Horns - Necessary to complete the Audiophile Journey?

[Kw6]

Recommendations?

I was kind of eyeing the Klipsch La Scala, or maybe even the Klipschorns.

I like my Hersey IV and may upgrade to La Scalas or Avant-garde Finish in the future. Keep in mind sounds best with SET. I also have Class D & A. SET let me just listen music now

 

[Ro808]

Are those passive?

The JBLs are Bi-Amp/Passive switchable.

 

[Sal1950]

I like my Hersey IV and may upgrade to La Scalas or Avant-garde Finish in the future. Keep in mind sounds best with SET. I also have Class D & A. SET let me just listen music now

As a 30+year owner of La Scala's I will only make a short comment here against your conclusions..
I did run VTL tube power amps for close to 20 years and my choice of them was done using somewhat unusual criteria in the face of today's trends. The VTL's I purchased was done in large part due to the measurements and listening review done by John Atkison of Stereophile.
https://www.stereophile.com/tubepoweramps/1188vtl/index.html
He started off the measurement section of the review saying this,

"On the test bench, you could have been forgiven at first for thinking that this VTL was a solid-state design, the small-signal frequency response (1V into 4 ohms) extending from 2.2Hz to 77kHz (–3dB). Noise levels, too, were impressively low, measuring –90dB, unweighted, with the input shorted, this improving by just over 6dB when an A-weighting network was switched in circuit ahead of the meter."

Designer David Manley during this period of his life had been designing/building tube gear for the recording studio and was focused mainly on making his designs as transparent as possible rather than turning them into some sort of expensive distortion producer. Strictly ultralinear with no triode switching or SET builds. Another quote,

"Enough of the grumbles: what did the VTL 100W Monos do to the sound of my system?
First, as any power amplifier should, they did very little. I am not quite talking "straight wire with gain," but their faults were minor. This is one hell of a transparent amplifier. The high frequencies were neither granular nor dull, but held in check the slight propensity for the top two octaves of the SL700 to be too forward."

Combining the amps "sound" or rather the lack of it, along with some details like the amps high input sensitivity 0.775, and 137k input impedance, made it a perfect choice to be mated to a full passive preamp. This was another component blend choice that aimed to offer the La Scala's as transparent and neutral a source as possible.
Bottom line, IMHO good horns should not be mated to source electronics with the aim of attempting to smooth over any possible brightness or edginess with a bunch of harmonic distortion. But rather to take advantage of their ability resolve inner detail by using the most transparent and distortion free products available. Final results are incredible when playing the best of today's digital recordings.
YMMV

 

[Ro808]

Dave Collins of Manley Labs designed a decent DAC too.

 

[nick-v]

I've owned many speakers over the years with just about every different type of tweeter (chronological - off the top of my head):

-Paradigm Monitor 11v2, Mini Monitor, CC370 (aluminum dome tweeter)
-GR Research Diluceo, custom Selah Audio Center channel to match (Aurum Cantus aluminum ribbon tweeter)
-M Audio BX8 D2 (silk dome tweeter)
-Tannoy Reveal 501a (textile dome tweeter)
-Polk LSi9 w/ upgraded crossovers (Vifa ring radiator tweeter)
-Paradigm Studio 100v5, CC690, Studio 10 (aluminum dome tweeter)
-KEF LS50 (Uni-Q driver aluminum dome tweeter)
-GoldenEar Super Cinema 3D Array X ("High-Velocity Folded Ribbon Tweeter")
-BG Radia R-320, CA-300, PD-8I, PD-8CI (BG Neo 3, Neo 8, Neo 10 planar-magnetic transducers) **Currently in Master Bedroom System
-Dynaudio Bm6a mkII (silk dome tweeter) **Currently in Desktop System
-KEF Q100 (Uni-Q driver aluminum dome tweeter) *Just sold to make way for JBL HDI
-Selah Audio Vigore (ScanSpeak beryllium tweeter) *Just sold to make way for JBL HDI
-Definitive Technology AW6500 (aluminum dome tweeter) **Currently in Outdoor System on Deck
-Sonos One (silk dome tweeter) **Currently in Kitchen/Dining room System (x3)

I've yet to own any horns... until now. I sold off my Selah Audio Vigore and KEF Q100 (I was running 4.2.2) and ordered some JBL HDI series speakers with horn-loaded compression drivers for my media room (HDI-3800, HDI-4500, HDI-1600). They should be here within the next week or two. I'm excited to experience them and see what they're all about.

