Speakers that measure well and work in small room

[andrew]

Is there a view as to when a horizontal 3-way with an asymmetric set-up - such as KH310a or HEDD Type 20 - will and won't work well given that the left / right off-axis response will vary? Or is one just better off looking at a vertical three way (noting that a symmetric horizontal 3-way such as HEDD Type 30 is too big for my room)? The Adam S3V with its waveguides on the AMT tweeter + mid looks like another option that looks to have good horizontal dispersion whilst quite constrained in the vertical. Anyone have any experience on this model?

 

[Duke]

"Speakers that measure well and work in a small room."

The following is a mixture of opinion and sleep-deprived ramblings.

If we are going to look at measurements, we need to first figure out what "good measurements" would look like in this context. And unless we have a recording studio, imo "flat" is the wrong goal.

Bruel & Kjaer's ideal in-room curve is approximately flat for the first decade (20 Hz to 200 Hz), then slopes gently downwards by 3 dB per decade, so -3 dB @ 2 kHz and -6 dB @ 20 kHz. My understanding is that this curve is supported by Toole's findings.

Next let's take a quick oversimplified look at some arguably relevant psychoacoustic principles:

1. Early reflections tend to degrade clarity. The earlier they arrive, the more detrimental they tend to be.

2. Early sidewall reflections tend to broaden the apparent image size (considered desirable by Toole and undesirable by Geddes).

3. Relatively late reflections tend to enhance spaciousness and timbre without negative side-effects, assuming they are spectrally correct. "Relatively late" in the context of small rooms would be about 10 milliseconds after the first-arrival sound. Typically a wider pattern means more undesirable early reflections but also more desirable late reflections, so it's a trade-off either way.

4. Reflections whose spectral balance is significantly different from the first-arrival sound tend to degrade timbre.

5. Reflections are significantly more likely to be detrimental above 700 Hz than below 700 Hz. This is why the JBL M2 doesn't attempt to control the radiation pattern below 700 Hz; JBL could certainly have done so if they found it made a worthwhile improvement.

Let me try to translate the above into what sort of radiation pattern we might want in a small room:

1) A fairly uniform radiation pattern north of 700 Hz, or as close to that as we can get, is desirable. Given that early reflections are inevitable in a small room (assuming it's not virtually anechoic), let's at least get their spectral balance correct.

2) If you're a disciple of Toole, you may or may not want that uniform radiation pattern to be wide (imo you'll be trading off some clarity for a wider stereo image and a bit more spaciousness). If you're a disciple of Geddes, you will want the radiation pattern to be narrow enough to minimize early reflections. Toole and Geddes agree most of the time, but the desirability of early sidewall reflections is one place where they disagree. Yours truly sides with Geddes on this.

Imo if we have a small room and speakers whose off-axis response sucks (yes that's the proper technical term), nearfield listening is probably the best bet because it's really hard to fix the off-axis response through room treatments. But if the off-axis response is pretty good, then we have more options.

So to sum up: In my opinion we want a gently downward-sloping response curve along with a usefully uniform radiation pattern, and personally I'd tend to choose a narrow pattern over a wide one. Not that these are the only things that matter.

 

[andrew]

I

So to sum up: In my opinion we want a gently downward-sloping response curve along with a usefully uniform radiation pattern, and personally I'd tend to choose a narrow pattern over a wide one. Not that these are the only things that matter.

Thanks. Yes, I’m using Audiolense to tweak the overall profile to the B&K curve. Any ideas as to speakers with narrowish dispersion - unity horns or pro audio options?

 

[Duke]

Any ideas as to speakers with narrowish dispersion - unity horns or pro audio options?

Here are some narrowish-pattern professional studio monitor manufacturers: JBL, Ocean Way, Augspurger, Kinoshita, Meyer, Westlake, Reflexion Arts, and Dutch & Dutch. In addition most PA mains (excluding line array and J-array modules) have narrowish patterns.

And here are some on the home audio side that use horns or wavegudes: Classic Audio, JBL, GedLee, Klipsch, Volti, PBN (“M” series), Spatial Audio, Cessaro, Pi Speakers, Tyler Acoustics (Pro Dynamics line), Edgarhorn, Avantgarde, Oswald's Mill, eXemplar, Zingali, Seaton Sound, Usher (D2), Pure Audio Project, Hawthorne, Living Voice, Serious Stereo, Tannoy, Zu (I think), and yours truly. I consider a coaxial (or concentric) to be a horn or waveguide type speaker. (In this context, a "waveguide" is a type of horn optimized for pattern control rather than acoustic amplification).

Tekton doesn't use horns or waveguides but achieves a narrowish pattern by arraying multiple dome tweeters.

Many electrostats also have narrowish patterns, including Quad, Sanders Sound, SoundLab, Martin Logan, and King Sound. When positioned well out into the room, say about five feet in front of the wall, these speakers arguably have very favorable room interaction characteristics: Their narrow patterns minimize early reflections, and then there's enough time delay of the backwave's wall bounce that it is effectively a late-onset reflection (which is also spectrally correct). So we arguably get a "best of both worlds": Minimal detrimental early reflections + an extra helping of beneficial late reflections. Maggies and other ribbon or planar magnetic speakers likewise benefit from being placed well out into the room, but their patterns are fairly wide in the horizontal plane. Still imo their room interaction can be pretty good.

 

[andrew]

Here are some narrowish-pattern professional studio monitor manufacturers: JBL, Ocean Way, Augspurger, Kinoshita, Meyer, Westlake, Reflexion Arts, and Dutch & Dutch. In addition most PA mains (excluding line array and J-array modules) have narrowish patterns.

