A Broad Discussion of Speakers with Major Audio Luminaries

[MattHooper]

General comment as someone who has in recent years entered the world of recording orchestral music: the world of recording is as much subject to nonsense as the rest of the audio world. Music recording (and the whole production process) is extremely subjective, as it's as much a part of the creative process as the performance itself. Even simple recordings feature dozens of small and/or large adjustments with regards to EQ (static and dynamic), compression, reverb, etc..

(bolding mine)

Just to be clear, by “ nonsense” you aren’t referring to the manipulations used for recordings are you?

Do you mean that there are just as many people in recording falling for snake oil or something like that?

 

[kmanusa]

That was the folklore, to be sure, but scratch the surface and they were active experimentalists. I recall the early spaced-omni microphone phase, with very spacious delivery, ambiguous localizations - which was their objective for some recordings. But I just did a quick internet search and came up with these comments from some insiders:
“Jack Renner at Telarc used a 3 mike system with sometimes 2 more out in the auditorium for ambience. Bob Woods used 3 or sometimes 2 and again, sometimes 2 out in the hall. And yet on other recordings, particularly the jazz they used spot mikes. Sometimes when recording classical with individual singing, these individuals were also spot miked. On many of the Kunzel pieces which had sound effects, these effects were recorded separately by Mike Bishop and added in. It all just depended on the sound they were after.”

“Bobby Owsinski interviewed Michael Bishop - Episode #170 – Whiz Passes Psy, Gibson Buys Onkyo, And Engineer Michael Bishop | Bobby Owsinski's Inner Circle Podcast - recently. Mr. Bishop talked a fair bit about mic placement that was used while he was recording for Telarc. If I recall correctly he talked about 4 omni microphones spread out for "coverage" and possibly the use of some targeted mics. Much closer to The orchestra than you would expect.”

“Telarc recorded the Milwaukee Symphony performing Smetana's Ma Vlast back when we were recording them for Koss Classics. I recall part of the Telarc "sound" came from packing the musicians very tightly together, almost uncomfortably so.”

https://www.stereophile.com/content/jack-renner-telarc-direct-cleveland-page-2. This interview is an interesting perspective.

John Eargle, a noted recording engineer and much more, was a good friend of mine. https://en.wikipedia.org/wiki/John_M._Eargle
He recorded for the Delos label, and was highly regarded for his classical recordings. He described the process in detail, which began with some overall orchestra miking combined with "tasteful" close miking to amplify weak instruments, and to provide clarity for solo artists. These close pickup signals were adjusted in timing and electronic ambiance to make their contributions fit the overall soundstage. He was especially proud that these manipulations were not detected, and that the result was interpreted as a credible replica of a live experience. John was a significant factor in encouraging me to write the first edition of these books.

As I said, the circle of confusion is difficult to avoid. It is almost certain that all decisions were made while listening to forward firing loudspeakers in a small room.

I suspect Telarc was reacting, at least in part, to what the "High End" press was writing about in those days. The Absolute Sound in particular was enamored of the 3-channel, 3- microphone recording technique said to be used by Bob Fine for the Mercury Living Presence classical recordings of the 1950s and early 1960s. I enjoyed the early Telarc recordings quite a bit (e.g., the Jack Renner recordings) but was not upset when they expanded their approach to recording orchestras over time-- most of the time their recordings avoided sounding gimmicky (the 1812 Overture recording being an exception!) and they eschewed the often grotesque multi-miking techniques used by many of the major classical labels back then. Despite this, Telarc did not make the High End press totally happy as it was about that time that some of them began their anti-digital crusade. Which also coincided with me eventually cancelling my subscriptions to the "little magazines".