 

[Sal1950]

They should be here within the next week or two. I'm excited to experience them and see what they're all about.

Hope you like them, I'm in love with mine
https://www.audiosciencereview.com/forum/index.php?threads/sals-system.614/

 

[watchnerd]

I've owned many speakers over the years with just about every different type of tweeter

I've yet to own any horns... until now. I sold off my Selah Audio Vigore and KEF Q100 (I was running 4.2.2) and ordered some JBL HDI series speakers with horn-loaded compression drivers for my media room (HDI-3800, HDI-4500, HDI-1600). They should be here within the next week or two. I'm excited to experience them and see what they're all about.

There seems to be a fork in the road between:

Vintage horns + tubes / SET

vs

"Home theater" horns + solid state

I've yet to see someone using home theater horns with tubes.

Why is that?

 

[Sal1950]

I've yet to see someone using home theater horns with tubes.

Why is that?

We know better.
JK
Probably things like heat and cost of long term maintenance to have 5+ channels of tubes running.
Doesn't make much sense.

 

[Duke]

That sounds interesting but are you going for a 2- or a 3-way speaker?

The main speakers would be two-ways, bigger than the Ginas and going down to about 80 Hz ballpark, with rear-firing arrays. A distributed multi-sub system would take over south of 80 Hz.

I'll also do at least one less expensive two-way floorstander that doesn't rely on subs.

I wonder if the preference for the Ultima Salon2 over the M2 could partly have been due to the former being a 4-way.

Yes I think the Salon being a 4-way helped it to win. I think its pattern width contributed to the difference, and it's easier to get a wide pattern with conventional drivers than with horns and big prosound woofers. And having four "ways" with different driver diameters facilitated keeping that wide radiation pattern remarkably smooth.

In my opinion, a single-speaker listening test generally favors the speaker with the widest radiation pattern, assuming no off-axis anomalies, for several reasons:

First, timbre is enriched by spectrally correct reflections, whether in the concert hall or in the listening room. A good wide-pattern speaker will put more energy into the reflections.

Second, spatial qualities apparently matter a great deal to perceived sound quality (according to Klippel's studies which Toole cites). Reflections convey spatiality, again something which favors wide-pattern speakers, even in single-speaker evaluations.

Third, either single speaker will be far enough from the side walls that the first sidewall reflections will arrive after a much longer time delay than normal, whether its pattern is wide or narrow. So if there is a perceptual downside to strong early sidewall reflections, it won't show up in single-speaker auditions where the side walls are abnormally far away. Thus if there is a perceptual penalty for having a wide pattern, it is not being evaluated in single-speaker listening tests.

Finally, these apparent disadvantages of narrow-pattern speakers seem to be significantly diminished in stereo listening, or at least that's one way to interpret the data Toole presents comparing stereo evaluations vs mono evaluations. To put it another way, mono listening may disproportionately penalize narrow-pattern speakers.

Regarding the single-speaker M2 vs Salon2 shoot-out, my recollection is that some who preferred the Salon2 cited a tonal preference and some a spatial preference.

Regarding the tonal preference, in my opinion the M2 is voiced too close to "flat" on-axis and within the listening window. My evaluators prefer more of a gentle downward trend in that type of speaker. Also, the timbre was probably richer on the Salon2 from the greater amount of energy in the reflections.

