Is purchasing B&W 702 S2’s a bad idea?

[test1223]

This was shown by two binaural blind listening comparisons conducted here in the forum:
Binaural blind comparison test of 4 loudspeakers - I
Binaural blind comparison test of 4 loudspeakers - II

The test was conducted with the B&W 802D and not with the 802N. The estimated in room responses of different B&W speakers are not that similar all might provide the 3kHz dip but I haven't seen such a fitting diffuse field direct sound wiggle with other models. The newer models also might have a to bright sound they seems to be less neutral without equalization but I haven't heard or seen telling measurements of the 802D.
I have listend to the 802N with higher listening distance and not much angled and it sounded neutral to me even the BBC dip was almost none existent which is strange due to its measurements. Some say that the direct sound of the B&W 802N have to be compensated to be linear to get even more convincing neutral sound but I haven't tried this with that speaker. But even without that it sounded very neutral to me more natural than most studio monitors with a listening distance which was slightly to large.

You already mentioned that many headphones and speakers already use a BBC dip in some way and therefore a playback of the recording of a speaker with such a dip might be to much of a dip.

The extreme sound power dip of the B&W speakers, due to the different perception of direct and diffuse sound, cannot be justified with the "Zwicker curve", since it says exactly the opposite of what your PIR example shows.
In the frequency range 2-4kHz the sound pressure level in the diffuse field must be increased by about 2dB compared to the sound pressure level in the direct field to perceive equal loudness. Here is the text to your quoted diagram:

Source: Zwicker, Fastl - Psychoacoustics

Yes it is reversed it might have to do with the room acoustic cues from the recording space, which it typically bigger and which therefore have to be made more "diffuse sounding" or it might have to do with the higher amount of diffuse sound compared with the studio recording. I am not sure what exactly is the cause of this effect and I am more than happy if there are other theories. But for me and others who tested this the light wiggle (about 2dB) makes a speaker with a bigger listening distance than 2m sound more natural. The effect is much easier to observe with good omnidirectional speakers which sound very unnatural without such a compensation. I can only encourage everyone to test this if you have the chance.

Yes, there are many other effects which are overlaid: the original BBC dip was introduced from about 1 to 3kHz there a 2dB dip was proposed to the direct frequency response. The explanation was to compensate the perception of distance which was to near without this. Clearly both effect have some influence on each other.

to justify this would go beyond the scope here...)

Can you please elaborate or at least throw some key word or a link in?

 

[preload]

but B&W exhibits an extreme dip in this range, which may still sound pleasant to many, but is immediately identified as "not right" by many.

This was shown by two binaural blind listening comparisons conducted here in the forum:
Binaural blind comparison test of 4 loudspeakers - I
Binaural blind comparison test of 4 loudspeakers - II

Oh no, please, @ctrl, not this again - these types of binaural re-recording "tests" do not in any way allow you to draw such a conclusion about whether people specifically found the BBC dip "right" or "not right." The internet forum "tests" that you linked, in fact, did not test that hypothesis specifically, and the methodology is basically of "youtube influencer" rigor. For instance, the recordings weren't level matched, and preference rankings mirrored the binaural recording volume/loudness for each "blind" loudspeaker (surprise surprise surprise).

I think people here need to be very careful about what conclusions they draw from experiments to avoid creation and perpetuation of myths.

 

[ctrl]

You already mentioned that many headphones and speakers already use a BBC dip in some way and therefore a playback of the recording of a speaker with such a dip might be to much of a dip.

Did I really? Can't remember.

But it is the case that edge diffraction causes in the frequency range 1.5-5kHz (depended on speaker size) in many speakers a dip in the on-axis frequency response (in combination with a hump in SP/PIR). This on-axis frequency response dip should usually not be compensated, otherwise a hump (or a more pronounced hump) will appear in the SP and PIR frequency response.
Here some examples with normalized on-axis FR (on-axis FR is flat) to show that the on-axis dip needs to remain:

But this has less to do with a deliberate implementation of a "BBC-dip" than with common sense not to remove on-axis edge diffraction without checking different axis FR.

Can you please elaborate or at least throw some key word or a link in?

I don't know if there is a study on "The special role of the 2-4kHz frequency range in loudspeaker design" - guess everyone has their own theory on that or does not regard this frequency range as special.

