Blind Listening Test 2: Neumann KH 80 vs JBL 305p MkII vs Edifier R1280T vs RCF Arya Pro5

[goat76]

I am not a newbie in this field, but came for me to another conclusion.
Indeed, sometimes it is better not to excite the room with too much low-frequency energy.

I don't think you are a newbie, I just think you have never got the implementation of subwoofers right which is hard to get right.

No one will prefer a boomy sound caused by problematic room acoustics, and no one will like the sound of subwoofers that reveal themselves as obvious separated sound sources (a sound that is not attached seamlessly to the main speakers). But with those problems solved, I'm quite sure everyone, including you, will most likely prefer the full-range sound most main speakers can't reproduce by themselves.

 

[computer-audiophile]

I just think you have never got the implementation of subwoofers

I find it difficult to project one's own experiences or knowledge onto other people.
I myself am rather a friend of point-source loudspeakers (as an ideal), which is why I have experimented with full-range loudspeakers for decades (besides other things)

About 10 years ago I wrote the following about the subwoofers in my former audio blog. (Automatic translation which is not perfect)

"Sounding out the sonic shallows
My small Neumann monitors already produce a deep and powerful bass on their own. Nevertheless, I had the idea of relieving them with a subwoofer, so that the drivers don't have to make so much excursion and can therefore play even more cleanly in the midrange. However, the experiment was disappointing. Without the extreme low bass, the sound was more coherent and stable in most recordings.

Only on very few pieces of purely electronic music did the low bass sound reasonably clean, but at the same time I got more trouble with room modes.

On classical concert recordings, wobbling and rumbling background noises suddenly become audible, which are otherwise much less annoying. Vinyl Records don't work at all, there's only rumbling and groove noise at the low bottom. Especially with recordings of organ concerts, I had expected a better reproduction in the lower registers, but even here the sound is not obviously clearer. What is found at the lower end of the frequency band on recordings made with microphones is apparently often just sonic rubbish.

I hardly ever listen to modern electronic dance music (if that's the right term), but with the pulsating beats of juan atkins & moritz von oswald the sub was one of the best fits and was even fun for a few bars."

 

[Newman]

I am not a newbie in this field, but came for me to another conclusion.
Indeed, sometimes it is better not to excite the room with too much low-frequency energy.

Similarly to goat’s comment, plus a couple of other factors. If you implement full bass extension and feel something’s wrong, make sure you are:-

1. Ironing out bass peaks with EQ.
2. NOT boosting bass cancellation dips with EQ.
3. Not trying to get one sub to sound good in multiple listening positions. Get more subs if you want that.
4. Tweaking the LEVEL of bass to taste. Olive has written about the range of perceptual preferences in bass LEVEL, so tweak that to taste, rather than think that FULL EXTENSION is the problem, which it shouldn’t be.
5. Beware the Treble Trap. People without full bass extension can find themselves with a slightly softened treble output without realising it, because flat treble can sound sharp without full bass extension for balance and naturalness. So, when such a system has subs added, the sound is perceived as excessively bassy, when it’s not the bass’s fault but needs flat treble for balance. Not saying this is so for you, but it’s a factor that some subwoofer-averse people need to consider.

cheers

 

[computer-audiophile]

Similarly to goat’s comment, plus a couple of other factors. If you implement full bass extension and feel something’s wrong, make sure you are:-

1. Ironing out bass peaks with EQ.
2. NOT boosting bass cancellation dips with EQ.
3. Not trying to get one sub to sound good in multiple listening positions. Get more subs if you want that.
4. Tweaking the LEVEL of bass to taste. Olive has written about the range of perceptual preferences in bass LEVEL, so tweak that to taste, rather than think that FULL EXTENSION is the problem, which it shouldn’t be.
5. Beware the Treble Trap. People without full bass extension can find themselves with a slightly softened treble output without realising it, because flat treble can sound sharp without full bass extension for balance and naturalness. So, when such a system has subs added, the sound is perceived as excessively bassy, when it’s not the bass’s fault but needs flat treble for balance. Not saying this is so for you, but it’s a factor that some subwoofer-averse people need to consider.

cheers

All correct and also good information for all those who do not yet know. Thanks!