And here are some on the home audio side that use horns or wavegudes: Classic Audio, JBL, GedLee, Klipsch, Volti, PBN (“M” series), Spatial Audio, Cessaro, Pi Speakers, Tyler Acoustics (Pro Dynamics line), Edgarhorn, Avantgarde, Oswald's Mill, eXemplar, Zingali, Seaton Sound, Usher (D2), Pure Audio Project, Hawthorne, Living Voice, Serious Stereo, Tannoy, Zu (I think), and yours truly. I consider a coaxial (or concentric) to be a horn or waveguide type speaker. (In this context, a "waveguide" is a type of horn optimized for pattern control rather than acoustic amplification).

Tekton doesn't use horns or waveguides but achieves a narrowish pattern by arraying multiple dome tweeters.

Many electrostats also have narrowish patterns, including Quad, Sanders Sound, SoundLab, Martin Logan, and King Sound. When positioned well out into the room, say about five feet in front of the wall, these speakers arguably have very favorable room interaction characteristics: Their narrow patterns minimize early reflections, and then there's enough time delay of the backwave's wall bounce that it is effectively a late-onset reflection (which is also spectrally correct). So we arguably get a "best of both worlds": Minimal detrimental early reflections + an extra helping of beneficial late reflections. Maggies and other ribbon or planar magnetic speakers likewise benefit from being placed well out into the room, but their patterns are fairly wide in the horizontal plane. Still imo their room interaction can be pretty good.

Over the weekend I set-up a pair of Unity Horns and Bass Bins that had been put in storage. The set-up involved using Audiolense to create filters for x/o between the Horn, Mid-Bass and Subs that incorporated room EQ with target curve that has ~6dB rise from highs to bass. I'd forgot how good these speakers can sound in a small room and, interestingly, just how different they sound to 3-way speakers despite sharing the same overall room target curve. All this provides a good solution although it does present a near zero WAF solution so I'll take a look at the options above to see if any can solve this part of the puzzle.

 

[Duke]

Over the weekend I set-up a pair of Unity Horns and Bass Bins... I'd forgot how good these speakers can sound in a small room..

Thanks for trying that and posting the results. Yeah Unity horns in a 16 x 12 room probably looks like overkill, but imo their excellent pattern control is especially beneficial in a small room.

All this provides a good solution although it does present a near zero WAF solution so I'll take a look at the options above to see if any can solve this part of the puzzle.

You already have three subs in that room, right? Imo multiple subs are the right bottom-end solution for your situation because small rooms have the greatest need for modal smoothing. So if the bottom couple of octaves are already well taken care of, then I think you just need "main" speakers (with good pattern control) that go down far enough to meet the subs.

Given that there tends to be an inverse relationship between enclosure size and WAF, what ballpark enclosure size do you think you can get away with?

Also, which wall will the main speakers be on - the 16 foot wall, or the 12 foot wall?

Thanks!

 

[andrew]

You already have three subs in that room, right? Imo multiple subs are the right bottom-end solution for your situation because small rooms have the greatest need for modal smoothing. So if the bottom couple of octaves are already well taken care of, then I think you just need "main" speakers (with good pattern control) that go down far enough to meet the subs.

Two subs - one at front the other at back - that are fed with <80Hz mono signal. DIY super-chunks in one rear corner and MSR Springtrap (Helmholtz) bass traps in front corners. The bass could do with a bit of tweaking but, yes, the real issue is the "main speaker"

Given that there tends to be an inverse relationship between enclosure size and WAF, what ballpark enclosure size do you think you can get away with?

Open to floor-stand or stand-mount speakers. Main issue is speaker width with, say 400mm / 16" being about the maximum (although a really deep speaker won't work)

Also, which wall will the main speakers be on - the 16 foot wall, or the 12 foot wall?

The 12" wall. (At present the Unity Horns are toed in to the point that the their lines intersect ahead of the listening position).

Thanks

 

[Duke]

Two subs - one at front the other at back - that are fed with <80Hz mono signal. DIY super-chunks in one rear corner and MSR Springtrap (Helmholtz) bass traps in front corners. The bass could do with a bit of tweaking but, yes, the real issue is the "main speaker".

Very nice.

At present the Unity Horns are toed in to the point that the their lines intersect ahead of the listening position.

Perfect! But obviously you already knew that, otherwise you wouldn't have done it.

Open to floor-stand or stand-mount speakers. Main issue is speaker width with, say 400mm / 16" being about the maximum (although a really deep speaker won't work)... The 12" wall.

Okay, my suggestion is Spatial Audio. Maybe the M4 or M3 (the latter is 17" wide so maybe not). Both models use a large coaxial driver on an open baffle augmented by a woofer down low. Coming soon is the X5, six grand a pair, 15" wide, two open-baffle 12's + a large Heil air motion transformer. You may want to absorb the backwave to minimize early reflections, assuming you can't realistically pull them out into the room very far.

The dipole bass of the Spatials will further decorrelate the in-room bass energy when combined with your multiple monopole subs, theoretically improving the bass smoothness a bit more.

In a few months I hope to have something with a 12" square footprint tailored for small rooms, and designed to be used with a distributed multisub system south of 80 Hz or so.

 

[b1daly]

I can understand this but not how it's "non-electronic."

Because the original sound waves are generated algorithmically. You could hypothetically type in all the values.

Yes, the machines that generate these signals use electricity, but the electricity they use does not contribute to the waveform. You could compute the same sound on an old tube based digital computer or a Rasberry PI and the energy use patterns would be totally dissimilar.

“Traditional” electronic instruments actually shape an electrical signal that completely represents the sound.

So for a completely digital signal it will have never existed in a material form that has the musical shape embedded, until it hits the outputs of the DAC before the playback amps.

On a truly analog signal path of the old vinyl days there is a direct energy linkage (the “analog”) at every step. For acoustic instruments, sunlight>food>bio-mechanical-kinetic-electric...magnetic>electric...>mechanical (record cutter)...mechanical>electric...>mechanical>kinetic>bio-mechnical>electrical>signal received at brain.

Pure electronic instruments like analog synthesizers skip the bio-mechanical steps.

