• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

A Broad Discussion of Speakers with Major Audio Luminaries

If you read the paper this came from, it is clear that the baseline from which the listeners started is a flat in-room curve - meaning that 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. The also liked a small bass boost - possibly a room interaction - but we know that too much bass is a forgivable sin :). The inexperienced listeners were off on their own planet. This experiment needed many more listeners to contribute usefully to a perspective on personal preferences, and the authors admit it.

A bit late to the thread here, but in support of the study; I've had some discussions with an experimented studio builder (both building studios and installing systems in existing studios). Typically midfield / main monitors, not nearfield systems to my understanding.

He said that they almost always ended up with an in-room response that had a 3-6dB tilt from the bass to the midrange area, and a 6-8dB total tilt. Sometimes there was more of a bass hump, and other times a pretty gradual / continuous slope. This sounded most natural and "right" both to him and the various studio owners. He attributed the differences in total dB drop and type of slope in part to preference and in part to the room and the speakers and how they interacted.

I find the above to be an interesting data point, and it seems to imply that the findings in the paper (despite limited number of listeners) likely has merit.
 
a decently implemented low budget mass market design with continuously increasing DI can reach similar when equalised to flatish listening window response as the biggest SP change happens lower, exemplary done on the KEF Q7 Meta with 5 filters to make the LW flatish, this is corresponding EIR with those filters

This is with completely flat on-axis response? Looks rather unlikely to me. Achieving such a balanced EIR while keeping the direct sound neutral, seems impossible due to increasing d.i. of the speaker in the aforementioned range. The difference is ´only´ +3dB of directivity index between 500 and 8,000Hz (thanks to large distance between the midwoofers and resulting lobing), but with several steps up from one octave-broad plateau to the neighboring band. Particularly what happens above 3K is of relevance.

According to Dr. Toole, trained listeners would in this case decrease the overshoot direct sound treble using an EQ. The dull reverb cannot be compensated electronically, although it might look better than the EIR with factory response, which I would not expect to deliver a tonal balance like the CD speaker I have mentioned:

EIR_Q7Meta.jpg


You don't need though a special and expensive design to reach such a corridor in that region,

I agree, there are affordable speakers offering sufficient approximation towards constant directivity in the relevant band. But, in contrary to your suggestion, they do not show continuously increasing d.i., but, well, a more or less constant directivity. Surprisingly simple example (ignore the bass bump as it can be EQ´ed):

EIR_SBC700.jpg


He attributed the differences in total dB drop and type of slope in part to preference and in part to the room and the speakers and how they interacted.

Dr. Toole has explained that trained listeners show prioritized preference for a neutral direct sound response in the frequency range our brain can distinguish direct from reflected sound. If you keep that in mind plus calculate common increasing directivity index of many speakers, that kind of makes sense.

It does not say much, though, if neutral indirect soundfield would be preferred, if it is possible to achieve it at the same time with linear on-axis response.
 
This is with completely flat on-axis response? Looks rather unlikely to me. Achieving such a balanced EIR while keeping the direct sound neutral, seems impossible due to increasing d.i. of the speaker in the aforementioned range. The difference is ´only´ +3dB of directivity index between 500 and 8,000Hz (thanks to large distance between the midwoofers and resulting lobing), but with several steps up from one octave-broad plateau to the neighboring band. Particularly what happens above 3K is of relevance.
It isn't impossible and this is as said with per PEQs flattened LW response, the DI is referring to the total sound power while the EIR corresponds to more the early reflections and the corresponding index, here is the corresponding LW response for above EIR:

1750151197555.png


The total DI is not a good indicator for the EIR and thus also not for the tonality in good/serious listening conditions where the direct sound is in balance with the reflected sound.
 
It isn't impossible and this is as said with per PEQs flattened LW response, the DI is referring to the total sound power while the EIR corresponds to more the early reflections and the corresponding index, ...
The total DI ...
Hi guys, just a quick thought, could you consider the RT60 performance of ordinary listening rooms? So when it comes to counting individual dBs and the frequency dependence in detail, I would be quite curious to see what comes out.

Perhaps it would help to measure the diffuse sound, say 30 cm from the main listening position. Depending on the circumstances, it may look different, just for comparison.
 