 

[goat76]

I find it difficult to project one's own experiences or knowledge onto other people.
I myself am rather a friend of point-source loudspeakers (as an ideal), which is why I have experimented with full-range loudspeakers for decades (besides other things)

About 10 years ago I wrote the following about the subwoofers in my former audio blog. (Automatic translation which is not perfect)

"Sounding out the sonic shallows
My small Neumann monitors already produce a deep and powerful bass on their own. Nevertheless, I had the idea of relieving them with a subwoofer, so that the drivers don't have to make so much excursion and can therefore play even more cleanly in the midrange. However, the experiment was disappointing. Without the extreme low bass, the sound was more coherent and stable in most recordings.

Only on very few pieces of purely electronic music did the low bass sound reasonably clean, but at the same time I got more trouble with room modes.

On classical concert recordings, wobbling and rumbling background noises suddenly become audible, which are otherwise much less annoying. Vinyl Records don't work at all, there's only rumbling and groove noise at the low bottom. Especially with recordings of organ concerts, I had expected a better reproduction in the lower registers, but even here the sound is not obviously clearer. What is found at the lower end of the frequency band on recordings made with microphones is apparently often just sonic rubbish.

I hardly ever listen to modern electronic dance music (if that's the right term), but with the pulsating beats of juan atkins & moritz von oswald the sub was one of the best fits and was even fun for a few bars."

I like those rumbling background noises like an occasional unintended bump to the floor and things like that, I think it adds some realness and atmosphere to the recording and can be one of those small important cues that add a three-dimensional feeling to the recorded musical event.

A similar thing that can add to the atmosphere is when the low rumble of a foot pedal from a close-up recorded piano is heard, mechanical sounds like that can be argued not to be part of the musical tones of the instrument, but they sure are part of what one hears if being there at the same distance.

And apart from little things like that, most speakers can't reproduce the last octaves of the instruments which makes their musical tones incomplete.

But that's me and I respect your view on the subject if the bass content mostly contains annoying things for you.

 

[computer-audiophile]

I have tried to differentiate some sub-bass problems as they occur in real recordings, but this does not seem to go down well.

Perhaps we should rather return to the topic in which practical listening experiences with Neumann KH 80 / JBL 305p MkII / Edifier R1280T / RCF Arya Pro5 ore one of these, which are discussed. I did my part.

 

[WillBrink]

Great work and effort. Much appreciated! The rotating table is a thing of brilliance! I do wonder if listening to those speakers in pairs would fair any differently.

 

[Guido1427]

Lovely to see an attempt at a decent experimental set up. Just a note on the statistics you used - you used ANOVA and I'm not sure that's 100% correct - it'll approximate the correct result but wouldn't stand up to peer review (at least not in my discipline). That's because ANOVAs are parametric tests that make assumptions about the underlying data that aren't actually supported when you use a likert-type preference scale. It appears that the speakers were rated on a 1-10 scale, and so the units are ordinal (in that they are ordered but the size differences between the numbers might vary - for example 2 is larger than 1 and three is larger than 2, BUT 2 is not necessarily twice the size of 1, and 3 is not necessarily three times larger than 1 etc). The correct statistics in this case are ordinal logistic regressions (you can use package ordinal by Christensen in program R for this https://cran.r-project.org/web/packages/ordinal/ordinal.pdf). These can be performed as repeated measures analysis if needed. I work a LOT with survey data in my day job, and this is the correct statistical method for data derived from preference scores. I'll stop being a smart a*se now - thanks again for all the reviewing. Much appreciated!

The scale could be interpreted as a measurement on intervalscale-level. The sample is the problem. However - a great idea to do something like that.

 

[Floyd Toole]

I hope this is not too much to ask, but Is there a correlation in blind listening tests when using a single loudspeaker on axis to compare it to another one when:

For example, loudspeaker A is flat on axis, also designed to be well behaved off axis, whereas loudspeaker B is only flat on axis but much worse off axis than loudspeaker A? Would a listener be able to discriminate the flaws of loudspeaker B if the setup is such that direct sound is dominant over the reflections? Would a chance to do so be highly dependent on the program material?

In other words, is there a chance that we give a higher score to loudspeaker B because of, say a little bit more bass extension, and only when we use stereo pairs we find out that we've made a poor selection? Would evaluating loudspeakers in mono, whilst revealing problems with resonances and flaws in on axis behavior, also be able to hide some problems off axis? Is spinorama the only reliable way to predict in room loudspeaker behavior?

Earl Geddes emphasizes the importance of loudspeaker polar response and having a flat DI at some degree off axis, ranging from 700-7000 Hz. He explains it in the video here:

Can you comment on this?