The electronic guitar and electric pianos like Fender Rhodes are actually “electromechanical” instruments. Some of the physical energy from the players body is used to energize a vibrating metal element, which is amplified

The introduction of digital technologies involves converting analog signals to digital and back, usually several times.

What is new about the latest digital instruments and modern production processes is that we have digitally generated music that does not go through a single analog to digital conversion, ever, in its entire existence! It will have only ever been anything remotely resembling sound at your playback system.

People work on the sound by listening to the most recent copy of the data. It’s basically a tap off the digital signal path.

What this means is that the philosophical foundations of signal transmission “accuracy” that are so often debated here are put on even shakier grounds. When you hear such sounds on your playback systems they are first generation, and by definition cannot be an innacurate representation of the “sonic signal”.

 

[b1daly]

Thanks for the interesting perspective, and I think there's logic to what you're saying, but I disagree with this point.

What I find impossible to come at is the idea that there is a specific (or a specific range of) "boxy speaker" sound, and that this is what is missing from accurate monitors.

Boxes can do all sorts of arbitrary things to the output of a speaker. In particular:

• Baffle dimensions and box depth determine the box's effect on the speaker's polar response. If you want to recreate a specific "boxy speaker" polar response to be consistent with how the music was supposedly intended to sound, wouldn't you need the box to have specific dimensions similar to those on which it was originally mastered?
• Box resonances similarly vary from box to box - indeed far more so than baffle dimensions. One box will create an entirely different set of completely arbitrary box resonances than another. How can replacing one arbitrary set of resonances with another entirely arbitrary set of resonances be more faithful to the recording?

In the end, I think that if you prefer the specific colourations of a box speaker, that's because you like that speaker's colourations - not because it gives you a more faithful representation of the original recording.



This is an interesting possibility I think, and the aspect of your discussion that I find a lot more convincing.

I don't know enough about analogue electronics, but I would have thought "non-deterministic" might be going a bit too far? What do you mean a mixer can't reproduce the same output to the same input? That if you turn it on one day and play the same recording through it, the output will be significantly/audibly different than it would be on the next day?

In terms of speakers, I would actually say that boxes tend to be among the more deterministic of speaker components, since box resonances tend to be determined by materials, structure, and dimensions, etc, all of which remain relatively stable regardless of external factors. Box damping materials may behave less deterministically as they tend to heat up to a more significant extent than MDF, I would say. Drivers and passive crossover components are actually likely to behave far less deterministically than boxes IMHO. Perhaps this behaviour in passive crossovers is something that tends to audibly distinguish them from active speakers..

Multitracked music used to be mixed in a fairly complicated analog device, for which great effort was put into getting reasonable linearity and low noise. But these were always limited. There was always crosstalk between the channels, which was affected by how close on the board they were. The whole system was run from a single power supply, so if you had a lot of channels running at high levels, it could affect power available to individual channels.

Basically these systems were chains of op amps that would cascade into mix busses. So by definition, since op amps are not perfectly logical linear, when sound A was mixed with sound B what resulted was not A+B but something more like A’+B’. When you added signal C you would get A’’+B’’+C’’. The original signals are gone.

With digital mixing, which dominates starting in about 2005, A+B=A+B exactly. You can literally extract the original A signal by mixing in the inverted version of B (-B).

This is completely “unnatural” and can result in very unconvincing mixes. A lot of effort has been put into creating the chaotic element of analog signal processes in DSP.

But most of such algorithms are “deterministic”. Sometime random elements are introduced, but these are not properly “chaotic.”

Not only is analog mixing not deterministic, it can only approach this “standard” asymptotically at ever increasing cost.

An area where this comes up a lot that really grinds my gears is that pop music has become completely dominated by cut and paste and looping techniques. So nowadays once you have one chorus of a song done, nobody bothers with any more recording. The exact same copy is pasted every chorus.

In the “olden days” if you wanted “pop perfection” you had to sing that chorus, harmonies and all, close to exactly the same every chorus.

Likewise with say a strummed guitar part. People will just play through the chords enough time to get a single iteration of the music block: the verse progression, chorus progression, then paste those in down the line. Or literally even just play one measure of a guitar part and loop that.

I’m telling, every song of and commercial significance is being produced like this, all over the world, pretty much every genre of pop that uses repetitive parts.

It’s not the same as a real-time linear performance, and I can tell. It doesn’t “feel” the same. The small imperfections, even in the most perfect performance, render each one unique.

So I’m extending this phenomenon deeper, to the problem of having bit identical data in a mix. Most people have no idea this is happening. They think new songs on the radio are kind of like old songs. They focus on the melody, words, chords, things like that. But I notice, and I’m sure it affects everybody.

This is the musical version of “the uncanny valley” people who work with digital graphics talk about.

Sounds generated this way are especially benefitted in my ear by imperfect playback systems. The more “uncolored” the playback system, the more clearly the icy digital unreality comes through.

Honestly, this type of music sounds awful to my ears on most systems, but man on studio monitors it’s unbearable. Clearly my taste on this matter is way out of fashion.

On the epic ASR thread on “Can You Trust Your Ears” which I read much of, Analog Scott stood mostly alone in making what I thought were some very excellent observations.

Most interesting was in the consideration of the Harman research about listeners prefer “accurate” speakers. This is a rather surprising result because, as he pointed out, music recordings themselves are full of all sorts of distortions.

It was pointed out that people are sensitive to and don’t like resonant peaks in speakers. Well resonant filters are used extensively in music production. Someone could put a very gnarly resonant peak in the mix, yet somehow a speaker that produced this accurately would be preferred over one that hid it better?

The underlying point he was trying to make is that the notion of “accuracy” which applies so well to devices that deal in electronic signals does not map well to the final stage of speakers.

His point is that there is no original signal to which accuracy can be maintained, and I agree with this. (Not taken to the extreme of course.)