Regarding the spatial preference, more in-room reflections will do two things: They will tend to de-emphasize the loudspeaker as the apparent sound source, and they will more effectively convey the venue cues (reverberation) in the recording. Obviously we'd expect this advantage of wide-pattern speakers to extend to stereo listening as well, UNLESS strong early sidewall reflections are detrimental... and recall that early sidewall reflections are simply not present when single speakers in the middle of a big room are being compared.

Imo, ime, ymmv, etc.

From a conceptual point of view I would like a speaker with a driver doing the bass duties, a midrange driver to cover the mids and another for the high frequencies. Sub-bass would ideally require a dedicated driver or the speaker could be assisted by subwoofers but I've lived hapilly with a large 3-way. In essence a 4-way or a 3-way + sub. Preferably all sealed or if ported with very low tuning.

The specifics of what I do are different, but then I want a large diameter midbass driver for the sake of pattern-matching with the horn in the crossover region.

I plan to do multiple pluggable ports, for adaptation to different room acoustic situations. The main speakers will have sufficient headroom to forego protective highpass filtering in normal (less than 120 dB/1 meter) use.

 

[tuga]

The main speakers would be two-ways, bigger than the Ginas and going down to about 80 Hz ballpark, with rear-firing arrays. A distributed multi-sub system would take over south of 80 Hz.

I'll also do at least one less expensive two-way floorstander that doesn't rely on subs.



Yes I think the Salon being a 4-way helped it to win. I think its pattern width contributed to the difference, and it's easier to get a wide pattern with conventional drivers than with horns and big prosound woofers. And having four "ways" with different driver diameters facilitated keeping that wide radiation pattern remarkably smooth.

In my opinion, a single-speaker listening test generally favors the speaker with the widest radiation pattern, assuming no off-axis anomalies, for several reasons:

First, timbre is enriched by spectrally correct reflections, whether in the concert hall or in the listening room. A good wide-pattern speaker will put more energy into the reflections.

Second, spatial qualities apparently matter a great deal to perceived sound quality (according to Klippel's studies which Toole cites). Reflections convey spatiality, again something which favors wide-pattern speakers, even in single-speaker evaluations.

Third, either single speaker will be far enough from the side walls that the first sidewall reflections will arrive after a much longer time delay than normal, whether its pattern is wide or narrow. So if there is a perceptual downside to strong early sidewall reflections, it won't show up in single-speaker auditions where the side walls are abnormally far away. Thus if there is a perceptual penalty for having a wide pattern, it is not being evaluated in single-speaker listening tests.

Finally, these apparent disadvantages of narrow-pattern speakers seem to be significantly diminished in stereo listening, or at least that's one way to interpret the data Toole presents comparing stereo evaluations vs mono evaluations. To put it another way, mono listening may disproportionately penalize narrow-pattern speakers.

Regarding the single-speaker M2 vs Salon2 shoot-out, my recollection is that some who preferred the Salon2 cited a tonal preference and some a spatial preference.

Regarding the tonal preference, in my opinion the M2 is voiced too close to "flat" on-axis and within the listening window. My evaluators prefer more of a gentle downward trend in that type of speaker. Also, the timbre was probably richer on the Salon2 from the greater amount of energy in the reflections.

Regarding the spatial preference, more in-room reflections will do two things: They will tend to de-emphasize the loudspeaker as the apparent sound source, and they will more effectively convey the venue cues (reverberation) in the recording. Obviously we'd expect this advantage of wide-pattern speakers to extend to stereo listening as well, UNLESS strong early sidewall reflections are detrimental... and recall that early sidewall reflections are simply not present when single speakers in the middle of a big room are being compared.

Imo, ime, ymmv, etc.



The specifics of what I do are different, but then I want a large diameter midbass driver for the sake of pattern-matching with the horn in the crossover region.

I plan to do multiple pluggable ports, for adaptation to different room acoustic situations. The main speakers will have sufficient headroom to forego protective highpass filtering in normal (less than 120 dB/1 meter) use.