One hundred percent agree that John Eargle was a superb recoding engineer. Anyone wanting to hear a wonderful Christmas recording should check out the Delos recording "Christmas With Roger Wagner" by the Roger Wagner Chorale engineered by John Eargle. This choral recording has real dynamic range-- don't get fooled by turning up the volume to hear the quiet parts because the chorus and pipe organ will surprise the heck out of you during the climaxes! John's meticulous processes enhance the beauty of this recording and are largely transparent; coupled with the wonderful singing, this remains my favorite holiday CD! The whole classical series John did for Delos was uniformly excellent!

Based on your John Eargle anecdotes related in another forum (AVS?), I suspect John was well aware of the circle of confusion. Perhaps such considerations were behind his decision to send a bunch of JBL loudspeakers for you to evaluate/measure when you were still working at NRC?

 

[TimVG]

(bolding mine)

Just to be clear, by “ nonsense” you aren’t referring to the manipulations used for recordings are you?

Do you mean that there are just as many people in recording falling for snake oil or something like that?

Yes, exactly. Not snake oil necessarily, but other pitfalls coming down to (questionable) subjective opinions, without feeling the need to consult, much less appreciate the science on sound reproduction that has been done.

It’s a bit like with sound reproduction itself: speakers are by far the most defining factor in what we hear, not so much the ancillary equipment. Sure, electronics can differ too, but if they measure correctly, that’s pretty much the end of the story — at least from a scientific point of view. Similarly in recording, microphones (along with obviously the performance itself and venue acoustics) play the dominant role, if your recording chain is technically sound.

Yet, just like with playback, an excessive amount of attention often goes to all sorts of ancillary gear and tweaks which sometimes reaches mythical proportions, that are mainly subjective and not grounded in thorough objective peer reviewed evidence.

You could even argue that a microphone’s coloration is part of the artistic process — but that’s precisely why playback needs to be as neutral and accurate as possible, to truthfully render what was captured. Yet many of these so called golden ears, skilled in other aspects as they may be, simply refuse to believe this fact.

 

[Floyd Toole]

I suspect Telarc was reacting, at least in part, to what the "High End" press was writing about in those days. The Absolute Sound in particular was enamored of the 3-channel, 3- microphone recording technique said to be used by Bob Fine for the Mercury Living Presence classical recordings of the 1950s and early 1960s. I enjoyed the early Telarc recordings quite a bit (e.g., the Jack Renner recordings) but was not upset when they expanded their approach to recording orchestras over time-- most of the time their recordings avoided sounding gimmicky (the 1812 Overture recording being an exception!) and they eschewed the often grotesque multi-miking techniques used by many of the major classical labels back then. Despite this, Telarc did not make the High End press totally happy as it was about that time that some of them began their anti-digital crusade. Which also coincided with me eventually cancelling my subscriptions to the "little magazines".

One hundred percent agree that John Eargle was a superb recoding engineer. Anyone wanting to hear a wonderful Christmas recording should check out the Delos recording "Christmas With Roger Wagner" by the Roger Wagner Chorale. This choral recording has real dynamic range-- don't get fooled by turning up the volume to hear the quiet parts because the chorus and pipe organ will surprise the heck out of you during the climaxes! John's meticulous processes enhance the beauty of this recording and are largely transparent; coupled with the wonderful singing, this remains my favorite holiday CD! The whole classical series John did for Delos was uniformly excellent!

Based on your John Eargle anecdotes related in another forum (AVS?), I suspect John was well aware of the circle of confusion. Perhaps such considerations were behind his decision to send a bunch of JBL loudspeakers for you to evaluate/measure when you were still working at NRC?

John definitely was aware of the circle of confusion, and he was aware of the inconsistencies in sound quality among JBL offerings when he sent the samples for measurement at the NRCC. I used those measurements as examples at the end of my four-hour lecture to assembled Harman engineers and management, including Sidney Harman, for my job "interview". At that point in the lecture it was evident what the curves meant. Not all of the engineers were pleased. I was hired. John sat in the front row, smiling.