One reason is that for all but the most purist specialty recordings, the closest one could get to the original “aural” event would probably be the final mix of a recording, before mastering.

But I can assure you that anyone who has tried to make recordings and distribute them knows that what they are delivering to the world is not “sound” but data. The sound coming out of the monitors is simply a tool to make the data manifest so it can be modified and shaped.

If all you had to do was make the mix sound good in the control room, this would be a lot easier of a task. But mixes don’t naturally “translate” from system to system. It’s complicated, but the engineers are trying to wrangle the creative work into a recording that will be effective across multiple playback systems. That will include the studio monitors but that’s more of a by product.

Especially in mastering, engineers try to shape the signal into something that translates well.

When this process works well, what you wind up with is a sort of “meta product” a kind of idealized version of the mix that doesn’t really exist anywhere as sound. A good mix will get the “essence” of the song across a bizarrely wide range of systems.

A paradigm example are the Beatles recordings. Most rock pop music lovers of certain generations “know” what Sargeant Peppers “sounds like”

If you listen to Sargeant Peppers on good studio monitors, especially the stereo mixes, it not only doesn’t sound so good, it sounds kind of “wrong”.

I need to read Tools book, but I remain surprised by the results of the research. (My guess would have been that people would have preferred some amount of boost of bass and treble.)

But in any case, because of the reasons I outlined above, attributing the listening preference to “accuracy” seems dubious. My first thought was that some speakers just sound better, specifically those that have a relatively flat frequency response and consistent off axis response.

It might have something to do with steady state resonant ringing being an unnatural type of sound, combined with familiarity of how acoustic sounds are generated in the natural world.

As far as your points about my preference for the coloration of specific speaker, well yes. But my “perception” of my perception is that I preferred “embodied” sound. The ideal of a speaker that “disappears” is bizarre to me. I see no reason for that to be desired outside of the goal of trying to create some kind of “holographic” illusion which could be neat.

But clearly the idea of “accuracy” when it comes to things like rock music is non-sensical. Rock music production is designed to create the sense of energy, literally sonic energy, when played at any level. Something cannot be “accurate” in a musical sense if the SPL relationship between any original acoustic source and the final recording is destroyed, deliberately, in production.

In the Harman research also wonder if the training, environment, familiarity with test material, and the task of generating a declared “preference” skewed the results to the surprising outcome.

While I am certain Harman makes use of this research across product lines, it remains the case that their is virtually no market demand for the kind of playback systems that were preferred in testing.

My observation is that the increasing expertise in audio engineering is being mostly applied to allowing for better playback experience in mobile environments, computer speakers, headphones, various types of portable and miniature speakers.

I don’t have any real world acquaintances who pay a lick of attention to experiencing “good sound” as exemplified by Harman. People are chucking “hi fi” systems. I’m far from an audiophile, yet I’m the one who goes around trying to dial my friends systems into something approaching neutral.

I change the speakers in our living room often and not once has anyone in my family noticed, except if the new speaker is much larger. It’s crazy, they literally don’t notice and don’t care how it sounds, as long as they can understand the words on TV, and get enough volume to have fun playing video games.

My step son listens to music in AirPods, or by walking around holding his little blue tooth speaker in front of his face.

I’m gettin old man...

 

[RayDunzl]

An area where this comes up a lot that really grinds my gears is that pop music has become completely dominated by cut and paste and looping techniques. So nowadays once you have one chorus of a song done, nobody bothers with any more recording. The exact same copy is pasted every chorus.

I don't care much for what you describe.

It's got no life in it, just sound (general opinion). Exceptions occur.

Sound is not a substitute for composition.

---

One of the very few "modern" recordings I bought was Random Access Memories.

I don't remember if it was before or after researching it a little. Anyway, I liked some of the tunes...

https://www.soundonsound.com/people/recording-random-access-memories-daft-punk

"Since their classic first two albums, Homework (1997) and Discovery (2001), Daft Punk have been regarded as leading lights of electronic dance music, yet they have waxed lyrical about the music of the '70s and early '80s, which they claim represents "the zenith of a certain craftsmanship in sound recording” and criticised music made with laptops, which "aren't really music instruments”. For Random Access Memories they had, they announced, gone back to the recording methods of the '70s and '80s, which involved not only a huge recording budget, but also the employment of great musicians from the era, and the use of high-end recording studios full of analogue equipment, all in order "to make music that others might one day sample”."

Seems like I listened to some of the other two albums mentioned, and discarded the result. I'll try them again.

*likes Donna Summer/Girogio Moroder I Feel Love

 

[andreasmaaan]

But clearly the idea of “accuracy” when it comes to things like rock music is non-sensical. Rock music production is designed to create the sense of energy, literally sonic energy, when played at any level. Something cannot be “accurate” in a musical sense if the SPL relationship between any original acoustic source and the final recording is destroyed, deliberately, in production.

Easy there! Nobody here is claiming that a speaker should be accurate to the original acoustic source.

The claim is that the speaker should be accurate to the recording. If a rock recording is mixed in such a way that “the SPL relationship between any original acoustic source and the final recording is destroyed, deliberately, in production,” then the speaker should reproduce that deliberate destruction, precisely as it is on the recording.

An area where this comes up a lot that really grinds my gears is that pop music has become completely dominated by cut and paste and looping techniques. So nowadays once you have one chorus of a song done, nobody bothers with any more recording. The exact same copy is pasted every chorus.

Some of this is no doubt due to lazy production practices, but have you considered the possibility that this cut/paste method could also be an intentional aesthetic? Albeit not one I'm particularly fond of either...

As far as your points about my preference for the coloration of specific speaker, well yes. But my “perception” of my perception is that I preferred “embodied” sound. The ideal of a speaker that “disappears” is bizarre to me. I see no reason for that to be desired outside of the goal of trying to create some kind of “holographic” illusion which could be neat.