There are people who claim, based on the "Zwicker curve", that for best sound, the SP frequency response in the range 2-4kHz have to have a hump, because at a normal listening distance the diffuse field is dominant and therefore the speaker has to reflect the "Zwicker curve" (ZC) for best sound. Other say on-axis frequency response should reflect the "equal-loudness-contour" (ELC) for best sound.... and so on.

ELC is part of our everyday life as it is a feature of our hearing. This means that every music recording mixed by a human contains the effects of ELC - which is reflected in the spectrum analysis of recordings (otherwise the spectrum of a recording would look like "white noise"). So no need for a speaker to copy the pattern of the ELC.

The ZC is valid for listening in a diffuse field, i.e. at any point in space the sound arrival from any direction in space is equally likely for all frequency, which applies to very few listening situations. Therefore, there is actually no reason for a home audio speaker to reproduce the ZC via SP or PIR frequency response because at the listening position in a normal listening room there is no ideal diffuse field, always a mixture of direct and reflected sound (additionally depended on the speaker directivity) - the critical distance provides an indication of the ratio of direct to reverberant sound.

Therefore what has been applied to loudspeaker design for decades is valid (IMO), flat on-axis FR and smoothly falling SP response.
If you look at the HRTF arriving at the eardrum, the frequency range of the ear canal resonance (around 2.7kHz, depending on the individual ear canal architecture) is already something special, as this is the "most prominent" resonator of the outer ear.

Source: Signals and Systems for Speech and Hearing, Stuart Rosen, Peter Howell (with colored extensions by me)

I think it is not unreasonable to assume that many people are particularly sensitive to imbalances in the frequency range of the ear channel resonance.

On the image above you can also see that the ear canal resonance is particularly strongly stimulated at 45° (horizontal plane). This could be an indication that, in addition to the direct sound, the first-order lateral reflections could have a particularly strong effect on the sound, since these are hardly attenuated in the sound pressure level (the travel distance is only slightly longer than the direct sound) in a normal listening room.
As example my listening room with first order reflections:


Therefore, when designing loudspeakers, I always make sure that SP and PIR in the 2-4kHz range do not have any hump, no matter how small, and the horizontal side wall reflections as well.
I see no reason for a "BBC dip" in a well designed speaker, but would always prefer a speaker with dip in the 2-4kHz range over one with a hump.

 

[test1223]

Did I really? Can't remember.

But it is the case that edge diffraction causes in the frequency range 1.5-5kHz (depended on speaker size) in many speakers a dip in the on-axis frequency response (in combination with a hump in SP/PIR). This on-axis frequency response dip should usually not be compensated, otherwise a hump (or a more pronounced hump) will appear in the SP and PIR frequency response.
Here some examples with normalized on-axis FR (on-axis FR is flat) to show that the on-axis dip needs to remain:

But this has less to do with a deliberate implementation of a "BBC-dip" than with common sense not to remove on-axis edge diffraction without checking different axis FR.

Yes a very small enclosure or a very wide enclosure is most of the time better since you don't want the diffraction to take place in the most sensible area of our hearing system.
In addition there is the crossover to the tweeter from 2kHz to 3kHz which created a power di dip in almost all cases due to the distance between the drivers.

If you look at the HRTF arriving at the eardrum, the frequency range of the ear canal resonance (around 2.7kHz, depending on the individual ear canal architecture) is already something special, as this is the "most prominent" resonator of the outer ear.

Source: Signals and Systems for Speech and Hearing, Stuart Rosen, Peter Howell (with colored extensions by me)

I think it is not unreasonable to assume that many people are particularly sensitive to imbalances in the frequency range of the ear channel resonance.

Yes if I remember correctly the high difference of the right and left ear in this frequency band is also used to get a better SNR for localisation and to perform the cocktail party effect. I am also certain that a better understanding of the desired tonality of the first sidewall reflection should improve the understanding which speaker sounds good and I am not sure if a curve which pleases the eye is the best fit for a good stereo experience.