John and I discussed such things many times over lunch while I was at Harman. He intended to write a foreword for my first book, and was scheduled to come for dinner to collect the first chapters for his review when he died two days before, in his home, listening to music. I was greatly saddened. He was a significant loss to the audio world - a technical and an artistic engineer; a powerful combination. RIP.

 

[MattHooper]

You could even argue that a microphone’s coloration is part of the artistic process — but that’s precisely why playback needs to be as neutral and accurate as possible, to truthfully render what was captured.

Just a slight bit of departure from the above.

I certainly agree that playback needs to be as neutral and accurate as possible if the goal is to render what was captured on the recording as faithfully as possible.

I’m a big fan not just of music but of the nature of recording themselves. I really like the nuances of production and recording style… I can even enjoy “ bad” recordings (eg thin, coarse sounding or whatever ) if they are nonetheless interesting in character. So I’m not generally inclined to reach for tone controls or EQ to fix the nature of the original record recordings to a significant degree. I actually had a digital parametric EQ in my rig for something like 20 years that went essentially unused so I sold it a few years ago.

That said, it’s also my view that really accurate neutral playback is unnecessary to hear most of the nuances in recordings, including the effects of different microphones, mic placements, specific reverbs added to instruments and all the choices that go into musical production.

Just as Dr Toole suggests that the character of acoustic cues in recordings tend to dominate over playback room acoustic cues, I would say that the recording and production choices in a recording tend to dominate over speaker or playback colorations.

One of countless examples : The new and interesting production techniques in the Beatles Sgt. Pepper translated through any number of consumer playback systems at the time, and through the years.

As a Rush fan, I’ve listened to the track Cygnus X-1 on countless different systems, from my family’s original stereo system, to boomboxes, Sony Walkman, car stereos, the various systems owned by my friends, all the systems I’ve been able to afford through my life… and the distinct production sounds of the opening of that track, the very particular percussive attack on Geddy Lee’s Rickenbaker bass, emerging out of that very distinctive alcove of trailing reverb, is always distinctive and recognizable.

And that’s why I’ve been able to hear and enjoy those distinctions across a wide variety of different types of loudspeakers in my room as well.

So it’s true that as an enthusiast I do sweat the small details, and I like to hear even the tightest distinctions in tracks. But this is a personal preference thing. I still try and keep the big picture view that most production choices translate across a wide variety of speaker/system performance, IMO.

 

[kmanusa]

John definitely was aware of the circle of confusion, and he was aware of the inconsistencies in sound quality among JBL offerings when he sent the samples for measurement at the NRCC. I used those measurements as examples at the end of my four-hour lecture to assembled Harman engineers and management, including Sidney Harman, for my job "interview". At that point in the lecture it was evident what the curves meant. Not all of the engineers were pleased. I was hired. John sat in the front row, smiling.

John and I discussed such things many times over lunch while I was at Harman. He intended to write a foreword for my first book, and was scheduled to come for dinner to collect the first chapters for his review when he died two days before, in his home, listening to music. I was greatly saddened. He was a significant loss to the audio world - a technical and an artistic engineer; a powerful combination. RIP.

I'm not trying to be patronizing but in IMHO, the industry greatly benefited from you moving to Harman and continuing your work there. I'm especially thankful for the measurements you did for John that helped so much in your "interview"-- in particular, the JBL 250Ti measurements shown in the 3rd edition of your Sound Reproduction book.

Having owned the 250Ti for quite a number of years, certain issues bothered me about it-- the kind of issues that are sometimes difficult to pick up in a dealer show room or brief home trial. It seemed a bit distant in the mids, a little warm in the bass, and too hot on top. Also, I always thought there were some possible resonance/crossover issues in the upper treble.

After coming across the 250Ti measurements in Chapter 18, I said to myself 'it sounds the way it measures'! If I had that info before I bought it, I might have passed in favor of something more neutral. Not to unduly pile on the 250Ti, it was a pretty decent transducer for its time. But I know now there are many more neutral transducers available today thanks to the technical tests done by ASR and on the other sites across the web. Thanks for advancing the art!