As my earliest JAES (1985-86) papers showed, listeners in double-blind evaluations preferred loudspeakers with flat and smoothish on-axis (direct sound) frequency responses. Beyond that, those that were preferred exhibited the smoothest off-axis (reflected sound) responses. None of those loudspeakers were horns, so the DI varied with frequency. Now that we have loudspeakers with excellent constant directivity horns, it is seen that these too receive elevated sound quality ratings (see the cover of my book). I don't know that we have sufficient data to be able to define the tolerances on variability in directivity, but it seems fairly obvious that smooth gradual changes would be preferable to uneven, changes - aiming for reasonable timbral matching of direct and reflected sounds - remember that the reflecting surfaces can modify the spectra of reflected sounds. BTW, sound power is a metric that is of questionable merit in conventional rooms with reverberation times of the order of 0.3- 0.5 s - i.e. acoustically relatively "dead". Early reflections dominate, to the point that in spinorama data, the early-reflections curve alone is a good approximation to a steady-state room curve at the listening location. All of this is in my journal papers and book.

So, to answer your question, the direct sound always has perceptual precedence over reflected sounds - that is why it is called the "precedence effect" or "law of the first wavefront". If there is a timbral mismatch between it and the later arriving reflections, the difference, according to double-blind listening tests, is audible and detrimental to the sound quality.

The audible effects of early reflections in typical listening rooms was investigated in: Olive, S. E. and Toole, F. E. (1989). “The Detection of Reflections in Typical Rooms”, J. Audio Eng. Soc., 37, pp. 539-553. which is summarized in Chapter 7 of the 3rd edition. All the data are supported by blind listening tests - something that is glaringly absent from most discussions of these topics. Personal opinions are just that. Interestingly, the data show that sensitivity to early reflections reaches a peak around 10 ms, being lower before and after. The "image shift" effect is really quite small, but still audible. One has to wonder whether the "precise" horizontal positioning of a phantom image is truly that important. For me it isn't. However, the beneficial effect of some amount of horizontally reflected sound - adding spaciousness - does seem to flatter a high percentage of recordings, especially those with hard panned (mono) left and right images. But I grant you, this can be personal, as some listeners strive for "pinpoint" imaging, something not always heard in real life performances.

I know both David Griesinger and Earl Geddes, and respect their contributions to the field. To comment on Earl's video would be a massive undertaking. I would like to think that reading my book would settle the issue, as we looked at very much the same concerns. I and my colleagues over the years had the advantage of facilities within which we could acquire abundant, accurate anechoic measurements, as well as listing rooms with which we could test various hypotheses as to what was audible in unbiased double-blind listening. While at Harman we attempted to acquire a sample of one of Earl's loudspeakers but none were available when we enquired. However, the notion of neutral direct sound and similarly neutral reflected sounds - his argument - is one with with I have no disagreement. We just have different ways of quantifying the parameters.

 

[Omid]

As my earliest JAES (1985-86) papers showed, listeners in double-blind evaluations preferred loudspeakers with flat and smoothish on-axis (direct sound) frequency responses. Beyond that, those that were preferred exhibited the smoothest off-axis (reflected sound) responses. None of those loudspeakers were horns, so the DI varied with frequency. Now that we have loudspeakers with excellent constant directivity horns, it is seen that these too receive elevated sound quality ratings (see the cover of my book). I don't know that we have sufficient data to be able to define the tolerances on variability in directivity, but it seems fairly obvious that smooth gradual changes would be preferable to uneven, changes - aiming for reasonable timbral matching of direct and reflected sounds - remember that the reflecting surfaces can modify the spectra of reflected sounds. BTW, sound power is a metric that is of questionable merit in conventional rooms with reverberation times of the order of 0.3- 0.5 s - i.e. acoustically relatively "dead". Early reflections dominate, to the point that in spinorama data, the early-reflections curve alone is a good approximation to a steady-state room curve at the listening location. All of this is in my journal papers and book.

So, to answer your question, the direct sound always has perceptual precedence over reflected sounds - that is why it is called the "precedence effect" or "law of the first wavefront". If there is a timbral mismatch between it and the later arriving reflections, the difference, according to double-blind listening tests, is audible and detrimental to the sound quality.