Like you I am also somewaht perplexed by some of Toole's interpretations of the research data and wonder if this may have driven the theory in a particular direction.
Sometimes it looks as though he is making a case for wide-directivity to support, or at least coloured by, his own listening preferences, but it may have been accidental.

I didn't know that the M2 vs. Salon2 match had been performed in mono.
I find that it makes absolutely no sense to use mono for preference testing.
If so, your comments regarding wide-directivity make sense and are supported by Toole's preference research in regard to mono and wide directivity (i.e. Kef vs Quad).

As for your new design, I quite like the idea of having a sole driver covering the whole of the midranges and even down to the midbass, my only concern, albeit theoretical, would be intermodulation distortion.
I'd love to hear it, hopefuly someday...

 

[Duke]

As for your new design, I quite like the idea of having a sole driver covering the whole of the midranges and even down to the midbass, my only concern, albeit theoretical, would be intermodulation distortion.

Thank you!

The midbass section I have in mind for the big speaker consists of two 15" midwoofers in a low-tuned ported or possibly sealed enclosure (depending on room acoustics and/or amplifier output impedance). So intermodulation distortion would be on the low side.

 

[thewas]

I didn't know that the M2 vs. Salon2 match had been performed in mono.
I find that it makes absolutely no sense to use mono for preference testing.

I am also not sure what to think about it, but on the other hand we must also show the argumentation of Toole:

"I guess you have to experience it to believe it. I did my first stereo vs mono test in 1985 and it was a very thorough and carefully conducted blind test. All of this is in JAES papers and in both books, so I won't get into regurgitating old research. The amazing reality is that listeners reported extensively on spatial/soundstage characteristics when listening in mono - which everyone, myself included, thought would be commented on only in stereo. The spatial ratings in mono closely tracked the sound quality ratings, and both were more strongly differentiated in mono than in stereo listening. We have learned since then that sound quality and spatial quality are closely linked.

In mono, the highest rated loudspeakers came closest to "disappearing" behind the screen, leaving impressions of image size and distance/depth to information in the recordings. Because stereo is mono L, mono R and double-mono amplitude and/or delay panned phantom images (including of course the featured artist), it is understandable that the soundstage is improved if one's attention is not drawn to the loudspeakers. This tends to be an advantage for wide dispersion loudspeakers. It was interesting to see that the mono ratings agreed with stereo ratings for close miked, pan potted stereo pop (truly multiple mono). With more complex pop and classical music there is a huge amount of uncorrelated information in both channels (to generate the desired spaciousness) and the spatial/soundstage ratings were not strongly differentiated. The dominant factor in the stereo tests was the recording itself, which, if you know how the signals are captured and processed, is not surprising. These are control room creations.

Over the years we have done a few stereo vs. mono tests to convince skeptics with the same result: the highest rated loudspeakers in mono, have been the highest rated loudspeakers in stereo but the differentiation in stereo was not always as easily discerned. Highly directional loudspeakers tend to stand out as lacking in both stereo and mono. Recently, the comparison has been expanded to multichannel, and it is even more forgiving than stereo. The more channels that are simultaneously active, the less the room interactions can be heard. But we listen in mono much of the time, in multi-mono stereo in non-classical music, a dominant center channel in movies and any time a signal is hard panned to a single channel. So how a loudspeaker sounds in mono matters, and mono tests yield the most critical comments. In stereo expect a speaker to be no worse, possibly better, but the stereo soundstage is definitely engaging."

Source: https://www.avsforum.com/forum/89-s...viewed-speakers-ever-made-8.html#post54603712

and his comment on this image:

"If you look at the scatter of individual responses it is clear that there were random differences of opinion. Do a little math and it will be clear that the differences in the stereo tests are not statistically significant. These were averages for all programs, and program is a variable - more obvious in stereo than in mono. Bear in mind that these tests were done by me in 1985 - 32 years ago. I know of no other tests of comparable thoroughness and published in a refereed journal since then. It may or may not have been flawless, but it remains the best evidence to date.
Toole, F. E. (1985). “Subjective measurements of loudspeaker sound quality and listener preferences”, J. Audio Eng. Soc., 33. pp. 2-31.
Toole, F. E. (1986). “Loudspeaker measurements and their relationship to listener preferences”, J. Audio Eng. Soc., 34, pt.1, pp. 227-235, pt. 2, pp. 323-348.