Very sad to hear how John passed. He left us too soon but I suppose there are worse ways to go than dying at home listening to music.

 

[TimVG]

Just a slight bit of departure from the above.

I certainly agree that playback needs to be as neutral and accurate as possible if the goal is to render what was captured on the recording as faithfully as possible.

I’m a big fan not just of music but of the nature of recording themselves. I really like the nuances of production and recording style… I can even enjoy “ bad” recordings (eg thin, coarse sounding or whatever ) if they are nonetheless interesting in character. So I’m not generally inclined to reach for tone controls or EQ to fix the nature of the original record recordings to a significant degree. I actually had a digital parametric EQ in my rig for something like 20 years that went essentially unused so I sold it a few years ago.

That said, it’s also my view that really accurate neutral playback is unnecessary to hear most of the nuances in recordings, including the effects of different microphones, mic placements, specific reverbs added to instruments and all the choices that go into musical production.

Just as Dr Toole suggests that the character of recorded acoustic cues in recordings tend to dominate over playback room acoustic cues, I would say that the recording and production choices in a recording tend to dominate over speaker or playback colorations.

One of countless examples : The new and interesting production techniques in the Beatles Sgt. Pepper translated through any number of consumer playback systems at the time, and through the years.

As a Rush fan, I’ve listened to the track Cygnus X-1 on countless different systems, from my family’s original stereo system, to boomboxes, Sony Walkman, car stereos, the various systems owned by my friends, all the systems I’ve been able to afford through my life… and the distinct production sounds of the opening of that track, the very particular percussive attack on Geddy Lee’s Rickenbaker bass, emerging out of that very distinctive alcove of trailing reverb, is always distinctive and recognizable.

And that’s why I’ve been able to hear and enjoy those distinctions across a wide variety of different types of loudspeakers in my room as well.

So it’s true that as an enthusiast I do sweat the small details, and I like to hear even the tightest distinctions in tracks. But this is a personal preference thing. I still try and keep the big picture view that most production choices translate across a wide variety of speaker/system performance, IMO.

I tend to agree. I like music first, the rest comes secondary. When I'm working on a recording myself however, then I'll do whatever I feel it needs. I use EQ, I use compression, I've even used saturation! All of my recordings are archival, so you're always working within a set of boundaries anyhow.

 

[MattHooper]

That was the folklore, to be sure, but scratch the surface and they were active experimentalists. I recall the early spaced-omni microphone phase, with very spacious delivery, ambiguous localizations - which was their objective for some recordings. But I just did a quick internet search and came up with these comments from some insiders:
“Jack Renner at Telarc used a 3 mike system with sometimes 2 more out in the auditorium for ambience. Bob Woods used 3 or sometimes 2 and again, sometimes 2 out in the hall. And yet on other recordings, particularly the jazz they used spot mikes. Sometimes when recording classical with individual singing, these individuals were also spot miked. On many of the Kunzel pieces which had sound effects, these effects were recorded separately by Mike Bishop and added in. It all just depended on the sound they were after.”

“Bobby Owsinski interviewed Michael Bishop - Episode #170 – Whiz Passes Psy, Gibson Buys Onkyo, And Engineer Michael Bishop | Bobby Owsinski's Inner Circle Podcast - recently. Mr. Bishop talked a fair bit about mic placement that was used while he was recording for Telarc. If I recall correctly he talked about 4 omni microphones spread out for "coverage" and possibly the use of some targeted mics. Much closer to The orchestra than you would expect.”

“Telarc recorded the Milwaukee Symphony performing Smetana's Ma Vlast back when we were recording them for Koss Classics. I recall part of the Telarc "sound" came from packing the musicians very tightly together, almost uncomfortably so.”

https://www.stereophile.com/content/jack-renner-telarc-direct-cleveland-page-2. This interview is an interesting perspective.