I think the B&W 802N provide a better sense of the room of the recording. A lot of people involved in orchestral music use the 802N and say something similar to that claim. Here is a collection of monitoring systems of classical music pros. https://www.audiosciencereview.com/...ers-are-the-classical-music-pros-using.12225/
You often hear the same claim form reviewers that you can better hear the room with the flagship B&W speakers. I think if you apply some eq to a better B&W speaker it might be a different experience and a endless dispute like using a direct sound or diffuse sound equalized headphone. In one of the older stax amps you can switch between both eq settings and both sounded right but different with any good recording.

 

[ferrellms]

Hi. I’m considering purchasing B&W 702 S2 speakers. I heard them at Best Buy recently, then again at a local shop. I kept coming back to them over others. Then, while researching them on this forum, I saw a post from someone about having wasted money on them prior to knowing the importance of measurements. That surprised me. Is that a known thing? I’d love to know more so I don’t waste money. Isn’t hearing them all you need (versus measuring)?

I appreciate any advise you can give!
-Justin

Well, you are reading this forum. Why not look at the test results? B&Ws are pretty poor. I liked my B&W 805s until I heard a pair of Mackie monitors that cost 1/2 as much, included amps, and sounded quite a bit less colored.

 

[ferrellms]

Unless one is specifically looking for a speaker which accurately reproduces the recorded signal, measurements are only useful if one is able correlate measured performance with one’s preference.

So the best option in my view is to test the speakers at home in their intended location with one’s favourite music.

As others have mentioned, some aspects of measured performance indicate if a speaker can be tuned to taste or whether it shows major flaws in performance or how difficult it is to drive.
But some people actually enjoy such flaws and many, perhaps most, are not using any kind of EQ.

A good starting point would be to visit a dealer and compare the BnWs to a few alternatives.

I will go out on a limb and disagree with "whatever you like is best" regardless of measurements. It is easy to say "whatever you like", but unhelpful. In the case of speakers there is such variation in sound and program content that listener preferences may differ in a lot of ways and may not be repeatable. You may indeed never hear anything really good sound (uncolored) at all with which to compare the BWs.

Do you want to hear what is on the recording? Get one of the powered monitors recommended in this forum. "Listener preference" is a rabbit hole - measurements are better and you will end up enjoying really accurate speakers. You will hear all of the program material (good and bad).

 

[Bob from Florida]

Traditional wedding question by usher - Bride or Groom?

Redneck wedding - Chevy or Ford?

Audiophile wedding - Objective or Subjective?

 

[tuga]

I will go out on a limb and disagree with "whatever you like is best" regardless of measurements. It is easy to say "whatever you like", but unhelpful. In the case of speakers there is such variation in sound and program content that listener preferences may differ in a lot of ways and may not be repeatable. You may indeed never hear anything really good sound (uncolored) at all with which to compare the BWs.

Do you want to hear what is on the recording? Get one of the powered monitors recommended in this forum. "Listener preference" is a rabbit hole - measurements are better and you will end up enjoying really accurate speakers. You will hear all of the program material (good and bad).

I avoid helping or suggesting altogether. People should go out and listen for themselves. That is how preference works.
Let's say I suggest a "transparent" monitor and the person ends up disapointed. What have I gained, and more importantly what has he gained?
That's nonsensical.

The ultimate goal of a system is to provide listening enjoyment to its owner, it's about what I think it's right or wrong, if he is getting good value for money or being conned, if 'I' wan't to hear what's on the recording...

 

[ctrl]

Yes a very small enclosure or a very wide enclosure is most of the time better since you don't want the diffraction to take place in the most sensible area of our hearing system.
In addition there is the crossover to the tweeter from 2kHz to 3kHz which created a power di dip in almost all cases due to the distance between the drivers.

The DI dip due to the increasing beam of the bass-midrange driver can be partially compensated, so that no or only a "shallow" dip occurs in the SP and PIR response.

There are possibilities to avoid a dip in the sound power response (SP). One can use a "constant power crossover" like third order Butterworth (instead of the typical "constant voltage crossover" like Linkwitz-Riley) or "smear" the SP/PIR dip with a second order crossover...

It is also often the case that edge diffraction causes a hump in the SP. Then it makes sense to place the crossover frequency in this frequency range to get a more smooth SP/PIR response or one with a shallow dip.

I am also certain that a better understanding of the desired tonality of the first sidewall reflection should improve the understanding which speaker sounds good and I am not sure if a curve which pleases the eye is the best fit for a good stereo experience.