I guess this is just a personal taste question. The creation of a holographic illusion is IMO exactly what stereo reproduction is about, so it seems strange to me to hear it called a "trick" or something undesirable

I don’t have any real world acquaintances who pay a lick of attention to experiencing “good sound” as exemplified by Harman. People are chucking “hi fi” systems. I’m far from an audiophile, yet I’m the one who goes around trying to dial my friends systems into something approaching neutral.

I change the speakers in our living room often and not once has anyone in my family noticed, except if the new speaker is much larger. It’s crazy, they literally don’t notice and don’t care how it sounds, as long as they can understand the words on TV, and get enough volume to have fun playing video games.

I guess our circles differ. Many of my acquaintances are interested in good sound, and of course many aren't.

Also, just because people are not aware of the Harman research, and wouldn't be able to understand it if they were aware, doesn't mean that their opinions about what sounds good don't more or less conform to Harman's findings. For example, whenever I've corrected systems to get a flatter anechoic response, all listeners I’ve dealt with have preferred the sound. It became quickly apparent to me when designing speakers that flat axial response and smooth off-axis response always sounds better (with any competently mixed music). People I’ve worked with, both technically minded and not, have tended to agree, even those who didn’t even understand what “flat response” meant.

My observation is that the increasing expertise in audio engineering is being mostly applied to allowing for better playback experience in mobile environments, computer speakers, headphones, various types of portable and miniature speakers.

The Harman research has application in these devices too. Why wouldn’t it?

If all you had to do was make the mix sound good in the control room, this would be a lot easier of a task. But mixes don’t naturally “translate” from system to system. It’s complicated, but the engineers are trying to wrangle the creative work into a recording that will be effective across multiple playback systems. That will include the studio monitors but that’s more of a by product.

IME the vast bulk of the work is done on monitors. Speakers that are thought to be representative of deficient everyday systems are used as a double-check, but that’s about it. Even this practice is questionable IMHO for the reasons I outlined in my earlier post: the deficiencies between different systems are always going to be different, so a single system's set of deficiencies can’t possibly be representative of all deficient systems.

 

[MetalheadRich]

PSB makes highly regarded and high quality speakers in your price range. Definitely worth checking out.

 

[b1daly]

Easy there! Nobody here is claiming that a speaker should be accurate to the original acoustic source.

The claim is that the speaker should be accurate to the recording. If a rock recording is mixed in such a way that “the SPL relationship between any original acoustic source and the final recording is destroyed, deliberately, in production,” then the speaker should reproduce that deliberate destruction, precisely as it is on the recording.

.

These are some good points.

But I don’t think folks here are realizing how surprising the finding on speaker preferences really are.

There is a lot more going on here than “meets the ear.”

The fact that the speaker as transducer will recreate the input signal with some fidelity is not sufficient to explain the results of the listening test.

This is easiest to see by considering that the “accuracy” of the speaker can’t be the cause of a listener preference because they don’t (shouldn’t) have any knowledge about the nature of the original signal.

This hints at where a form of bias could creep in with the use of constant test tracks.

 

[Juhazi]

^"The fact that the speaker as transducer will recreate the input signal with some fidelity is not sufficient to explain the results of the listening test.
This is easiest to see by considering that the “accuracy” of the speaker can’t be the cause of a listener preference because they don’t (shouldn’t) have any knowledge about the nature of the original signal. This hints at where a form of bias could creep in with the use of constant test tracks."

Yepyep, I don't put much weight on listening impression published by media or random hifi listeners, or even my friends. We do need test reports that include standardized measurements, to help choosing best speker for one's needs. Measurements easily lead to some kind of competition of "the best", but it would be misleading - based only on some criteria.

It has been said many times in this thread already, that eg. choosing low/high directivity in a small room is a matter of taste. Low bass extension with help of bass-reflex tuning might lead to excessive bass level and boominess too, depending on room. High spl with low distortion is hardly ever a real challenge at home.

Anyway, loudspeakers are the components that have widest variation of measured and perceived sound quality in hifi! So many choices and chances to go wrong or right way...

 

[Krunok]

We do need test reports that include standardized measurements, to help choosing best speker for one's needs.

The problem I see that comes before independent testing is that even the manufacturers of well known and expensive brands are publishing very basic specs. Pretty much the only thing you can find is frequency range with linearity (but without actual graph, specified smoothing etc), sensitivity, nominal impedance (that one is usually a joke, right?) and crossover frequency (usually withuot any additional details about crossover). I would like, for a start to see distortion graphs, imedance and phase vs frequency graphs, lateral and vertical response etc.

 

[b1daly]

Easy there! Nobody here is claiming that a speaker should be accurate to the original acoustic source.

The claim is that the speaker should be accurate to the recording. If a rock recording is mixed in such a way that “the SPL relationship between any original acoustic source and the final recording is destroyed, deliberately, in production,” then the speaker should reproduce that deliberate destruction, precisely as it is on the recording.



Some of this is no doubt due to lazy production practices, but have you considered the possibility that this cut/paste method could also be an intentional aesthetic? Albeit not one I'm particularly fond of either...



I guess this is just a personal taste question. The creation of a holographic illusion is IMO exactly what stereo reproduction is about, so it seems strange to me to hear it called a "trick" or something undesirable



I guess our circles differ. Many of my acquaintances are interested in good sound, and of course many aren't.

Also, just because people are not aware of the Harman research, and wouldn't be able to understand it if they were aware, doesn't mean that their opinions about what sounds good don't more or less conform to Harman's findings. For example, whenever I've corrected systems to get a flatter anechoic response, all listeners I’ve dealt with have preferred the sound. It became quickly apparent to me when designing speakers that flat axial response and smooth off-axis response always sounds better (with any competently mixed music). People I’ve worked with, both technically minded and not, have tended to agree, even those who didn’t even understand what “flat response” meant.



The Harman research has application in these devices too. Why wouldn’t it?