Hi guys, just a quick thought, could you consider the RT60 performance of ordinary listening rooms? So when it comes to counting individual dBs and the frequency dependence in detail, I would be quite curious to see what comes out.

Perhaps it would help to measure the diffuse sound, say 30 cm from the main listening position. Depending on the circumstances, it may look different, just for comparison.
Different to what? When I do measurements around my listening position I mainly use the MMM and there is usually at good match to the EIR above 500 Hz and other people here have shown similar.
 
... good match to the EIR above 500 Hz and other people here have shown similar.
Does your room resemble the average the EIR is based on? Is MMM method compatible to what the standard prescribes? I wonder because the standard, if my recollection holds, wants the microphone to be directed towards the ceiling, while being compensated for that. Alas, I don't have the paper at hand.
 
Does your room resemble the average the EIR is based on? Is MMM method compatible to what the standard prescribes? I wonder because the standard, if my recollection holds, wants the microphone to be directed towards the ceiling, while being compensated for that. Alas, I don't have the paper at hand.
Two times yes, see exemplary in below links (there are more from others and me if you search this forum), although the orientation and compensation of typical 1/4" measurement mics mainly makes a difference above 8 kHz where it doesn't much matter (as below they are quite omnidirectional).


 
Two times yes, see exemplary ...
Thanks for pointing me to the standard CEA-2034 (again), really appreciated. Section 13 shows a comparison of calculated data versus a measurement. The calculation weighs different contributions, e/g from the listenng window, early reflections ... . The weighting seems to represent the real measured grand total pretty well, while the room is small considering a Schroeder frequency around 350Hz. Weighting confirmed, but so far not so in a 'typical' room. Anyway, your measurements confirm the weighting in your room(s), nice.

I'm a bit lost now. I was under the impression that e/g listening window is about the direct field - clearly a fallacy of mine. Have to re-read, sorry.
 
Hi guys, just a quick thought, could you consider the RT60 performance of ordinary listening rooms? So when it comes to counting individual dBs and the frequency dependence in detail, I would be quite curious to see what comes out.

Perhaps it would help to measure the diffuse sound, say 30 cm from the main listening position. Depending on the circumstances, it may look different, just for comparison.
There is no significant diffuse sound contribution in domestic size rooms, that's why RT60 is not really relevant.

Cheers
 
A bit late to the thread here, but in support of the study; I've had some discussions with an experimented studio builder (both building studios and installing systems in existing studios). Typically midfield / main monitors, not nearfield systems to my understanding.

He said that they almost always ended up with an in-room response that had a 3-6dB tilt from the bass to the midrange area, and a 6-8dB total tilt. Sometimes there was more of a bass hump, and other times a pretty gradual / continuous slope. This sounded most natural and "right" both to him and the various studio owners. He attributed the differences in total dB drop and type of slope in part to preference and in part to the room and the speakers and how they interacted.
This is an example of what NOT to do: i.e. "calibrate" or evaluate a room based on steady-state room curves at frequencies above transition/Schroeder. Did he know the anechoic performance of the loudspeakers he was using? If not the room curves cannot be properly interpreted. My observation of many practices in pro audio is that there is skepticism about "science" - it is the art that matters, so why are measurements relevant? It gets especially dangerous when one hears that recording engineers must "like" their monitors - judged by playing their own masters of course, thus perpetuating whatever biases exist. Rarely are small room bass problems adequately corrected. I would like to think that these "traditions" are fading, and that rational scientific principles are penetrating the field. There are some excellent neutral monitors in the marketplace, and measurements to prove it.

In general discussions we tend to focus on mid/high frequency loudspeaker directivity in explaining the downward tilt, but forget that it is basically established by low-frequency directivity and room reverberation. This is discussed at monotonous length in an open-source JAES paper - anybody can read it.

Toole, F. E. (2015). “The Measurement and Calibration of Sound Reproducing Systems”, J. Audio Eng. Soc., vol. 63, pp.512-541. This is an open-access paper available to non-members at www.aes.org http://www.aes.org/e-lib/browse.cfm?elib=17839.
The paper dwells on cinema sound because it is in serious need of help, but the principles apply equally to small rooms, and examples are shown - including, mea culpa, the much quoted Olive et al. personal preference curves (Figure 14) which have not been well explained and have too few listeners to be usefully predictive. As explained earlier in this thread, the trained listeners adjusted the tone controls to restore a flat direct sound at mid/high frequencies. End of story.