One should not completely ignore the life's work of Toole and Olive. Both in the evaluation of loudspeakers (work of Toole, Olive, et al.) and headphones (work of Olive, et al.), the frequency range 2-4kHz does not stand out as "special".

If the curves are pleasing to the eye, then this is a byproduct of scientific work

The correlation of listener preference and flat on-axis, listening window and evenly sloping sound power and predicted in-room response is there.
One can criticize in detail (keywords: lateral reflections, monophonic versus stereophonic), but more extensive work is not known to me.
This has already been discussed endlessly here in the forum, so I'm sure everyone has their own opinion on the subject.

But even if you look at the target curves for headphones, the 2-4 kHz frequency range is quite inconspicuous and does not show a clear frequency response drop like B&W uses for its speaker models. Etymotic IEM target is based on a modified diffus field HRTF and Harmann IEM target based on the "transfer" of a reference speaker in a listening room and listener preference tests.

Again no special treatment of the 2-4kHz range. Both IEM target curves reproduce the ear canal resonance without extra dip in the range 2-4kHz.


Contrast this with a measurement of the B&W 802 Diamond (HifiTest 1-2012, first image below) speaker, from the list of B&W loudspeakers linked above that are said to be popular in recording studios. Unfortunately, we do not have complete measurements of the B&W reference series. But all incomplete measurements show a partly considerable drop of the sound power frequency response in the range 1.5-5kHz.

I think the B&W 802N provide a better sense of the room of the recording. A lot of people involved in orchestral music use the 802N and say something similar to that claim. Here is a collection of monitoring systems of classical music pros. https://www.audiosciencereview.com/...ers-are-the-classical-music-pros-using.12225/
You often hear the same claim form reviewers that you can better hear the room with the flagship B&W speakers.

Yes, this impression might be directly related to the sharp drop in sound power and on-axis frequency response in the 1.5-5kHz range.

Blauerts directional bands could play a big role in this. According to Blauert there are certain frequency ranges where it hardly matters if the sound comes from the front, back or top, most people will always mention the same sound source direction.

The frequency range around 3 kHz is perceived by most people as coming from the front, regardless of whether the sound comes from the front, back or top.
If this frequency range is lowered in a speaker, then everything that is "present, to the front" is weakened and shifted "to the rear".
Better and with more details I tried to explain this here.

 

[test1223]

But even if you look at the target curves for headphones, the 2-4 kHz frequency range is quite inconspicuous and does not show a clear frequency response drop like B&W uses for its speaker models. Etymotic IEM target is based on a modified diffus field HRTF and Harmann IEM target based on the "transfer" of a reference speaker in a listening room and listener preference tests.

Again no special treatment of the 2-4kHz range. Both IEM target curves reproduce the ear canal resonance without extra dip in the range 2-4kHz.

I don't understand where you assumed to find a dip?? I was referring to the following difference (or a similar difference since it is very difficult to get a real diffuse sound field and a direct sound field which is flat all over the head area).

(https://www.headphonesty.com/2020/04/harman-target-curves-part-3/)
And as you can see the Harman target curve is a mixture of both curves. If you equalize a headphone to a more diffuse target curve or a more direct sound curve you typically will like both curves. It depends on the recording and on your preference what you like better but most of the time you only can say that it is different and both sound good with a different focus.

The DI dip due to the increasing beam of the bass-midrange driver can be partially compensated, so that no or only a "shallow" dip occurs in the SP and PIR response.

There are possibilities to avoid a dip in the sound power response (SP). One can use a "constant power crossover" like third order Butterworth (instead of the typical "constant voltage crossover" like Linkwitz-Riley) or "smear" the SP/PIR dip with a second order crossover...

It is also often the case that edge diffraction causes a hump in the SP. Then it makes sense to place the crossover frequency in this frequency range to get a more smooth SP/PIR response or one with a shallow dip.

Yes but you either sacrifice some early reflection fr smoothness or power di smoothness. Getting both right with a typical midwoofer tweeter arrangement is only possible if the distance is atypical small or you use a coaxial driver.

One should not completely ignore the life's work of Toole and Olive. Both in the evaluation of loudspeakers (work of Toole, Olive, et al.) and headphones (work of Olive, et al.), the frequency range 2-4kHz does not stand out as "special".