The audible effects of early reflections in typical listening rooms was investigated in: Olive, S. E. and Toole, F. E. (1989). “The Detection of Reflections in Typical Rooms”, J. Audio Eng. Soc., 37, pp. 539-553. which is summarized in Chapter 7 of the 3rd edition. All the data are supported by blind listening tests - something that is glaringly absent from most discussions of these topics. Personal opinions are just that. Interestingly, the data show that sensitivity to early reflections reaches a peak around 10 ms, being lower before and after. The "image shift" effect is really quite small, but still audible. One has to wonder whether the "precise" horizontal positioning of a phantom image is truly that important. For me it isn't. However, the beneficial effect of some amount of horizontally reflected sound - adding spaciousness - does seem to flatter a high percentage of recordings, especially those with hard panned (mono) left and right images. But I grant you, this can be personal, as some listeners strive for "pinpoint" imaging, something not always heard in real life performances.

I know both David Griesinger and Earl Geddes, and respect their contributions to the field. To comment on Earl's video would be a massive undertaking. I would like to think that reading my book would settle the issue, as we looked at very much the same concerns. I and my colleagues over the years had the advantage of facilities within which we could acquire abundant, accurate anechoic measurements, as well as listing rooms with which we could test various hypotheses as to what was audible in unbiased double-blind listening. While at Harman we attempted to acquire a sample of one of Earl's loudspeakers but none were available when we enquired. However, the notion of neutral direct sound and similarly neutral reflected sounds - his argument - is one with with I have no disagreement. We just have different ways of quantifying the parameters.

Thank you so much for sharing your insights.

I was wondering whether you ever looked at intra-observer scoring variability over time. My impression is that our subjective enjoyment of sound varies tremendously based on our mood, type of music, playback volume, expectations (big expensive, impressive looking systems), whether we’re putting on our analytic hats out vs just relaxing etc. My sound system sounds fantastic at times and uninspiring at other times…

 

[RDoc]

Flat anechoic/Klipple response is what is preferred, in a typical room that speaker will have a slope.
In his book Toole does not seem to be any sort of advocate for 'house curves' at all.
He just notes that speakers that are anechoically flatish on axis with smooth off axis have a downward trend from bass to highs in a typical real room.

This downward trend varies even with the best measuring speakers as traits such as dispersion, listening distance and room variables affect it and how it ultimately 'looks' on a graph. There is no one slope that a speaker should be forced into let alone a 'house curve' that all speakers would be forced into. It is not the 'summary' or sum of what you hear.

Maybe he will chime in as he has a few times on this thread.

For speaker response, flat, that is 1:1 with the input, sure. However, when actually listening to music, people like non-linear response curves beyond what averaging and smoothing many rooms very jagged individual responses predict. In Toole's book, Fig 12.7 (pp 351 in my edition) shows the "Subjectively-preferred steady-state room curves" from Olive, and none of those are straight lines, much less flat. For trained listeners, above 500Hz or so, it's pretty linear. Below that though, it has a pronounced hump, and for untrained listeners, which includes almost everyone, the hump is very pronounced plus there's significant treble boost. I'd expect that in real rooms, people's tendency to compensate for room acoustics would keep this non-linear preference. So, for systems set up for a pleasing listening experience for most people, I'd expect the house curve, room curve, balance, whatever you want to call it, won't be linear, and certainly not flat. It also is likely somewhat idiosyncratic, perhaps evolving over time and experience.

My point isn't to be pedantic about this. IMHO, with the widespread advent of pretty much audibly perfect DSP corrected speakers and electronic components, the emphasis in home HiFi testing is going to shift more in the direction of "presence". That is, the overall system's multichannel effects that make recorded sound more like hearing live music in a good venue. My suspicion is that the "presence" effect is why Olive's curves, tested in mono, aren't linear.

Here's a presentation by Sean Olive, et all, discussing room curves for trained listeners: https://drive.google.com/file/d/0B9...ew?hl=en&resourcekey=0-uSgBBin0zcCtmpgvEsdskQ

 

[Floyd Toole]

Thank you so much for sharing your insights.

I was wondering whether you ever looked at intra-observer scoring variability over time. My impression is that our subjective enjoyment of sound varies tremendously based on our mood, type of music, playback volume, expectations (big expensive, impressive looking systems), whether we’re putting on our analytic hats out vs just relaxing etc. My sound system sounds fantastic at times and uninspiring at other times…

The simple answer is "no", listener sound quality ratings are remarkably consistent in double-blind tests. We use relatively constant collections of programs, so there is no "pleasure" involved, just commentary on the extent to which the loudspeakers in the test modify/color/distort the timbre of the music. The emotional content of program material obviously affects our mood and our mood can affect our pleasure in the program material. These "enjoyment" factors have been pretty much eliminated from the sound quality evaluations.