IME the vast bulk of the work is done on monitors. Speakers that are thought to be representative of deficient everyday systems are used as a double-check, but that’s about it. Even this practice is questionable IMHO for the reasons I outlined in my earlier post: the deficiencies between different systems are always going to be different, so a single system's set of deficiencies can’t possibly be representative of all deficient systems.

I know I haven't convinced you of my heterodox theories, and I'm not necessarily convinced myself. More like a working hypothesis.

But I thought I would lay out the basis of my ruminations.

-------------
The start of my current speaker research was the frustration I have with how bi-amped nearfield monitors sound. They work OK as monitors, but as far as an aesthetic experience, I find they lack, almost totally. I find it annoying to work on speakers that never sound great to me.

The speakers I have been gravitating to for home use are generally vintage designs, essentially rectangular boxes with drivers on the front baffle. Let's call these "hi-fi speakers" for lack of a better term.

I really didn't put much effort into my home system until just recently, so I'm kind of picking up from where I stopped with this in my early 20s.

So when I came upon the discussion of the speaker listening tests conducted by Harman here, I was really surprised, as my perception was that my personal preferences didn't seem to be lining up well with the results, to the extent I could draw a conclusion.

Since the speaker ranking order was consistent across different categories of listeners, this gave me a great pause. I've been around long enough to know that I am not exceptional in most things, and as a believer in science, I need to consider this point of view seriously.

The thing I am most lacking would be a chance to hear something like the Revel's that have been so well reviewed and appreciated here. I have some towers from the lower lines of Energy and PSB, and while I found these speakers unobjectionable, I'm looking for more than that!

My current contenders for a great home system are a pair of ADS L1290, for which I have had all the drivers rebuilt. It sounds quite excellent but does have some limitations. Mostly with poorly recorded sources, they can sound overly bright and harsh, and the tweeter becomes perceptable as a separate driver.

I've EQd them with a very high quality parametric to tweak some ringing in certain frequencies. Not much, just little notches. And then I roll off the treble just a bit by running a lowpass filter up around 22khz, which dips into the very high frequencies. I just don't like a lot of high frequencies, and some records just kill me with the treble.

-----------------------

There are two obvious differences between the "monitors" and traditional "hi-fi speakers."

--Monitors- bi-amped, self powered
--Hi-Fi speakers- passive crossover, separate amp

--Monitors- cabinets constructed of molded materials, with non-square edges, cabinets are heavy and highly damped, generate little discernable character
--Hi-Fi speakers- wooden boxes, rectangular, usually have a perceptible cabinet resonance


For the sake of discussion, I'll limit my comments as being relevant to the music I like the most, produced primarily between 1965 and 1990 or so.

This is the era where studio monitors were generally either soffit mounted with horns, or some kind of near-field or mid-field rectangular box, made from wood. Generally these used passive crossovers. Genelec started the trend of the self-powered, bi-amped nearfields, I think sometime in the 80s, but it didn't take off until mid-90s. In the 80s and 90s by far the most common studio monitor was the Yamaha NS-10.

Here's an interesting discussion of this weird little speaker.

https://www.soundonsound.com/reviews/yamaha-ns10-story

It didn't sound like much anything else, but it was a passive speaker, two way, in a rectangular box.

There's a lot of opinions about why this became almost ubiquitous in the 80s and 90s. My theory is that it is actually pretty "easy-to-please". It never sounds great, but it filters out a lot of lows and does a good job of "integrating" the sound as I've been trying to introduce the idea.

Especially when cranked, the whole box would kind of come alive, and your mix would "pop out" sounding very tight and forward. While this characteristic is not so great in a studio monitor, the reality is every recording engineer is under pressure to please the artist/client/producer, so working on a speaker that kind of makes everything sound OK is a crutch. (Many people had the theory that they sounded bad, so if you could get it sounding good on the NS-10 it would "sound good on everything." I never found this to be the case.)

Here's a discussion on Gearslutz about what were the popular studio monitors in the 1970s. Big horn-mounted monitors tended to be the "mains" but as far as smaller monitors go, I would say the JBL lines dominated. These were almost a classic "monkey coffin" design.

https://www.gearslutz.com/board/high-end/41263-what-were-they-monitoring-70s.html

I'm going to leave the soffit mounted mains with horns out of the discussion, as these speakers are unlike anything in the consumer world. They were also notorious for not being neutral, as they were essentially part of the structure of the room.

This is why the trend towards near field monitors started. They were an attempt to bring more standard and neutral monitoring to the production process. The theory is that by monitoring closer, at lower volumes, the direct sound of the speaker would dominate the reflected sound, giving a more consistent reference.

-------------------

Let's assume that the goal of the home playback system should be to accurately convey the art and intention of the artist.

As a thought experiment, my contention that by far the most accurate representation of recorded music would come from the primary set of monitors it was mixed on. (There's room for a variance here, as the monitors albums were recorded on, mixed on, and mastered on were usually different.)

Take a classic 70s record, like Fleetwood Mac Rumours, find out what monitors it was mixed on. Put them in your living room, and this is going to be a spectacular playback experience.

-------------------
The essence of such sound reproduction experience is not some attempt at creating an "invisible speaker." On the contrary, as the band was doing blow off the console, they were rocking out on actual speakers. That's the medium.

So the goal of a "disappearing speaker" seems antithetical to the artistic intention of such a recording.

------------------
My current theory about the issue I'm having with the studio monitor type speakers is that, while they are generally very forgiving to mixes, and sound "good" on most music, they never sound exactly "right" to me for most pop/rock music of this era.

In contrast, the vintage speakers I favor for fun often sound pretty damn close to what I think the music "should" sound like. But they are more hit and miss, as they have colorations and resonances that can make certain mixes sound not so great.

The ADS L1290 are actually fairly flat, so they represent at least a kind of neutral starting point. But they do have a characteristic sound, and I can generally perceive some cabinet resonance.