If the curves are pleasing to the eye, then this is a byproduct of scientific work

The correlation of listener preference and flat on-axis, listening window and evenly sloping sound power and predicted in-room response is there.
One can criticize in detail (keywords: lateral reflections, monophonic versus stereophonic), but more extensive work is not known to me.
This has already been discussed endlessly here in the forum, so I'm sure everyone has their own opinion on the subject.

I didn't ignore the research at all. But I am aware that the research is far away from being completed. A lot of people here don't see this fact. I haven't seen meaningful research e.g. on the alternation of the frequency response of one individual first reflections while it don't affect all other influences. There is plenty more research to do where only one influence is altered (listening distance, wave front curvature, amount of diffuse sound, ...). Since in every part the none linear head related transfer functions are involved and an inherently "insuffizient" stereo system it would be more than stunning if every single curve should be a straight line in all relevant different listening room scenarios.
Since it is very hard to conduct such experiments we have to wait a few years to get such research and insides. Maybe the advancement in artificial 3D headphone sound rendering will bring all this research to us.

Yes, this impression might be directly related to the sharp drop in sound power and on-axis frequency response in the 1.5-5kHz range.

Blauerts directional bands could play a big role in this. According to Blauert there are certain frequency ranges where it hardly matters if the sound comes from the front, back or top, most people will always mention the same sound source direction.

The frequency range around 3 kHz is perceived by most people as coming from the front, regardless of whether the sound comes from the front, back or top.
If this frequency range is lowered in a speaker, then everything that is "present, to the front" is weakened and shifted "to the rear".
Better and with more details I tried to explain this here.

The tendency of the Blauert bands is definitely also one piece of the puzzle which is also directly connected to the HRTFs, but you have to be aware that the findings are only true in the median plane with a distinct experimental setup and the probabilities are even not that high with this setup. Or brain can't be tricked that easily if there is also sound from the horizontal plane (like stereo speakers). E.g. the panning of sound in the median plane is a big problem with binaural headphone 3D sound, since your brain is very capable to don't get fooled with artificial median plain localisation.

 

[ctrl]

I don't understand where you assumed to find a dip?? I was referring to the following difference (or a similar difference since it is very difficult to get a real diffuse sound field and a direct sound field which is flat all over the head area).

I refer to the starting point of our discussion and the B&W practice of extrem on-axis and SP dip around 2.5kHz, where you said that a dip around 2.5kHz in on-axis FR would sound more "natural" and you justified this with the different perception of free and diffuse field sound as shown by the "Zwicker curve" (here):

To my experience also these tracks should sound better with the B&W. As I claimed earlier the circle of confusion is on part but the BBC dip around 2.5kHz is also a part which is derived by testing the playback of a neutrally recorded voice. And the dip is a part of the perception difference between direct sound and diffuse sound. The loudspeaker at a certain distance where some amount of diffuse sound is created needs to adjust the sound power for that (at least the 2.5kHz part).

Which, as already said, is not correct, because according to the Zwicker curve with increasing diffuse sound in the 2-4kHz range should be more of a hump (in SP response) not a dip.

The free and diffuse field HRTF you showed is nothing more than another way of representing the "Zwicker curve" results.

There are different representations of the two curves (free and diffuse HRTF) and also their measurement can be done in different ways (in-ear mic in humans or dummy head ear,...), therefore the frequency ranges do not match exactly with the results of Zwickers equal-loudness for free and diffuse field sound (Zwicker used narrow band noise as sound source and had the test participants set equals loudness manually).

In the range 0.8-2kHz the diffuse field HRTF has to be lowered for "equal loudness" (compared to free field),
in the range 2-4kHz slightly raised
and in the range around 8kHz strongly lowered - same as the Zwicker curve predicts too.
So here again the transition from the free to the diffuse field means an increase of SP in the frequency range around 2.5kHz for equal loudness sensation.

For well designed speakers there is no reason for such an extreme on-axis FR dip (and SP/PIR dip) around 2.5kHz as B&W practices.
Hence the examples in my previous post (preferred behavior of loudspeaker and headphone target curves of different origin), to show that there is no evidence that an extreme on-axis FR dip (and SP dip) around 2.5kHz is sound wise advantageous.
Personally, I don't mind a shallow SP/PIR dip in the 2-4kHz range.