This is not just my opinion. Evans et al. (2009 ) reviewed several studies examining the effects of loudspeaker directivity, concluding that there were several fundamental questions remaining to beproperly investigated. They did note that the author’s work described here “appears to be the most relevant to date, with regard to loudspeaker directivity effects.”
Evans, W., Dyreby, J., Bech, S., Zielinski, S. and Rumsey, R. (2009). “Effects of loudspeaker directivity on perceived sound quality – a review of existing studies”, Audio Eng. Soc., 126th Convention, paper 7745.

Most people don't have the facilities, the money, the time, or the attention span, to undertake such meticulous tests. Nevertheless, since then there has been a never ending stream of unsubstantiated opinion from all possible quarters. But that is audio.

.

Progress is slow, but there has been progress. A major problem is that enthusiasts in general have not taken the trouble to find and read the existing science. They will not be guided to it by audiophile publications because for many of them the scientific findings run contrary to their business interests. Fortunately a few have acknowledged that the complete answers may not lie in "take it home and listen to it" sighted and biased subjective evaluations. But for them the lack of facilities and finance are serious limitations. So, in the absence of published comprehensive anechoic data, consumers and professionals are seriously handicapped when trying to make rational decisions.

The last chapter in my new book is a 50 year retrospective on loudspeakers - lots of curves. It is clear that there were some excellent loudspeakers years ago, and some mediocre ones even today. Price is not a factor. Professionals seem to have had the greatest problems embracing measurements. Several consumer audio magazines show measurements in their reviews - not perfect, but usefully accurate, and in one case (soundstagenetwork.com) nearly so as they use my old NRCC facilities. Professional magazines almost always leave evaluations of monitor loudspeakers to sighted listening tests by individuals who have a high probability of having hearing loss. Go figure."

Source: https://www.avsforum.com/forum/89-s...iewed-speakers-ever-made-10.html#post54613648

 

[tuga]

I am also not sure what to think about it, but on the other hand we must also show the argumentation of Toole:

"I guess you have to experience it to believe it. I did my first stereo vs mono test in 1985 and it was a very thorough and carefully conducted blind test. All of this is in JAES papers and in both books, so I won't get into regurgitating old research. The amazing reality is that listeners reported extensively on spatial/soundstage characteristics when listening in mono - which everyone, myself included, thought would be commented on only in stereo. The spatial ratings in mono closely tracked the sound quality ratings, and both were more strongly differentiated in mono than in stereo listening. We have learned since then that sound quality and spatial quality are closely linked.

In mono, the highest rated loudspeakers came closest to "disappearing" behind the screen, leaving impressions of image size and distance/depth to information in the recordings. Because stereo is mono L, mono R and double-mono amplitude and/or delay panned phantom images (including of course the featured artist), it is understandable that the soundstage is improved if one's attention is not drawn to the loudspeakers. This tends to be an advantage for wide dispersion loudspeakers. It was interesting to see that the mono ratings agreed with stereo ratings for close miked, pan potted stereo pop (truly multiple mono). With more complex pop and classical music there is a huge amount of uncorrelated information in both channels (to generate the desired spaciousness) and the spatial/soundstage ratings were not strongly differentiated. The dominant factor in the stereo tests was the recording itself, which, if you know how the signals are captured and processed, is not surprising. These are control room creations.