John Eargle, a noted recording engineer and much more, was a good friend of mine. https://en.wikipedia.org/wiki/John_M._Eargle
He recorded for the Delos label, and was highly regarded for his classical recordings. He described the process in detail, which began with some overall orchestra miking combined with "tasteful" close miking to amplify weak instruments, and to provide clarity for solo artists. These close pickup signals were adjusted in timing and electronic ambiance to make their contributions fit the overall soundstage. He was especially proud that these manipulations were not detected, and that the result was interpreted as a credible replica of a live experience. John was a significant factor in encouraging me to write the first edition of these books.

As I said, the circle of confusion is difficult to avoid. It is almost certain that all decisions were made while listening to forward firing loudspeakers in a small room.

I have no problem with any of the augmented techniques above, even if they are meant to produce a documentary style sonic presentation.

Some people seem to assume that certain minimalist micing will automatically capture a more natural sound when that may not be the case.

As you often point out, microphones don’t
“ hear” like two ears and a brain, and so the type of augmentation cited above (and used in many orchestral recordings) is very often in the service of making up for the liabilities of microphones, and trying to achieve a balance that actually sounds more natural as it might if you were there.

Anybody who has done many recordings, certainly recognizes this.

I’ve been recording sound effects for film and TV since the mid 80s, and especially when it comes to recordings in the wild, you become achingly aware of how the microphone is going to pick up things differently than how your naked ears are hearing it.

If I’m outside recording for for instance, a neighbourhood ambiance and somebody is walking even fairly distant but jangling some keys, while my brain can disentangle those keys from the environmental sounds I’m concentrating on, to the microphone those high frequency clashing keys may as well be right beside the microphone. You play the recording and all you’re aware of are those damned keys!

Often when I’m manipulating sound, it’s in the service of massaging recordings to sound more natural, real or believable than they sound untouched.

 

[Travis]

In short, for consumers it is not easy. Sorry. Fortunately, there is abundant evidence that we humans can adapt to enjoy music through significantly flawed loudspeakers - until we hear something better

Thanks you so much for your detailed answer.

I assume when we do ”hear something better” whether T audio show, friend’s house, dealer, we would want to verify with measurements?

 

[thewas]

That's still 'science' of course, but debateable. I'm leaning towards another reasoning along some lines Dr Guenther Theile draw into the blue. (He's also an academic with a ph/d, many prices, well reknown, but denialed on this board.)

Why do you tend to such incorrect accusations and generalisations? Where did this whole bord denial him?

or example, recommendation for slope of power response was published few decades ago in some books (I have). Some say it was by Harman, but I haven't found the reference from AES eLib. There are many significant, but missing features in Olive's preference work. So "preference" argument does not look very strong imo.

Could you please post some more specific details to such as this is really interesting.

 

[amirm]

Can you tell us a bit more about your experience there?

Sure. I have been there twice.

First time was as part of "Dealer Training" when I set up Madrona. What it turned out to be was an incredible full day of deep education from the entire Harman research team. The tail end was sitting through the double blind test with the speaker shuffler. Prior to testing, we had no idea what speakers were being tested/played. I heard one sample and thought it was a control as the sound was awful. Then came the second one and it sounded so much better. The third was close to it but maybe a bit better. We were given sheets to write our scores. Screen is pulled aside and I see that the worst sounding speaker was Martin Logan! Couldn't believe my eyes. It wasn't until that comparison that I realized the flaws it had. The second was hard to believe too as it was a JBL horn speaker. Somehow I considered those PA speakers and not worth considering. Yet it scored slightly better than the B&W that was double its price. That is the shot I post earlier:

Interestingly, all but one of us voted similarly. The one exception was a poor new Harman field engineer who was just hired. I think he picked the Martin Logan the best!