And as you can see the Harman target curve is a mixture of both curves. If you equalize a headphone to a more diffuse target curve or a more direct sound curve you typically will like both curves. It depends on the recording and on your preference what you like better but most of the time you only can say that it is different and both sound good with a different focus.

Agree with you, HRTF and listening habits are individually different and the Harman target curve represents the "average" preferred target curve. Sean Olive does not claim otherwise, as can be read here - 64% of listener are fine with the harman target curve.

The interesting thing is that the basis for the Harman headphone target curves is not a simply a mix of free and diffuse HRTF, but a dummy head recording of a reference speaker in a slightly treated listening room, equalized to the Harman in-room target curve for speakers. Plus small corrections of the recorded HRTF by listener preference tests.
==>
If you are really picky, you could say that there is a small 1 dB dip in the 2-3 kHz range of the equalized in-room response of the reference speaker (black curve). Which is in line with my personal experience of not allowing an SP or PIR hump in this frequency range under any circumstances.
But nothing as drastic as B&W does.

Since in every part the none linear head related transfer functions are involved and an inherently "insuffizient" stereo system it would be more than stunning if every single curve should be a straight line in all relevant different listening room scenarios.

Nobody claims that it applies to speaker in "all relevant different listening room scenarios". It is, roughly speaking, meant for "average" living rooms with a reverberation time around 0.4s.
A straight line is supposed to represent "only" for on-axis and listening window response. SP and PIR evenly downward slope, DI evenly upward.

but you have to be aware that the findings are only true in the median plane with a distinct experimental setup and the probabilities are even not that high with this setup. Or brain can't be tricked that easily if there is also sound from the horizontal plane (like stereo speakers).

In my post I linked to another post of mine where it describes in more detail how this affects sterophonic listening. To put it very briefly:

With loudspeaker stereophony (i.e. in the horizontal plane) the original direction-determining bands v (front), h (back) and o (top) of the median plane can be easily reinterpreted as two directionless hearing sensations “ present ” and “ diffuse ”. This is also noteworthy for the equalizer settings for the sound processing in stereophony and the surround sound of broadband signals. Where:

• front v = present in the sound , close, direct, superficial - this can be achieved by increasing the frequencies 300 to 400 Hz and 3 to 4 kHz as well as lowering frequencies by 1 kHz.
• back h (and above o) = diffuse , distant and spatial in the sound - this can be achieved by increasing the frequencies by 1 kHz

Source: directional bands introduced by J. Blauert

 

[test1223]

I refer to the starting point of our discussion and the B&W practice of extrem on-axis and SP dip around 2.5kHz, where you said that a dip around 2.5kHz in on-axis FR would sound more "natural" and you justified this with the different perception of free and diffuse field sound as shown by the "Zwicker curve" (here):

I worded it in a nice manner the first time but that causes more confusion I guess. You claimed that there is no such huge 2.5kHz dip in the headphone direct vs. diffuse sound field equalization like in the B&W direct frequency response which is a wrong claim! There is an about 4dB dip in the example of the direct vs. diffuse equalization of a headphone and there is a dip of about 3dB in the measurement of the B&W 802N.

Which, as already said, is not correct, because according to the Zwicker curve with increasing diffuse sound in the 2-4kHz range should be more of a hump (in SP response) not a dip.

I have answered this already. Your claim: not correct is based on your interpretation what should be correct! Please refer to my earlier answer otherwise we are moving in circles.

Yes it is reversed it might have to do with the room acoustic cues from the recording space, which it typically bigger and which therefore have to be made more "diffuse sounding" or it might have to do with the higher amount of diffuse sound compared with the studio recording. I am not sure what exactly is the cause of this effect and I am more than happy if there are other theories. But for me and others who tested this the light wiggle (about 2dB) makes a speaker with a bigger listening distance than 2m sound more natural. The effect is much easier to observe with good omnidirectional speakers which sound very unnatural without such a compensation. I can only encourage everyone to test this if you have the chance.

I like to add to my earlier answer that the side wall reflections might be an important part where a wiggle might be beneficial since they can add to the envelopment but provide the tonality of the direct sound (more or less).