Over the years we have done a few stereo vs. mono tests to convince skeptics with the same result: the highest rated loudspeakers in mono, have been the highest rated loudspeakers in stereo but the differentiation in stereo was not always as easily discerned. Highly directional loudspeakers tend to stand out as lacking in both stereo and mono. Recently, the comparison has been expanded to multichannel, and it is even more forgiving than stereo. The more channels that are simultaneously active, the less the room interactions can be heard. But we listen in mono much of the time, in multi-mono stereo in non-classical music, a dominant center channel in movies and any time a signal is hard panned to a single channel. So how a loudspeaker sounds in mono matters, and mono tests yield the most critical comments. In stereo expect a speaker to be no worse, possibly better, but the stereo soundstage is definitely engaging."

Source: https://www.avsforum.com/forum/89-s...viewed-speakers-ever-made-8.html#post54603712

and his comment on this image:

View attachment 75071

"If you look at the scatter of individual responses it is clear that there were random differences of opinion. Do a little math and it will be clear that the differences in the stereo tests are not statistically significant. These were averages for all programs, and program is a variable - more obvious in stereo than in mono. Bear in mind that these tests were done by me in 1985 - 32 years ago. I know of no other tests of comparable thoroughness and published in a refereed journal since then. It may or may not have been flawless, but it remains the best evidence to date.
Toole, F. E. (1985). “Subjective measurements of loudspeaker sound quality and listener preferences”, J. Audio Eng. Soc., 33. pp. 2-31.
Toole, F. E. (1986). “Loudspeaker measurements and their relationship to listener preferences”, J. Audio Eng. Soc., 34, pt.1, pp. 227-235, pt. 2, pp. 323-348.

This is not just my opinion. Evans et al. (2009 ) reviewed several studies examining the effects of loudspeaker directivity, concluding that there were several fundamental questions remaining to beproperly investigated. They did note that the author’s work described here “appears to be the most relevant to date, with regard to loudspeaker directivity effects.”
Evans, W., Dyreby, J., Bech, S., Zielinski, S. and Rumsey, R. (2009). “Effects of loudspeaker directivity on perceived sound quality – a review of existing studies”, Audio Eng. Soc., 126th Convention, paper 7745.

Most people don't have the facilities, the money, the time, or the attention span, to undertake such meticulous tests. Nevertheless, since then there has been a never ending stream of unsubstantiated opinion from all possible quarters. But that is audio.

.

Progress is slow, but there has been progress. A major problem is that enthusiasts in general have not taken the trouble to find and read the existing science. They will not be guided to it by audiophile publications because for many of them the scientific findings run contrary to their business interests. Fortunately a few have acknowledged that the complete answers may not lie in "take it home and listen to it" sighted and biased subjective evaluations. But for them the lack of facilities and finance are serious limitations. So, in the absence of published comprehensive anechoic data, consumers and professionals are seriously handicapped when trying to make rational decisions.

The last chapter in my new book is a 50 year retrospective on loudspeakers - lots of curves. It is clear that there were some excellent loudspeakers years ago, and some mediocre ones even today. Price is not a factor. Professionals seem to have had the greatest problems embracing measurements. Several consumer audio magazines show measurements in their reviews - not perfect, but usefully accurate, and in one case (soundstagenetwork.com) nearly so as they use my old NRCC facilities. Professional magazines almost always leave evaluations of monitor loudspeakers to sighted listening tests by individuals who have a high probability of having hearing loss. Go figure."

Source: https://www.avsforum.com/forum/89-s...iewed-speakers-ever-made-10.html#post54613648

This is a Toole interpretation of the data I disagree with.
The fact that the Quad rates higher in stereo for both sound and spatial quality and as high as the other speakers is enough in my view to invalidate Toole's claim on the effectiveness of mono listening for preference evaluation purposes because speakers are listened to in stereo (I don't deny its usefulness for assessing particular aspects of performance). One can conclude that wide-directivity speakers sound more pleasing in mono but that is all.
Speakers are commonly listened to in stereo pairs so that is how they should be evaluated. No one walks into a store and asks to listen to a single speaker then buys a pair... That is absurd.