 

[Floyd Toole]

For speaker response, flat, that is 1:1 with the input, sure. However, when actually listening to music, people like non-linear response curves beyond what averaging and smoothing many rooms very jagged individual responses predict. In Toole's book, Fig 12.7 (pp 351 in my edition) shows the "Subjectively-preferred steady-state room curves" from Olive, and none of those are straight lines, much less flat. For trained listeners, above 500Hz or so, it's pretty linear. Below that though, it has a pronounced hump, and for untrained listeners, which includes almost everyone, the hump is very pronounced plus there's significant treble boost. I'd expect that in real rooms, people's tendency to compensate for room acoustics would keep this non-linear preference. So, for systems set up for a pleasing listening experience for most people, I'd expect the house curve, room curve, balance, whatever you want to call it, won't be linear, and certainly not flat. It also is likely somewhat idiosyncratic, perhaps evolving over time and experience.

My point isn't to be pedantic about this. IMHO, with the widespread advent of pretty much audibly perfect DSP corrected speakers and electronic components, the emphasis in home HiFi testing is going to shift more in the direction of "presence". That is, the overall system's multichannel effects that make recorded sound more like hearing live music in a good venue. My suspicion is that the "presence" effect is why Olive's curves, tested in mono, aren't linear.

Here's a presentation by Sean Olive, et all, discussing room curves for trained listeners: https://drive.google.com/file/d/0B9...ew?hl=en&resourcekey=0-uSgBBin0zcCtmpgvEsdskQ

I would add only one comment, that these "preferences" really amount to what we used to call bass and treble "tone control" adjustments. Such adjustments were quite common in years past, as program material was quite variable - the "circle of confusion" in action in a period of huge variations in recording control room systems. Nowadays things have definitely improved, but I still find occasions when I would like simple tone controls to reach for - but my complex, computer based, processor doesn't offer such a simple capability. A single calibration cannot be appropriate for all programs, and then we add personal preferences. I am moving later this year and in my new system I will have easily accessible "old fashioned" tone controls, probably in a hand-held device or smart phone.

My present system has a superb multi-sub Sound Field Managed system that plumbs the low-bass depths wonderfully. But, it is occasionally evident that a recording was monitored on, or deliberately mixed/equalized to compensate for typical consumer systems that don't have much bass extension - the low bass is exaggerated through my system, and it needs tweaking in real time. As it is I occasionally live with the "forgivable sin": too much bass

 

[RDoc]

The simple answer is "no", listener sound quality ratings are remarkably consistent in double-blind tests. We use relatively constant collections of programs, so there is no "pleasure" involved, just commentary on the extent to which the loudspeakers in the test modify/color/distort the timbre of the music. The emotional content of program material obviously affects our mood and our mood can affect our pleasure in the program material. These "enjoyment" factors have been pretty much eliminated from the sound quality evaluations.

Mr Toole, I'd just like to thank you and the other serious people you worked with for bringing some actual factual research into a field filled with wizards peddling snake oil.
Along with Amirm's testing, it's actually helped me significantly in selecting and tuning my system and is much appreciated. My wife likes the sound a lot too!

 

[RDoc]

I would add only one comment, that these "preferences" really amount to what we used to call bass and treble "tone control" adjustments. Such adjustments were quite common in years past, as program material was quite variable - the "circle of confusion" in action in a period of huge variations in recording control room systems. Nowadays things have definitely improved, but I still find occasions when I would like simple tone controls to reach for - but my complex, computer based, processor doesn't offer such a simple capability. A single calibration cannot be appropriate for all programs, and then we add personal preferences. I am moving later this year and in my new system I will have easily accessible "old fashioned" tone controls, probably in a hand-held device or smart phone.

My present system has a superb multi-sub Sound Field Managed system that plumbs the low-bass depths wonderfully. But, it is occasionally evident that a recording was monitored on, or deliberately mixed/equalized to compensate for typical consumer systems that don't have much bass extension - the low bass is exaggerated through my system, and it needs tweaking in real time. As it is I occasionally live with the "forgivable sin": too much bass

My Denon based system uses Audyssey with two presets I can switch between, and I keep a small selection of curves on my phone to load if needed.

 

[BDWoody]

These "enjoyment" factors have been pretty much eliminated from the sound quality evaluations.

So, you play a lot of Nickelback?