Agree with you, HRTF and listening habits are individually different and the Harman target curve represents the "average" preferred target curve. Sean Olive does not claim otherwise, as can be read here - 64% of listener are fine with the harman target curve.

Yes as I already wrote there is no simple right or wrong when it comes to diffuse field headphone eq or direct sound headphone eq. My claim is that the same is true for good speakers. Some curves like the power frequency response (or others) might sound better or different but also right with a light wiggle at about 3kHz. Which curve provides the best results overall might be dependent on many different factors like the listening distance or the room information of the recording space. To investigate this behavior would be a step forward.

Personally, I don't mind a shallow SP/PIR dip in the 2-4kHz range.

This is exactly the point. To my experience a small dip did increase the preference. One or two dB is about perfect B&W might exaggerate it but I didn't feel that way with listening with the 802N. I can't speak for all other models. Without that dip it is still okay but it already moved toward worse and a peak is not good which is more or less consensus.

Nobody claims that it applies to speaker in "all relevant different listening room scenarios". It is, roughly speaking, meant for "average" living rooms with a reverberation time around 0.4s.
A straight line is supposed to represent "only" for on-axis and listening window response. SP and PIR evenly downward slope, DI evenly upward.

Then why are you that certain that the response of the 802N with higher listening distance in a not heavily damped room should sound wrong?

A straight line is supposed to represent "only" for on-axis and listening window response. SP and PIR evenly downward slope, DI evenly upward.

I don't want to challenge a tilt which isn't controversial at all. I want to discuss the assumption that a light wiggle e.g. like the Zwicker curve can bring a higher preference (in specific scenarios).

In my post I linked to another post of mine where it describes in more detail how this affects sterophonic listening. To put it very briefly:


Source: directional bands introduced by J. Blauert

Please have a closer look at the experimental setup where the Blauert bands where derived of (fixed head, anechoic chamber, no visual cues). The results can't be transferred 1 to 1 to a different set-up (stereo in a reflective room) since there are many different mechanisms which determine the localisation of a sound source and which "overrule" the Blauert bands. While some tendencies might be true, the influence of reflection cues and horizontal plane localisation mechanisms play an important role to not misinterpret tonal differences with localisation differences.

 

[ctrl]

There is an about 4dB dip in the example of the direct vs. diffuse equalization of a headphone and there is a dip of about 3dB in the measurement of the B&W 802N.

Yes, no one disputes that (up to four db difference for equal loudness in 2-4kHz range between free and diffuse field). But I'm trying to explain to you that the argumentation with free and diffuse field is not an argument to put a dip in the 2-4kHz range of the on-axis or sound power (SP) response of a loudspeaker.

Assuming a loudspeaker sounds optimal in a damped studio in the near field - free field dominant.
Then, with the same loudspeaker in a less damped living room at a normal listening distance, the diffuse field would be dominant and up to 4 dB less sound pressure level would reach the eardrum in the frequency range 2-4 kHz.
For the same tonality as in the studio, the frequency response in the 2-4 kHz range would have to be increased up to 4dB if one argues with free/diffuse fields.
Which, however, would contradict both of our experiences of speaker in "normal" listening rooms (where a small dip in the 2-4kHz range is advantageous in many cases).

This is exactly the point. To my experience a small dip did increase the preference. One or two dB is about perfect B&W might exaggerate it but I didn't feel that way with listening with the 802N. I can't speak for all other models. Without that dip it is still okay but it already moved toward worse and a peak is not good which is more or less consensus.

I agree with you one hundred percent, a slight dip in the sound power (SP) and predicted in-room response (PIR) in the 2-4kHz range is often beneficial - whether this is achieved by a slight dip in the on-axis FR or via special speaker/XO design reducing output of the off-axis FR in the 2-4kHz range.

I don't know the big B&W floor-standing speakers, only binaural recordings of these speakers. In all recordings that I know, you can hear the slightly too much on-axis/SP dip immediately. Of course, this does not mean that there can be environments in which this is not the case.

The results can't be transferred 1 to 1 to a different set-up (stereo in a reflective room) since there are many different mechanisms which determine the localisation of a sound source and which "overrule" the Blauert bands. While some tendencies might be true, the influence of reflection cues and horizontal plane localisation mechanisms play an important role to not misinterpret tonal differences with localisation differences.