 

[thewas]

I think the only step that would take us really forward would be some more and newer mono and stereo blind tests to verify or falsify his 1985 results as the current data basis is just too limited but I am afraid that Harman research interest in loudspeakers preference has gone down in the last years, which is a pity as with the new variable directivity Lexicon (even a Harman group porduct) and Beolab loudspeakers such a comparison would be more expedient then using different loudspeakers.
Hopefully ASR will grow more to do such in the future.

 

[Duke]

Duke, I wonder if you may be a tiny bit off in your summary of Harman's findings. Since Sean Olive's prediction model was published, I personally believe that the Revel speakers have been actually designed for flatter listening window than on-axis measurements, based on Kevin Voeck's (@AVKV) comments ((https://www.avsforum.com/forum/89-speakers/710918-revel-owners-thread-408.html#post56177990): "As our research has long indicated, the listening window is a far better indicator of direct sound quality than is any on-axis curve.".

I tend to think in terms of "first-arrival sound", but have been careless in using terms like "on-axis" which is almost always different. As a subscriber to Earl Geddes' ideas, my "first-arrival sound" is something like 20 degrees off-axis, which may not even be within Harman's "listening window".

One implication of listening from 20 degrees off-axis is that there is less spectral discrepancy between the first-arrival sound and the reverberant sound. This is a (probably minor) area where I disagree with Harman's conclusions. My understanding is that Harman finds listeners prefer an approximately flat listening window response and then a smooth but downward-sloping off-axis response, and from that conclude this to be the ideal. I do not think they have ever evaluated a speaker whose first-arrival and reverberant field responses were very similar (moreso than for the Salon2 and M2), with BOTH being smooth and gently downward-sloping, but imo that may be the ideal.

Consequently, the Harman approach seems to have resulted in more than one approach, since it has produced the JBL Array 1400 which I regard in some ways as an earlier, more extreme version of the M2 in its relatively constant but very high directivity and flattish sound power above 800 Hz, then the M2 with flatter sound power and directivity, though not as high.

I have never heard the JBL Array 1400 but it has started to fascinate me, because of some ideas I'm currently exploring in-house. For instance, the Array 1400 will much weaker sidewall interactions...

From what little I understand, the double-blind loudspeaker evaluations were done using single (mono) speakers in the multichannel listening lab, which highly emphasizes direct sound (https://www.harman.com/documents/HarmanWhitePaperMLLListeningLab_0.pdf: "In the MLL room, the only significant first order reflections are from the floor, and these are attenuated at higher frequencies by the carpet. At listener-loudspeaker distances greater than 2 m any reflection with a path length greater than 6.34 m will be attenuated 10 dB by spreading loss [18]. This effectively eliminates all second order reflections since their path length exceeds this value")

Imo the single-speaker evaluations do not adequately evaluate the early reflections, in particular the early sidewall reflections. They have been eliminated by path lengths and/or room treatments. The sidewall reflections in the single-speaker audition tests arrive after a long enough time delay to be clearly beneficial, in my opinion. In a normal stereo setup, those sidewall reflections will arrive much earlier and will be stronger.

I understand that wide-pattern speakers are still generally preferred in stereo listening, but that could be because the benefits of their relatively abundant later-arriving reverberant energy more than offsets any detriment from their stronger early reflections. There are unorthodox approaches which deliver abundant later-arriving reverberant energy without a corresponding increase in the early reflections... and I suppose multichannel is arguably one of them.

 

[Duke]

View attachment 75071

Thanks for posting Toole's side of the story and that image.

Imo an alternative way of interpreting the data in the image might be: The Sound Quality of BB, and the Spatial Qualities of BB and E, were penalized by mono listening.

Personally, I'd be interested in the devilish details of WHY the scores for BB and E went up dramatically in stereo versus mono, even if the actual ranking order only changed in one instance (E moving from second to first place in Spatial Quality). Imo there may well be something of psychoacoustic significance going on there.

 

[youngho]

One can conclude that wide-directivity speakers sound more pleasing in mono but that is all.
Speakers are commonly listened to in stereo pairs so that is how they should be evaluated. No one walks into a store and asks to listen to a single speaker then buys a pair... That is absurd.

