below are the target slope values (i.e. slopes most closely correlating with those speakers in the studies that were most preferred) calculated from Olive's studies:

I've plotted the target slopes for the larger, wider-directivity sample set using Olive's formula:

Here are those plots:

As I noted in the other thread, the target slopes of the wider-directivity sample set actually match those of KEF speakers more closely than they match those of Revel.

Moreover, these target slopes do not represent what I think many people have in mind when they use the term "wide-directivity", although ofc they are not "narrow directivity" either.

As I also mentioned in the other thread, it might be interesting to speculate that wider directivity was the reason 8 out of 11 listeners preferred the Salon2 over the M2 in the oft-cited shootout, but the sample is too small and the variables too numerous to call such speculation science (by a long stretch IMHO).

As was pointed out to me, in the same paper he states:

The degree of tilt varies depending upon the type of loudspeakers being tested. For example, 3-way and 4-way loudspeaker designs tend to have wider dispersion...at mid and high frequencies than 2-way loudspeakers. This suggests that the ideal target slope may depend on the loudspeaker's directivity.

 

[andreasmaaan]

As was pointed out to me, in the same paper he states:

The quote came out garbled for some reason @MZKM, but I presume these are his comments in which he basically notes that preferred target slopes are dependent on the directivity of the sample set and therefore conclusions can't be drawn concerning preferred target slopes from his studies?

I totally agree.

EDIT: and would add that this is the primary reason why applying the target slopes to loudspeakers measured here to derive preference ratings is fraught.

 

[MZKM]

The quote came out garbled for some reason @MZKM, but I presume these are his comments in which he basically notes that preferred target slopes are dependent on the directivity of the sample set and therefore conclusions can't be drawn concerning preferred target slopes from his study?

I totally agree.

I fixed the spacing errors. But yes.

 

[amirm]

Great vertical directivity should be enough to seriously consider coaxial over i.e. elac dbr for bizarre speaker placement scenarios (like one near the ceiling and one placed in reasonably standard position).

Should it? Here is directivity of human voice in horizontal and vertical planes:

On the left is horizontal and we see that it is symmetrical left and right. The one on the right is vertical and it definitely is not symmetrical up and down, nor is it that similar to what comes out on-axis. Absence of any direct research, it seems to be our brain would tolerate vertical directivity errors.

 

[andreasmaaan]

Should it? Here is directivity of human voice in horizontal and vertical planes:

On the left is horizontal and we see that it is symmetrical left and right. The one on the right is vertical and it definitely is not symmetrical up and down, nor is it that similar to what comes out on-axis. Absence of any direct research, it seems to be our brain would tolerate vertical directivity errors.

Are you suggesting that loudspeaker radiation should be modelled on the radiation pattern of the human voice?

 

[amirm]

Are you suggesting that loudspeaker radiation should be modelled on the radiation pattern of the human voice?

Not directly as there are other instruments in music. I am saying from evolutionary point of point, the central cortex would need to be tolerant of human voices have different response vertically than horizontally and filter out that out. Or else the conflict would cause annoyance.

It is a theory.

 

[Trouble Maker]

I've had a question bouncing around in my head about the PIR curves that I think relates to this.
The PIR curve is calculated at the MLP right?
If so, my key takeaway about (wide) directivity is about tolerance to moving outside of the MLP and keeping that curve, or at least the shape. I would imagine the same conclusion can be drawn about speaker placement in room. Someone mentioned this idea earlier about 'hanging' speakers from the ceiling. So, when I see big drops in the directivity plots close to on axis it makes me worry that the speaker would be very dependent on placement, the room and listener position. I don't have a dedicated listening room, it's a system in my living room that lives with us, not us with it. Tolerance to placement in room and listening position are paramount for me.
This is just my theory at this point.
Whether vertical directivity is nearly as important as horizontal seems to be a whole other thing too.

 

[Senior NEET Engineer]

Looks like we're mixing up sound production with sound reproduction

 

[napilopez]

The quote came out garbled for some reason @MZKM, but I presume these are his comments in which he basically notes that preferred target slopes are dependent on the directivity of the sample set and therefore conclusions can't be drawn concerning preferred target slopes from his studies?

I totally agree.

EDIT: and would add that this is the primary reason why applying the target slopes to loudspeakers measured here to derive preference ratings is fraught.

I agree as well, though more specifically he says the target slopes may vary depending on directivity rather than dismissing the target slopes altogether. For example, a wide directivity design with a highly negative slope would clearly be tuned dark.

However, to your earlier point about KEFs slopes vs the revels, I don't know how many people are looking at the PIR and ERDI curves for determining directivity. SPL plots or normalized SPL plots are for more useful for this purpose. The Spinorama curves have too much vertical influence and make directivity difficult to visualize. The Revels most certainly have wider horizontal directivity than the KEfs overall.

Of course this depends on the frequency range we're discussing too. I personally tend to focus on 2Kish to 8Kish for spatial cues. Waveguides speakers will usually be 'wider' in the top octave but that usually doesn't mean too much imo.

EDIT: Though me using that frequency region for directivity started out with my own correlation with measurements, Toole's book presents this image (Figure 6.3) which roughly backed up my subjective impressions

The "image shift, image source/broadening" bit.

 

[jhaider]

Interestingly lots of people have written that they vastly prefer the Beolab 90 in Narrow Mode vs Wide or Omni, which is contradicting the hypothesis.

Interesting point. A blind test of directivity preferences using those speakers would be extremely interesting.

 

[Dennis Murphy]

Not directly as there are other instruments in music. I am saying from evolutionary point of point, the central cortex would need to be tolerant of human voices have different response vertically than horizontally and filter out that out. Or else the conflict would cause annoyance.

It is a theory.

Interesting, but not necessarily true. It seems equally plausible to me that we've just come to regard this particular disparity between horizontal and vertical dispersion as normal for the human voice. I'm not sure it means that we wouldn't have come to accept a different sound of the human voice as "normal" if this disparity didn't exist. And, of course, a speaker with poor and/or restricted vertical response would exacerbate this disparity and fail to recreate what the brain has come to accept as normal. I hope you're right, however, because it's much more difficult and restricting to design a speaker with even vertical dispersion.