That a one-to-one transfer is possible is explicitly not claimed. Therefore, a directional band determined by Blauert with the localization "front" becomes the impression "present" during the transfer for stereo listening.

The whole thing goes back to the electrical engineer and recording engineer Eberhard Sengpiel (two times Grammy award winner). Sengpiel had compared Blauerts directional bands (valid for the median plane) with the equal-loudness contour and found "similar" frequency ranges.

directional bands: vorn=front hinten=rear oben=above

Directional bands as rectangular frequency ranges and plotted equal-loudness contour at 80 phon:

He commented (1995):

It is not coincidental that the maxima and minima of these curves lie in the middle of Blauert's bands. So far I am not aware of any literature that points this out. In stereo hearing, the directional localization is no longer in front, but one has to choose the appropriate words present and diffuse for it.

In loudspeaker stereophony, present and diffuse no longer mean direct directional localization, but rather proximity or distance. In the frequency ranges where the hearing is more sensitive, the level must be amplified to appear more present.

A summary can be found at "The importance of Blauert's directional bands for sound recording" - unfortunately the text is only available in German.

 

[ctrl]

Another small remark about the frequency range around 2-4kHz.

That this range requires special attention B&W has correctly recognized (only, as already said, IMO the crossover is tuned too extreme) and if you look at the development of the Harman OE target curves, then it is confirmed that this frequency range is "special".

Have collected from various sources the Harman target curve for OE headphones 2017 and 2018 and compared each with the 2013 version:

If we now normalize to the 2013 target curve and compare the deviations of the 2017 and 2018 target curves with it, we notice that in the frequency range 2-4kHz the target sound pressure level was lowered particularly strongly (ignore >12kHz):
2013 vs 2017 2013 vs 2018

Unfortunately, I don't have access to Harman's original sources, so I can't guarantee that the curves shown are one hundred percent accurate.

 

[pogo]

I appreciate any advise you can give!

Link

 

[ferrellms]

I avoid helping or suggesting altogether. People should go out and listen for themselves. That is how preference works.
Let's say I suggest a "transparent" monitor and the person ends up disapointed. What have I gained, and more importantly what has he gained?
That's nonsensical.

The ultimate goal of a system is to provide listening enjoyment to its owner, it's about what I think it's right or wrong, if he is getting good value for money or being conned, if 'I' wan't to hear what's on the recording...

Well, just because somebody likes something does not mean it is good! If a suggestion steers somebody in the direction of real quality, make the suggestion. He will thank you when he hears various systems other than his and realizes his is better.

 

[tuga]

Well, just because somebody likes something does not mean it is good! If a suggestion steers somebody in the direction of real quality, make the suggestion. He will thank you when he hears various systems other than his and realizes his is better.

Maybe he will never realise it's better, and won't like the sound. What then?

 

[Armance]

Just to share my little experience with the B&W 702 Signature.

I went from Naim / Sonus Faber (which I like a lot) to a McIntosh / 702 Signature.

The first days, it was a nightmare : harsh sound & painful trebles. I concluded that I have made a very poor decision and was willing to sell them. It was a demo pairs, so I supposed that they have been used and brunt…Which they were not. I used a CD (Jean-Marie Reynaud, Magical CD) and the musicality was better. The Sonus Faber are very friendly relatively to the room placement, the B&W are not. They sound was much much better once I fined a good positioning.

The sound signature can be described as a mix of openness, details and dynamic. Compared to Sonus Faber, they are less « emotional » and definitively less colored.

So my conclusion is that they can be a tremendous pair of speakers if :

1- The most important : you like the sound signature.

2- You have a powerful amp. The are presented as 8 ohms but they behave like 6 ohms speakers in an important part of the spectrum with some downs to 3.1 Ohms.
If the amp is not powerful, they will sound awful and thin.
The sound signature of the amp must be complementary; even if i didn’t listen to them with an Accuphase amp, I am pretty sure that they will sound awful with the amp of this brand.

3- Be very careful with the placement and the room. The Sonus Faber performed very well in a room with big windows without curtains. The B&W required me to install a heavy curtains to perform very well.

So, they can be fantastic or awful. You have to check criteria before the investment.

Best regard