KEF Q100 Speaker Review

[bigjacko]

Room reflections will make the sound more enveloping but they will also reduce image sharpness.
As for musicians playing in the room, that has got more to do with the recording than with the playback.

I do agree on dipoles reduce image sharpness. Musician playing in the room is just an analogy, it tries to show that music is not from two front speakers, but from the whole room.

 

[edechamps]

Just to clarify: your position is that everybody should search through scads of forum threads for the necessary context and supplemental information

It's looking more and more like a "Speaker Measurement FAQ" pinned thread would be useful. With answers to recurrent questions regarding the measurement setup, "yes, it's done in Amir's garage, and yes, it is anechoic data", questions about the Olive score and what it means, questions about the harmonic distortion measurements, etc.

 

[amirm]

I think a brief paragraph before @MZKM posts his scores would be handy. It would say what the score is and a link to the research threads. I had to add such a thing to my reviews prior to measurements and is saving repeat question/answers.

 

[bobbooo]

hardly a falsifiable theory, not yet science

Of course it's science, and it is falsifiable - accurately measure around 70 speakers, do controlled double-blind listening tests with them, calculate the preference ratings from the measurements using Olive's formula and see if the correlation with listener ratings is comparable to what he found. The reason it hasn't been replicated is until now it's been very expensive and time-consuming to measure that many speakers accurately, as few people have access to a high quality anechoic chamber. Now Amir has the Klippel NFS, that side is much easier and quicker. But that still leaves the other side - highly controlled double-blind listening tests of a large sample of speakers involving trained listeners, which is a huge undertaking, requiring time, expense, specialist equipment, resources and properly trained scientists and psychoacoustics researchers to do properly.

The number of parameters used to fit the experimental data is also rather high, which leads to more questions.

No it's not. As Olive states in his AES paper:

A common problem with regression models is that the models are over-fitted and are not very generalizable to other samples. This can happen when the ratio of observations to number of independent variables falls below 5:1. Ideally there should be 15-20 observations for each independent variable.

The number of samples (70), is perfectly in this ideal range (60-80 samples for 4 variables as are used in the model), which means it is not overly fitted to the particular sample set used and so has good generalisability to other samples outside the data set.

 

[stevenswall]

KEF R3, as old R300, does not quite integrate the additional woofer with the coaxial. Of course, it has improved compared to the old version.

What do you mean by this? Is it like what Buchardt talks about matching the dispersion of the tweeter to the woofer?

 

[b1daly]

I think Linkwitz said that dipole speaker's imaging is "realistic", whether it is perfect or not is not known. The sound comes from the room and walls not from two point sources like box speakers, that is what he means "realistic". I never heard of dipoles, so does not have any comment on perfect imaging.

This ignores that the vast majority of commercial recordings are produced on two point source speakers. Speakers that lack this architecture have very little chance of reproducing the intended effect of the creators of the recording.

 

[b1daly]

It definitely sounds like IMD distortion is what you and Amir are describing as this buzzing sound. I guess I just don't listen as loud as some people because I've never heard it in the LS50 or Q150, I also think these smaller speakers should be used with a high pass in place during listening tests. Playing small speakers loud and full range isn't indicative of how they are used in practice most of the time so I don't think it's fair to rate them badly because of that.

Also just a random comment about the preference score. At one time, I compared the Q100, Q150 and LS50 in my room and I preferred them in that order. The Q100 aren't bad speakers but I can't imagine many people preferring them over the LS50, so I'm also starting to question the preference ratings.

FWIW, my listening is always at a low volume. When listening the the Q150 I had a "what the heck" moment with the distortion sound. I only mentioned it because Amir noted it. It could just be that there are resonant peaks in frequencies that have excess distortion from the original music recording.
---
On a related topic, one of the many things I'm struggling to get my mind around with the measurements of speakers is that frequency response is the dominant factor in user preference, and the related off axis characteristics. My subjective experience is that there are other elements besides frequency response, and I have whiled away many hours EQ'ing speakers, with mixed results.

My current area of ruminating on the subject of what I prefer in listening speakers is that some amount of cabinet resonance is helpful, especially for vintage/classic rock albums. Modern speakers that are made with very dense cabinets and shaped to minimize any diffraction sound incorrect to me, they sound "uncanny." (think of recordings like Back In Black)

Many people champion the quality of speakers "disappearing" but I find this is not appropriate for some music, as it leads to a poorly localized image, that kind of "floats" around.

Paradoxically, speaker cabinet resonance helps localize the speaker in the room, and this actually stabilizes the image as you move around.

Cabinet resonance should show up in a spinorama measurement, but I have no intuition as to what it would look like. I suppose it should cause deviations in the shape of the curves.

Almost all commercial recordings are produced on a pair of point source stereo monitors, and historically, these were resonant boxes. Any imaging involved in the recording is built around this architecture.

 

[maty]

What do you mean by this? Is it like what Buchardt talks about matching the dispersion of the tweeter to the woofer?

That comment came from looking at old measurements of R300 and R3. In Amirm's it is clear that this is not the case. Which graphs are more confiable? I do not know.

I made my PEQ adjustments based on the Zvu measurements, which are consistent with all the others I have / saw on the Q100 -> therefore I consider them more reliable (Q100).