The second time I was there as part of elite acousticians who were given access to JBL Synthesis and its ARCOS EQ tools. I was there representing our company. Two other companies where there with their key employees who were installing very high end theaters. By then I had very much studied a ton of papers from Dr. Toole and his team and other researchers. So I was able to answer all the quiz questions easily during the class period. We then attended the same blind test but then followed to the other Room where Dr. Olive tested us with How to Listen software. It was there that I was able to easily outperform these other company employees but was beat by Dr. Olive.

These were incredible opportunities to touch and feel the research and spend time interacting with the very people who produced them. It was very generous of the team at harman to go so far out of their way to try to educate their dealers. Other companies would have just given you some product specs and what products to push for best margin.

Above got me started on a journey of reading some 300 to 400 research papers, joining ASA, etc. to get as much information as I possibly could. And of course ASR where I put all of this to practice and realizing the power of the research in predicting listener preference.

 

[amirm]

It would have been an experience, but not necessary. For example principles you and many others respect have been my basic design targets for 40+ years. Smooth or otherwise "controlled" directivity was praised in Finland already four decades ago. Addition to very early years is calculated power averages such as ER and PIR, but those don't give much new info to already smooth on and off-axis and power. Measurements have always been available, but reliability was lower in early years though my first DIY project was measured in the largest anechoic in Finland. Not in the reflecting room to get power response. Background and fundamentals should be quite okay...

But something started to happen ca. 2005. Some of my, and some commercial designs sounded unbelievably lame, attenuated or lost front edges of transients though on-axis, directivity and tonality were close to perfect. SPL capacity was also adequate comparted to listening level. A few locals participated in the listening sessions. Even 5" full-ranges, 2-ways and 3-ways with much lower order XO slopes - all with carefully balanced sound, were more dynamic in transients such as percussions and piano (left hand too). This much difference was not expected. Just small tonality change and vertical travel at HF (with non-coaxial design) as a part of precedence effect. The findings were discussed on a local forum.
I'm not sure, but public data and discussion might start something good also at Genelec, though criticism was not directy aimed to their product in question and it's excess group delay ca. three times longer than audibility threshold. They studied audibility with multiple files imitating different multi-way models and slope orders. Next generation (current) products have input FIR stages for phase linearisation. It's not necessarily full range minimum-phase due to possible latency requirements in HT and pro audio, but user can select longer latency to fix more lower frequencies. This might do some difference with 3-ways.

My advice for designers is that use your changes to eliminate also timing errors if possible - no matter what (obsolete and limited) studies or some authorities say. It's not on me if that is ignored or not even measured.

You seem to have good command of English language but what you write is extremely hard to follow. Partially it is your ego getting in the way constantly but it goes beyond that. It just seems like stream of consciousness at times.

But no, there is no substitute to sitting in a controlled double blind test of multiple speakers. Only arrogance would say that when it comes to speakers, listening in controlled testing has little value. What you say was known 40 years ago was likely due to the research Dr. Toole conducted at NRC, not because you or some ghost invented it that you can't name. Or reference.

Anyway, until you start writing clearly, linking to references, and lowering the bar of arrogance oozing out of your posts, there is little you can accomplish.

 

[MarkS]

Fortunately, there is abundant evidence that we humans can adapt to enjoy music through significantly flawed loudspeakers - until we hear something better.

I wonder if there is any research on whether "hearing something better" still works when the flaws are mild.

My personal experience is that my brain adapted to the (somewhat non-neutral but not terribly so) frequency response of my current speakers, and now more neutral speakers sound off to me.

I listen to a lot of live unamplified classical music (around two performances per month, on average), and my current speakers (purchased in 2021 before discovering ASR) sound to me like they are producing the correct tonality.

I do not believe that this effect "wears off" over time.

Is there any (scientific, non-anecdotal) evidence that I am wrong about this?

 

[Holmz]

…

The most significant aspect of a speaker is its tonality. No other aspect remotely comes close as everyone is sensitive to it. If I put in a dip at 1 kHz at -3 dB, every person will hear the difference in AB testing. Screw around with the parameters you are talking about, and you are lucky if they are detected at all.
…

Maybe, but probably not.
A 3dB peak may be more easily heard, but a dip is generally harder… especially when it is narrow.