According to Toole, the Revel Salon 2 and JBL Array 1400 tested extremely similarly in Harman's listening tests, despite very different directivity:
https://www.audiosciencereview.com/...review-measurements.14310/page-17#post-442549

 

[youngho]

I think the only step that would take us really forward would be some more and newer mono and stereo blind tests to verify or falsify his 1985 results as the current data basis is just too limited but I am afraid that Harman research interest in loudspeakers preference has gone down in the last years, which is a pity as with the new variable directivity Lexicon (even a Harman group porduct) and Beolab loudspeakers such a comparison would be more expedient then using different loudspeakers.
Hopefully ASR will grow more to do such in the future.

Have you read Toole's most recent edition where he discusses one paper using the Beolab speakers you mention?

[edit] sorry, I was mistaken about the Beolab speakers mentioned in the book.

 

[Duke]

Have you read Toole's most recent edition where he discusses one paper using the Beolab speakers you mention?

Do you know if anyone has evaluated the different "modes" (radiation patten widths) of the Beolab 90 in a single speaker configuration? My understanding is that the "Narrow" mode is preferred in stereo, but wonder whether that holds true in single-speaker mono listening.

 

[youngho]

I tend to think in terms of "first-arrival sound", but have been careless in using terms like "on-axis" which is almost always different. As a subscriber to Earl Geddes' ideas, my "first-arrival sound" is something like 20 degrees off-axis, which may not even be within Harman's "listening window".

Listening window is defined as "average frequency response within a ±30º horizontal, ±10º vertical window."

One implication of listening from 20 degrees off-axis is that there is less spectral discrepancy between the first-arrival sound and the reverberant sound. This is a (probably minor) area where I disagree with Harman's conclusions. My understanding is that Harman finds listeners prefer an approximately flat listening window response and then a smooth but downward-sloping off-axis response, and from that conclude this to be the ideal. I do not think they have ever evaluated a speaker whose first-arrival and reverberant field responses were very similar (moreso than for the Salon2 and M2), with BOTH being smooth and gently downward-sloping, but imo that may be the ideal.

The JBL Array 1400 has relatively flat sound power above 1 kHz, similar to the Harman on-axis measurement. On page 190-191, Toole discusses the bipolar Mirage M1 speakers, which appear to be nearly flat from 200-10 kHz. I think it's important to note that Toole wrote that "[the M1" performed well in double-blind listening tests in the small NRCC room, and also in this large one" (referring to his own personal room at the time, for which he bought a pair).

I have never heard the JBL Array 1400 but it has started to fascinate me, because of some ideas I'm currently exploring in-house. For instance, the Array 1400 will much weaker sidewall interactions...

Probably helpful to distinguish between ipsilateral and contralateral sidewall interactions.

Imo the single-speaker evaluations do not adequately evaluate the early reflections, in particular the early sidewall reflections. They have been eliminated by path lengths and/or room treatments. The sidewall reflections in the single-speaker audition tests arrive after a long enough time delay to be clearly beneficial, in my opinion. In a normal stereo setup, those sidewall reflections will arrive much earlier and will be stronger.

If you're referring to the Harman single-speaker audition tests, I don't know that the relative strength of these reflections is likely to be meaningful. See my post above about Revel Salon 2 versus JBL Array 1400 listening tests.

I understand that wide-pattern speakers are still generally preferred in stereo listening, but that could be because the benefits of their relatively abundant later-arriving reverberant energy more than offsets any detriment from their stronger early reflections. There are unorthodox approaches which deliver abundant later-arriving reverberant energy without a corresponding increase in the early reflections... and I suppose multichannel is arguably one of them.

I think that it will probably be helpful to eventually make some distinctions, even if broad, about preferences, like Toole does regarding lateral reflections or perhaps like Tapio Lokki does for classical music: https://users.aalto.fi/~ktlokki/Publs/JASMAN_vol_140_iss_1_551_1.pdf.