In the real world circumstances differ from theoretical abstracts, and in this respect I like to show a photo of one of my listening rooms from years past, revealing what the loudspeaker "sees". It is evident that most of the mid/high energy has little opportunity to participate in an early-reflective sound field (small rooms have no "classical" reverberation and no critical distance). All rooms will be different of course.
Figure 12.3 views from the tweeter rev.jpg


Subjectively preferred bass balance is often influenced by playback sound level, the equal loudness curves causing listeners to prefer some bass boost just to be able to hear the lowest frequencies.
 

Attachments

  • 1750160041360.jpeg
    1750160041360.jpeg
    412.1 KB · Views: 30
There are interesting discussions in this thread. :)

How can you get a live recording of a large orchestra to sound as close to reality as possible at home in small listening room with a HiFi system?
I'm ignoring the slightly impossible task of getting the same dynamics, but rather the illusion that it sounds at least a little like a large live orchestra.
Is higher SPL at the expense of slightly worse FR preferable? Or the other way around?
Maybe it's possible to achieve this with omni speakers with their sound dispersion characteristic? Then at the expense of pinpointed accuracy in the sound image? Omni = sound porridge?

Then of course small listening rooms are designed differently in shape plus taste and preference in what you want and can furnish them with. That affects the sound. With that, I don't know if it's possible to generalize answers to my questions, but that doesn't stop me from asking them anyway.:)
 
Last edited:
There is no significant diffuse sound contribution in domestic size rooms, that's why RT60 is not really relevant.
Of course I was not referring to RT60 exactly but ...

In the real world circumstances differ from theoretical abstracts, and in this respect I like to show a photo of one of my listening rooms from years past, revealing what the loudspeaker "sees". It is evident that most of the mid/high energy has little opportunity to participate in an early-reflective sound field (small rooms have no "classical" reverberation and no critical distance). All rooms will be different of course.
I wonder what predictive powers the DI etc numbers in the spinorama actually have. There was a discussion about a dB or so, where the tilt begins and things like that.

I'm ignoring the slightly impossible task of getting the same dynamics, but rather the illusion that it sounds at least a little like a large live orchestra.
I also wonder, why the classic's orchestra is so prominent, but foremost why 'illusion' is the word. I never - actually, ever fell into an illusion when listening to a playback. Not even into the feeling of immersion. As that oddball, should I donate my ears to science (after)? I think that expactations regarding the powers of a stereo (surround) system are often pretty much exaggerated. I long for proper frequency response, which is easily achievable today, a smooth reverberant field without hot spots in space and frequency, sufficient dynamic range, lean but extended bass, combined with lowest non-linear distortion and all within reasonable limits and budget. That makes me 'decode' the music, and I'm happy.
 
This is an example of what NOT to do: i.e. "calibrate" or evaluate a room based on steady-state room curves at frequencies above transition/Schroeder. Did he know the anechoic performance of the loudspeakers he was using? If not the room curves cannot be properly interpreted. My observation of many practices in pro audio is that there is skepticism about "science" - it is the art that matters, so why are measurements relevant? It gets especially dangerous when one hears that recording engineers must "like" their monitors - judged by playing their own masters of course, thus perpetuating whatever biases exist. Rarely are small room bass problems adequately corrected. I would like to think that these "traditions" are fading, and that rational scientific principles are penetrating the field. There are some excellent neutral monitors in the marketplace, and measurements to prove it.

Without knowing exactly what he was doing (I wasn't there), I feel like I misrepresented what he meant. He custom builds speakers, so they are his own design. While I'm not 100% sure, I don't think he was "forcing" a curve, but likely adjusting below Schroeder when needed based on speakers that were likely anechoically pretty flat. I know his designs has been measured in anechoic chambers.

Anyway, my point was to confirm that speakers with these slopes worked well, not that he was tuning to match the curves (or that anyone else should).
 
Anyway, my point was to confirm that speakers with these slopes worked well, not that he was tuning to match the curves (or that anyone else should).
What slopes?
 