In any case, I think that Klippel can generate both phase and group delay charts from its measurements, Erin does generate the group delay chart. And Amir doesn't think it's relevant.

The impulse response may also be relevant.
I’m sure if it was not relevant, then DIRAC would not be doing so well.

…
But something started to happen ca. 2005. Some of my, and some commercial designs sounded unbelievably lame, attenuated or lost front edges of transients though on-axis, directivity and tonality were close to perfect. SPL capacity was also adequate comparted to listening level. A few locals participated in the listening sessions. Even 5" full-ranges, 2-ways and 3-ways with much lower order XO slopes - all with carefully balanced sound, were more dynamic in transients such as percussions and piano (left hand too). This much difference was not expected. Just small tonality change and vertical travel at HF (with non-coaxial design) as a part of precedence effect. The findings were discussed on a local forum.

There is a lot of stuff in ^here^ that is not really captured with a Klipple scan.

I'm not sure, but public data and discussion might start something good also at Genelec, though criticism was not directy aimed to their product in question and it's excess group delay ca. three times longer than audibility threshold. They studied audibility with multiple files imitating different multi-way models and slope orders. Next generation (current) products have input FIR stages for phase linearisation. It's not necessarily full range minimum-phase due to possible latency requirements in HT and pro audio, but user can select longer latency to fix more lower frequencies. This might do some difference with 3-ways.

My advice for designers is that use your changes to eliminate also timing errors if possible - no matter what (obsolete and limited) studies or some authorities say. It's not on me if that is ignored or not even measured.

Thumbs up.

You seem to have good command of English language but what you write is extremely hard to follow. Partially it is your ego getting in the way constantly but it goes beyond that. It just seems like stream of consciousness at times.

Finns, and other Europeans, can often come across that way.
They do not shy away from ego and do not pussyfoot around a whole lot.

 

[Arindal]

the treble on these Revel loudspeakers had been boosted. The "trained" listeners turned the treble down exactly the amount necessary to restore a flat direct sound from the loudspeakers.

This explanation absolutely makes sense, as a flat overall in-room-curve plus increasing directivity index basically means an increasing on-axis response towards higher frequencies which no trained listener would accept. I wonder what caused such an enormous imbalance between direct and indirect sound. If the treble had to be reduced by 5 or 6dB above 5K in order to restore flat on-axis response, my suspicion would be that the end result showed an even more significant lack of indirect sound in this frequency band compared to 500Hz. Bringing me to the conclusion that even in a controlled environment with speakers being undoubtedly flat on-axis such imbalance is on a scale that it is clearly audible and qualifying as a massive coloration.

Which leaves my question open, which potential variant you personally and the results of your research would suggest to meet the ideal: A relative attenuation of treble in the reverb in a region of -7dB or more (´Harman loudspeaker curve´ according to this graph), or a completely frequency-independent ratio between direct and indirect sound (´Linkwitz dipole in-room curve´) over the complete audible spectrum? If you have no major disagreements with Siegfried Linkwitz, it sounds to me like the latter (we both do not have to agree with his particular method of achieving his aim, the dipole, to accept the aim itself as an ideal). Which would put a lot of speaker R&D efforts of the last 30 years under doubt, focussing on the ´smooth directivity´, accepting a steep and significant increase in d.i.

Unfortunately those in the cinema side of the business have little choice, as the monied sponsors scream "louder".

Everyone has a choice to always wear hearing protection. The linear attenuation models dubbed ´hearsafes´ actually became pretty good.

Preserving one's hearing is a challenge. Responsible professionals have their hearing checked regularly

I wholeheartedly agree. But as mentioned, a conscious generation of audio engineers are now in responsible position, with many of them reaching retirement age, never having exposed their ears to dangerous SPL unprotected in their whole life. I have strong hope that I will enter this phase with almost full sensitivity up to 15 or 16K, and I am seemingly no exception among the responsible group of former colleagues.