----------------
For my thesis to be right, it has to be the case that there is enough a "family" sound in speakers of a given design/era. So that you could find a home speaker that on average gives you a closer representation of what the artists experienced as they work. I tentatively believe this to be the case, and my main focus on a possible explanation is focused on how the cabinet's role in the overall presentation of speakers.

----------------
Here's my point of view on this issue of a speaker kind of "disappearing" and instead presenting a kind of "holographic image" of some other acoustic reality. In addition to the fact that I think it is impossible to correctly represent some types of music in this manner, as I discussed above, such systems have other weakness.

In general, sounds naturally occur in the world emanating from physical objects. Systems that are very rigid and present a kind of open "sound field" can sound "uncanny" or "fake" to me. This is kind of subtle, and perhaps entirely subjective perception, but I think there is something to it.

The functioning of our auditory system is based on evolutionary capacities, developed long before electronic reproduction of music. I don't think we can escape its influence in how we interpret the meaning of what we hear.

---------------------

Playback systems that are designed to primarily communicate meaning by creating a "sound image" in the 3D space of the listening environment also suffer from image instability. I don't know enough about what effects this. The most dramatic example I've heard is on planar speakers which can present a striking illusion of an actual physical object creating sound in front of you, but have a narrow sweet spot.

The Genelec 8030s in our main studio actually present a pretty stable image, the most of any monitors we've had set up there, even when walking around the room.

But the system I prefer for listening has speakers which can be clearly identified in the sound field. In this case, we are not worried about creating an illusion of some other 3D space. We have an actual 3D soundfield in the very room we are in! That is likely closer to what the artist intended if they worked on a similar system.

-----------------

In my perception, cabinet resonance helps with localizing the speakers in the space. This allows for a more stable image all around, because we can use our auditory, visual, and kinesthetic systems to understand that when we move, we are moving, and our sound source is staying still.

This is less of a concern if the listening environment is designed for a front, center, facing speaker thing. Which is great when possible. My absolute favorite presentation of music. But often I am moving about while listening, or have to place speakers in sub-optimal places because of how the room is.

-----------------

The goal of presenting recorded music is largely a project of creating an experience that is interesting for the listener. It cannot be about presenting a realistic acoustic representation of another acoustic event except in very limited cases. (And there are systems which can do a very good job of creating the illusion of an acoustic event from another space. Our studio monitors do this quite well.)

But most music is not about such literal representation. For rock music (in the larger sense) the volumes of the actual instruments are much louder than typical playback. This presents a very hard challenge for the producers, and I find that it is easier to create this illusion on speakers that do have a sense of "embodied" presence. Ideally this experience could be entirely encoded in the stereo signal and brought back to life by a perfectly neutral speaker system, but my perception is that this remains an unobtainable goal. Though the best producers and engineers do a damn good job. As an example, listen to Andy Wallace's mix of Nirvana's Nevermind. That record jumps out of any damn speakers its played on and rages right there!
--------------------

What is to be said in favor of relatively flat playback systems, with flattish off-axis sound, "accurate" systems so to speak, is that on average, you can be confident that such a system will be the most flexible, and give the best experience overall.

So essentially, if you like certain kinds of music a very accurate system, one with flat frequency response and good off-axis behavior, and no to little cabinet coloration, you give up the most accurate representations possible, for representation that is never quite right, but always pretty close. Something like that.

--------------------
The studio monitor sounds I find so frustrating to most people sound rather stunning. Our studio monitoring is nothing to write home about in the world of studios, but when "civilians" stop by and we play them music they are familiar with, it is likely the best sounding system they have ever heard.

My frustration is that these systems work well as monitors, are very easy to describe as "sounding good" but never give me the playback experience I am confident is closest to the original intention of the artists I like the most.

---------------------

The vision that Toole lays out about more standardized monitoring in studios and playback systems is interesting and has a lot to recommend it. But I doubt that these super neutral, overly damped designs can ever really deliver the effect I'm talking about.

But my sense of modern recording that is it has almost thoroughly abandoned the aesthetic values of music production that I like best. I would venture the aesthetic values are lost on almost everyone my age (53) or older.

Much of it sounds quite terrible to my ear, especially rock music production.

In the pop and Hip-Hop realms, we have seen an explosion of imaginative production techniques. Which I stand in awe of, even though I don't connect emotionally with the music, I can hear a lot of interesting things going on. They are often using deliberate digital distortions that I never imagined would become popular. They actually hurt my ears. For example in the dubstep genre. It sounds cool, but it's just too much for me to process comfortably. Possibly, in the same way, the original designers of electric guitar amps could not have forseen that the deficiencies in their designs would become the most desirable attributes of guitar amps

I suffer some hearing damage, so I listen at low volumes, and limit my exposure to sounds that hurt my ears when I can, as part of a hearing preservation strategy.

Anyway, what is happening here is that recorded music has essentially departed completely from being related to acoustically generated signals.

Time marches on. My concerns will become less and less relevant (have already become), as the style of music I prefer for pleasure listening has lost popularity with the younger generation, who drive the trends of the music business.

------------------------

Ironically, when looked at a whole system, the process of music creating and reproduction, which is a complex technical and sociological process, has moved ever farther from the ideals of Toole and his colleagues.

I am sure we do hear the results of the research by Harman on preferences in speaker voicing. To my ear, the small, often portable systems, that dominate home listening today are much closer to flat than ever before, and in smaller packages. These are amazing accomplishments of engineering.

It is these devices, plus headphone/earbuds, sound bars, and car auto which will dominate the playback environment for the vast majority of listeners for the foreseeable future.

I've been amazed that stereo reproduction has been knocked from it's primacy as a playback format. The portable systems are often point sources, or close to it. What they are doing with the stereo signal is unclear to me. Cars present a quasi-stereo-surround environment. The sound in modern car systems can be quite good, though it seems people tend to adjust these systems to boost bass and treble.