 

[QMuse]

I made my PEQ adjustments based on the Zvu measurements, which are consistent with all the others I have / saw on the Q100 -> therefore I consider them more reliable (Q100).

Although you talk often about your PEQ corrections I don't remember ever seeing the result of them.

Can you pls post corrected in-room measurement of your KEFs?

 

[Dzhaughn]

This ignores that the vast majority of commercial recordings are produced on two point source speakers. Speakers that lack this architecture have very little chance of reproducing the intended effect of the creators of the recording.

Well, that sounds a bit crazy. Name one musical instrument that is a two-point source.

Ok, besides a bagpipe. The exception proves the rule.

Are we to reject headphones? Must we listen to Joy Division on cassette in an AMC Gremlin? (See "24 hour party people" if you don't get the reference. By all means.)

 

[b1daly]

Well, that sounds a bit crazy. Name one musical instrument that is a two-point source.

Ok, besides a bagpipe. The exception proves the rule.

Are we to reject headphones? Must we listen to Joy Division on cassette in an AMC Gremlin? (See "24 hour party people" if you don't get the reference. By all means.)

All acoustic instruments actually radiate in 3 dimensions.

Stereo sound reproduction works based on the phantom image generated in the middle between the two speakers. If you sit right in front of two reasonably placed stereo speakers and close your eyes, you will hear the lead vocal emanating from a space between the two speakers, even though the signal is actually come out of both speakers.

But you can still localize the point sources left right.

 

[tuga]

All acoustic instruments actually radiate in 3 dimensions.

Stereo sound reproduction works based on the phantom image generated in the middle between the two speakers. If you sit right in front of two reasonably placed stereo speakers and close your eyes, you will hear the lead vocal emanating from a space between the two speakers, even though the signal is actually come out of both speakers.

But you can still localize the point sources left right.

Instrument radiation pattern vs. frequency to illustrate your point and why close-mic'ing doesn't work:

And now human voice:

 

[tuga]

For those who can read French this presentation by Jean Michel LeCleac'h is a good read.

 

[bigjacko]

All acoustic instruments actually radiate in 3 dimensions.

Stereo sound reproduction works based on the phantom image generated in the middle between the two speakers. If you sit right in front of two reasonably placed stereo speakers and close your eyes, you will hear the lead vocal emanating from a space between the two speakers, even though the signal is actually come out of both speakers.

But you can still localize the point sources left right.

Some people like sharp image, some people like the sound coming from every where and think that is closer to a live concert. I personally use headphone and do enjoy the sharp image that headphone provides. I do notice that some songs are recorded with stereo image in mind, but there are still people prefer dipoles even the image is not sharp. The big sound stage and wide sweat spot are also advantages of dipoles.

 

[b1daly]

Instrument radiation pattern vs. frequency to illustrate your point and why close-mic'ing doesn't work:

And now human voice:

That's going a bit far to say close mic'ing doesn't work. Vocals are generally recorded with a close mic. Close mic'ing is used extensively for all kinds of instruments, often combined with distant mics.

I find you can usually capture a decent image of an acoustic instrument with two mics. Using more than two starts to get ugly because you've got to mix them down to dual mono streams, and phase issues becoming challenging.

Recording involves a lot of fudging

 

[bigjacko]

Instrument radiation pattern vs. frequency to illustrate your point and why close-mic'ing doesn't work:

And now human voice:

Why does different frequency radiate out in different directions? Is it because different note use different part of the instrument? This is very interesting phenomena.

 

[tuga]

In this link one can visualise how different "real" stereo two microphone techiniques distribute the phantom images of instruments between the speakers ("real" stereo cannot place images outside the speakers):

Visualization of all stereo microphone systems with two microphones
http://www.sengpielaudio.com/HejiaE.htm

 

[Frank Dernie]

Well, that sounds a bit crazy. Name one musical instrument that is a two-point source.

Ok, besides a bagpipe. The exception proves the rule.

Are we to reject headphones? Must we listen to Joy Division on cassette in an AMC Gremlin? (See "24 hour party people" if you don't get the reference. By all means.)

I suppose it depends whether you are attempting to produce a sound which you hope/believe is similar to hearing the performance in a room or to accurately reproduce the recording you have bought/rented.
Both are reasonable goals but only the second is possible to a reasonably high degree. The former is not achieveable or even defineable and will certainly not be achieved by bouncing sound around your own living room, however nice and ethereal it may sound.

 

[QMuse]

Why does different frequency radiate out in different directions? Is it because different note use different part of the instrument? This is very interesting phenomena.

He doesn't have a clue, he's only copying stuff from the Net.

 

[tuga]

That's going a bit far to say close mic'ing doesn't work. Vocals are generally recorded with a close mic. Close mic'ing is used extensively for all kinds of instruments, often combined with distant mics.

I find you can usually capture a decent image of an acoustic instrument with two mics. Using more than two starts to get ugly because you've got to mix them down to dual mono streams, and phase issues becoming challenging.

Recording involves a lot of fudging

Do note that my comments are in reference to documental recording of classical music.

In my opinion close mic'ing has three problems:
- it distorts the timbres
- it picks up mechanical noises which are not audible from a seat in the audience
- it doesn't pick up ambience cues

And when used in a stereo mix they will destroy the illusion of listening to the original soundfield (edit) soundscape. You can fabricate a soundfield (edit) soundscape but it will be something else.