I also wonder, why the classic's orchestra is so prominent, but foremost why 'illusion' is the word. I never - actually, ever fell into an illusion when listening to a playback. Not even into the feeling of immersion. As that oddball, should I donate my ears to science (after)? I think that expactations regarding the powers of a stereo (surround) system are often pretty much exaggerated. I long for proper frequency response, which is easily achievable today, a smooth reverberant field without hot spots in space and frequency, sufficient dynamic range, lean but extended bass, combined with lowest non-linear distortion and all within reasonable limits and budget. That makes me 'decode' the music, and I'm happy.
It's been so long since I've been to a concert with a big band/orchestra that it seems more like an illusion. In a way, what I listen to at home becomes the real thing, so to speak. It's a reality of course. That being said without delving into the philosophical aspects of what reality itself is.

In addition, you have to be in the right frame of mind to enjoy live music. When listening at home, you can enjoy music regardless of whether you're "turned on" for music listening or more in the mood for some background music or music at a not-so-high volume listening to some nice big band. :)
It's a waste of money, after all, not being an active listener at a concert, I think.
 
I wonder what predictive powers the DI etc numbers in the spinorama actually have. There was a discussion about a dB or so, where the tilt begins and things like that.
DI is an academic concept with its origins buried in history - describing a characteristic of the loudspeaker, not how it interacts with any specific room or predictive of the sounds arriving at a listener's ears - which is what matters. When I look at a family of spinorama curves I look first for evidence of resonances in the spatially averaged curves and then, if that looks good, I consider the spectral balance (broadband smoothness) of the off-axis curves as reflected in the directivity indices. Arguably the early-reflections DI is more important than the classic DI based on sound power. I have never specified numerical deviations, nor do I have any reliable experimental evidence to do so. As described in my last post, the tilt begins at bass frequencies and above mid/high frequencies the direct sound physically and perceptually takes precedence.
 
Of course I was not referring to RT60 exactly but ...


I wonder what predictive powers the DI etc numbers in the spinorama actually have. There was a discussion about a dB or so, where the tilt begins and things like that.

I also wonder, why the classic's orchestra is so prominent, but foremost why 'illusion' is the word. I never - actually, ever fell into an illusion when listening to a playback. Not even into the feeling of immersion. As that oddball, should I donate my ears to science (after)? I think that expactations regarding the powers of a stereo (surround) system are often pretty much exaggerated. I long for proper frequency response, which is easily achievable today, a smooth reverberant field without hot spots in space and frequency, sufficient dynamic range, lean but extended bass, combined with lowest non-linear distortion and all within reasonable limits and budget. That makes me 'decode' the music, and I'm happy.
With all due respect, maybe you should listen to the music, instead of the equipment. I'm only half-joking. A willing suspension of disbelief, necessary for the illusion, requires will.

I have attended and been a part of orchestra performances in a variety of halls my whole life, and listened to thousands of hours of recorded orchestral music. A neutral system will easily convey the space of the performance space, if the recording possesses it (and that is not assured by any means). If the recording doesn't possess it, and I can't tolerate the music without it, then I ought to buy a digital reverberation device, which is as close as it gets to a plausible facsimile of a reverberant performance space. But that lack of tolerance is, in turn, intolerable to me. Where am I sitting? If I'm on the back row of the orchestra, where the tuba player usually sits, my impression of the overall sound will be very intensely dominated by the trombone section and the tympani, and next in line by the string basses. Those instruments are around me and in my near field. Plus the stage usually sounds quite dry to the musicians in a good hall, because the reverberation heard by the audience is "out there," where the direct sound has had a chance attenuate down to something closer level of some of the reflected sound. On stage, the direct sound is far stronger in proportion to reflections, and stages are either highly reflective to the audience (by using a shell) or they are dead (with curtains that trap reflections). Thus, the back-row musicians only hear a bit of what reflects back to them from the far wall of the auditorium.

If I'm in the audience on Row 10, then I might actually be below the lip of the stage with the back-row musicians out of view. I'll heard the strings dominantly by direct sound, and the winds, brasses, and percussion to a greater extent by reflection from the stage shell or auditorium walls. If I'm on the front row of the mezzanine, I'll hear the whole orchestra in probably the most balanced way. The back row of the mezzanine will be affected by very close direct reflections from behind me. I've sat in all these places, and in all these places, the orchestra sounded like...an orchestra...and Shostakovich sounded like...Shostakovich.