 

[MediumRare]

For speaker response, flat, that is 1:1 with the input, sure. However, when actually listening to music, people like non-linear response curves beyond what averaging and smoothing many rooms very jagged individual responses predict. In Toole's book, Fig 12.7 (pp 351 in my edition) shows the "Subjectively-preferred steady-state room curves" from Olive, and none of those are straight lines, much less flat. For trained listeners, above 500Hz or so, it's pretty linear. Below that though, it has a pronounced hump, and for untrained listeners, which includes almost everyone, the hump is very pronounced plus there's significant treble boost.

Thanks for sharing that excellent presentation. I don’t understand your comments above. Per the EQ presentation the most preferred EQ was perfectly flat, with a downward slope. No hump. Can you explain, please?

 

[RDoc]

My comment was based on Toole's diagrams from his book. I don't know how he synthesized them but they show a definite bass boost for everyone and a treble boost for non-trained listeners.

 

[Floyd Toole]

My comment was based on Toole's diagrams from his book. I don't know how he synthesized them but they show a definite bass boost for everyone and a treble boost for non-trained listeners.

The caption to the figure (12.7) indicate that the data came from Sean Olive papers. Basically it took a known excellent loudspeaker and let listeners play with especially designed tone controls to adjust things to their "preferences". Of course the program variable is included, so the results are not "absolute". A problem with the test was that overall loudness was not adjusted in real time as the tone adjustments were made. It is possible that the younger inexperienced listeners turned up both the bass and treble just to make it louder. All of this is explained in the text surrounding the figure in the book.

 

[audiofooled]

As my earliest JAES (1985-86) papers showed, listeners in double-blind evaluations preferred loudspeakers with flat and smoothish on-axis (direct sound) frequency responses. Beyond that, those that were preferred exhibited the smoothest off-axis (reflected sound) responses. None of those loudspeakers were horns, so the DI varied with frequency. Now that we have loudspeakers with excellent constant directivity horns, it is seen that these too receive elevated sound quality ratings (see the cover of my book). I don't know that we have sufficient data to be able to define the tolerances on variability in directivity, but it seems fairly obvious that smooth gradual changes would be preferable to uneven, changes - aiming for reasonable timbral matching of direct and reflected sounds - remember that the reflecting surfaces can modify the spectra of reflected sounds. BTW, sound power is a metric that is of questionable merit in conventional rooms with reverberation times of the order of 0.3- 0.5 s - i.e. acoustically relatively "dead". Early reflections dominate, to the point that in spinorama data, the early-reflections curve alone is a good approximation to a steady-state room curve at the listening location. All of this is in my journal papers and book.

So, to answer your question, the direct sound always has perceptual precedence over reflected sounds - that is why it is called the "precedence effect" or "law of the first wavefront". If there is a timbral mismatch between it and the later arriving reflections, the difference, according to double-blind listening tests, is audible and detrimental to the sound quality.

The audible effects of early reflections in typical listening rooms was investigated in: Olive, S. E. and Toole, F. E. (1989). “The Detection of Reflections in Typical Rooms”, J. Audio Eng. Soc., 37, pp. 539-553. which is summarized in Chapter 7 of the 3rd edition. All the data are supported by blind listening tests - something that is glaringly absent from most discussions of these topics. Personal opinions are just that. Interestingly, the data show that sensitivity to early reflections reaches a peak around 10 ms, being lower before and after. The "image shift" effect is really quite small, but still audible. One has to wonder whether the "precise" horizontal positioning of a phantom image is truly that important. For me it isn't. However, the beneficial effect of some amount of horizontally reflected sound - adding spaciousness - does seem to flatter a high percentage of recordings, especially those with hard panned (mono) left and right images. But I grant you, this can be personal, as some listeners strive for "pinpoint" imaging, something not always heard in real life performances.

I know both David Griesinger and Earl Geddes, and respect their contributions to the field. To comment on Earl's video would be a massive undertaking. I would like to think that reading my book would settle the issue, as we looked at very much the same concerns. I and my colleagues over the years had the advantage of facilities within which we could acquire abundant, accurate anechoic measurements, as well as listing rooms with which we could test various hypotheses as to what was audible in unbiased double-blind listening. While at Harman we attempted to acquire a sample of one of Earl's loudspeakers but none were available when we enquired. However, the notion of neutral direct sound and similarly neutral reflected sounds - his argument - is one with with I have no disagreement. We just have different ways of quantifying the parameters.

Thank you very much, I appreciate it. Like most people here on ASR, I value what's measurable and audible in the hopes of better understanding rather complex matters of sound reproduction.

I agree that stereo is fundamentally flawed, but to me, good imaging with neutral on axis and reflected sounds is very helpful and indeed, as you say, flattering to many recordings.