 

[amirm]

A 3dB peak may be more easily heard, but a dip is generally harder… especially when it is narrow.

If you can't hear that, then you certainly not going to care about anything else.

 

[amirm]

Finns, and other Europeans, can often come across that way.

We have a huge EU population here and none are acting this way. I am used to cultural differences, having worked with many nationalities around the world. And had them on my teams as well. He needs to act more professionally and be respectful of others around him especially in this forum where we have so many knowledgeable people contributing.

 

[amirm]

The impulse response may also be relevant.

I have been posting its sister measurement, the step response for countless speakers. I have yet to see someone point out some insight from it.

I’m sure if it was not relevant, then DIRAC would not be doing so well.

You have heard the befits of that apart its frequency response correction? Mind you, we are not talking about channel timing correction.

 

[kharan]

As an uncouth European with English as a second language , I'm going to suggest the following.

With DACs, amps and reviews for other electronics, @amirm you don't limit yourself to performance data that has been proven to be audible in double blind testing. Anything with a SINAD of beyond 100 dB and quite likely even less is basically transparent. I would even go as far as to say that 99% of modern amps have inaudible harmonic and multi-tone distortion. This is especially inaudible for distortion figures of 10 kHz and 15 kHz signals.

Yet, that data is captured and shown. The reason you do this, among others, if I recall correctly, is that you said it was that it showed the technical perfection or lack of thereof of a device even if it went beyond the widely demonstrated criteria for audibility.

To me that makes a lot of sense; when considering high fidelity, fidelity to the recorded signal is one of the stated goals, whether it is audible or not. It shows technical performance, it demonstrates a manufacturer's capability and willingness to reproduce the soundwaves as they were recorded, mixed and mastered.

The same goes for speakers, there might be less of an incentive to show ancillary data for which audibility hasn't been demonstrated rigorously, and for which inaudibility has been demonstrated in specific listening conditions, but fundamentally, phase data and IMD also inform about a speaker's abilty to reproduce the signal it is fed.

Furthermore, both IMD and time domain infidelity figures for speakers are magnitudes above what they are for electronics. Broadly speaking, they are in the range of demonstrated other distorsions (in the general sense), that have been demonstrated to be audible even if IMD and phase distorsion have not been demonstrated to be audible in conditions you would accept.

I would also add that there as devices you evaluate purely for technical performance like reclocking boxes. Any decent DAC should be able to reclock incoming signals are deliver an output that is jitter-free, making those boxes redundant at best.

In summary, like for electronics, I do think technical perfection and accuracy to the signal is a strong argument for showing IMD figures and phase information.

It would even educate people about the trade offs of phase accuracy vs directivity in traditional crossovers.


Side note, I understand that IMD is correlated to THD, but showing the results from IMD testing is a very direct and obvious way of showing how a speaker distorts with musical content at different SPL levels. Likewise, deviation from linearity at higher SPLs is similarly correlated with THD at different levels, but a compression chart is a more direct way of showing the SPL limits of a speaker.

 

[Keith_W]

I have heard 2 of the 3 speakers in that Harman demo (have not heard the JBL). Those Martin Logans beam like lasers and they sound horrible if you are not sitting perfectly on axis. I am also reminded of an older experiment done by Toole where he compared a Quad ESL speaker against two monopoles. I may have read somewhere else that in the Harman room, the front wall is covered with acoustic foam behind the curtain so the dipole is effectively a narrow radiating monopole.

To be honest I don't understand dipoles and why they sound good. The dipole sound (or Martin Logan sound) is not for me, but I can see why some people would like it. I suspect that the "open" and "airy" sound is created by front wall reflections, and if you remove that, you just get a bad sounding monopole.

As for the B&W ranking behind the JBL ... that's no surprise to me IMO.

Do you remember why you ranked the speakers the way you did?