The range of differences I have experienced in person surround the experience I get at home with a good recording, where there is some effects of the performance hall reverberation recorded along with the music. But there is nothing I can do to put me on the front row of the mezzanine, because I can plainly see that I'm on my sofa holding a glass of Scotch and reading silly posts on ASR on my phone while listening. But even so, the orchestra sounds like an orchestra and Shostakovich sounds like Shostakovich. One of the points Dr. Toole makes in his book is that humans are exceptionally good at separating primary sound from reflected sound, no matter in what proportions they are presented.

But I'm extremely sensitive to percussion or high strings that sound crispy (from distortion caused by clipping) and tubas sounding like euphoniums or French horns sound like trombones (from colored frequency response or too much second-order distortion--the kind audiophiles often seem to admire). I can also detect very slight changes in spectral tilt cause by the use of a tone control, even when I don't notice the 72-Hz null in my room that is only 8 or 10 Hz wide. We should think of a broad spectral tilt in the same way as dissimilar levels in a comparison. The correct protocol for comparisons is to match levels accurately, using a capable RMS-reading voltmeter, because even very slight differences in level affect our perceptions and preferences. But if the choices under comparison have subtly different spectral tilts, then that is impossible, because the difference in levels will vary by frequency. Even a couple of tenths of a dB could make a difference. But for music, I can't otherwise hear harmonic distortion less than about 1%, which is good, because very few speakers can do better than that.

Here is my listening room as of a couple of years ago, without equalization:

1221_BeforeEQ.jpg


Notice the dip between 200 and 300 Hz. That's a room effect. Notice the dip at a little above 1 KHz--that's a small dip in the response of the speakers, as noted in the anechoic testing (this is the "listening window" response as reported by Soundstage from testing in the NRCC anechoic chamber):

frequency_listeningwindow.gif


This listening window is averaged from these on and off-axis measurements:

On axis,15, and 30 degrees off-axis:
frequency_on1530.gif


45, 60 and 75 degrees off-axis:
frequency_456075.gif


That can be fixed by equalization, as Dr. Toole suggests above. And I can also use EQ to do something about that dip in the mid-bass.

I use a Yamaha YDP2006 digital parametric equalizer, because I'm old-school and prefer real devices over software, especially for speaker and room correction that will be applied to all playback. After EQ:

1221_AfterEQ.jpg


There are still nulls in one channel at 52 and 90 Hz, and in the other channel at 72 Hz, caused by cancellations in the room. And there's a null next to a spike in the upper hundreds. Those nulls under 100 Hz might get help from properly integrated subs, if I would use them, but frankly none of these narrow nulls distract me in the least. I get more of a change from turning my head or reaching over to get my glass, and that much is true even if I'm sitting on the front row of the mezzanine in a concert hall.

If a recording needs a broad spectral adjustment because of the way it is mixed, that's what tone controls are for.

Rick "has to listen to stuff he doesn't get lost in to hear the equipment, which isn't much that is musical" Denney
 
Last edited:
The ones I referred to / described in my previous post.
But it makes no sense to say speakers have those slopes. And as Dr Toole said, it makes no sense to make the room have those slopes. So what exactly is meant when, quote, "speakers with those slopes work well"?
 
With all due respect, maybe you should listen to the music, instead of the equipment. I'm only half-joking. A willing suspension of disbelief, necessary for the illusion, requires will.
Thanks for the nice read about real experiences. (Yes, I read it all 8-)

What you tell in the first sentence could have been from me. See my "signature". Belief is everything, mostly. But I wouldn't call it 'illusion'. In regard to the orchestra, I'm not into old music, it's just the age. You know, as science progresses, so does art.

Back to slopes:
But it makes no sense to say speakers have those slopes. And as Dr Toole said, it makes no sense to make the room have those slopes. So what exactly is meant when, quote, "speakers with those slopes work well"?
All too true. The slope is of lesser concern, and with it it's shape. I would say, it's a matter of taste. Alas, without knowing the room's properties quite well, analytically, or from experience with other speakers, predictions from the spinorama are a bit speculative.

That is what I would have liked to point to. If the slope etc is of concern, the room's properties should be likewise to the same degree. But I don't see that often
 
Back
Top Bottom