I don't think true surround sound reproduction will ever become that popular, as it is too complicated and inconvenient to set up. The trend towards "sound bars" for audio-visual systems seems inexorable. The sound generated by devices is quite strange when I've heard them. We had a Sony soundbar with sub-woofer, and I could never get that thing to sound good on anything. The best that could be said was that it did sound better than the TV speakers.

But I know there is great determination in the industry to generate simple systems that provide a surround like experience, by using clever driver placement and DSP.

Headphones have seen a resurgence, ironically spearheaded by the Beats line, which is cool.

It is unclear to me what can be generalized as being the optimum type of playback system for the primarily synthetic sounds which I predict have forever replaced acoustic instruments as the driver of styles in pop music. (Acoustic sounds are still used, but are usually recontextualized.)

----------------
So there is a bunch of thoughts, loosely connected by the project of understanding what makes for a "good" speaker, for a given purpose.

If anyone bothers to read this far, thanks!

 

[RayDunzl]

You're welcome.

 

[andreasmaaan]

Interesting post @b1daly, thanks Too long to address each point but let me try to cover a few things you've said.

--Monitors- cabinets constructed of molded materials, with non-square edges, cabinets are heavy and highly damped, generate little discernable character
--Hi-Fi speakers- wooden boxes, rectangular, usually have a perceptible cabinet resonance

A thought occurred to me. Could it be that your preference is not for speakers that have perceptible cabinet resonances, but rather for speakers that have larger, wider baffles (and the resultant effects that has on polar response)?

Here's my point of view on this issue of a speaker kind of "disappearing" and instead presenting a kind of "holographic image" of some other acoustic reality. In addition to the fact that I think it is impossible to correctly represent some types of music in this manner, as I discussed above, such systems have other weakness.

In general, sounds naturally occur in the world emanating from physical objects. Systems that are very rigid and present a kind of open "sound field" can sound "uncanny" or "fake" to me. This is kind of subtle, and perhaps entirely subjective perception, but I think there is something to it.

The functioning of our auditory system is based on evolutionary capacities, developed long before electronic reproduction of music. I don't think we can escape its influence in how we interpret the meaning of what we hear.

The thing is, with stereo or multichannel reproduction, the system is being asked to create a phantom image at a point in space where there are no speakers. I don't see how a system that embodies the sound in the speakers could not hinder this.

But I understand you are talking only about your own perception/preference ofc (but also see my previous point).

The essence of such sound reproduction experience is not some attempt at creating an "invisible speaker." On the contrary, as the band was doing blow off the console, they were rocking out on actual speakers. That's the medium.

So the goal of a "disappearing speaker" seems antithetical to the artistic intention of such a recording.

You'll never get a pair of speakers in your room that happen to "appear" (i.e. distort, in the broadest sense of the term) in just the same way as the band's speakers did in the studio the recording was done in. Any differences are going to be more or less arbitrary. Like I argued earlier, the overwhelming likelihood is that your speaker/room's arbitrary flaws are going to be completely different from the arbitrary flaws of the speakers/room in which the recording was made.

I think that, as we both more or less agree, on average the most neutral speaker is most likely to sound best with most recordings.

Of course, within "neutral" there are a range of possibilities when it comes to speakers, primarily to do with polar response, which as I mentioned earlier is the very thing I suspect this whole discussion may boil down to...

I've been amazed that stereo reproduction has been knocked from it's primacy as a playback format. The portable systems are often point sources, or close to it. What they are doing with the stereo signal is unclear to me.

It's actually very simple. They are just mixing the two channels to one mono channel using either a simple analogue circuit or a DSP. Or in other cases, there are two small full range drivers on opposite ends of the device, playing a very narrow version of stereo, sometimes with additional DSP effects processing to try to increase the perceived stereo width.

I am sure we do hear the results of the research by Harman on preferences in speaker voicing. To my ear, the small, often portable systems, that dominate home listening today are much closer to flat than ever before, and in smaller packages. These are amazing accomplishments of engineering.

Agree

The goal of presenting recorded music is largely a project of creating an experience that is interesting for the listener. It cannot be about presenting a realistic acoustic representation of another acoustic event except in very limited cases. (And there are systems which can do a very good job of creating the illusion of an acoustic event from another space. Our studio monitors do this quite well.)

I totally agree with this. The speaker can't reproduce an original acoustic event, it can only reproduce a recording.

It is unclear to me what can be generalized as being the optimum type of playback system for the primarily synthetic sounds which I predict have forever replaced acoustic instruments as the driver of styles in pop music. (Acoustic sounds are still used, but are usually recontextualized.)

Also an interesting question, but basically following on from my previous point: speakers were never intended to reproduce acoustic events; they were only ever intended (or indeed able) to reproduce recordings (in the broadest sense, i.e. including material that was not actually recorded because it was synthesised, etc).

Back then, all we had for speakers were recordings. Today, still all we have are recordings. So speakers are still being asked to do exactly the same thing. It's only the content of the recordings that is ever-changing.

 

[b1daly]

Interesting post @b1daly, thanks Too long to address each point but let me try to cover a few things you've said.



A thought occurred to me. Could it be that your preference is not for speakers that have perceptible cabinet resonances, but rather for speakers that have larger, wider baffles (and the resultant effects that has on polar response)?



The thing is, with stereo or multichannel reproduction, the system is being asked to create a phantom image at a point in space where there are no speakers. I don't see how a system that embodies the sound in the speakers could not hinder this.

But I understand you are talking only about your own perception/preference ofc (but also see my previous point).



You'll never get a pair of speakers in your room that happen to "appear" (i.e. distort, in the broadest sense of the term) in just the same way as the band's speakers did in the studio the recording was done in. Any differences are going to be more or less arbitrary. Like I argued earlier, the overwhelming likelihood is that your speaker/room's arbitrary flaws are going to be completely different from the arbitrary flaws of the speakers/room in which the recording was made.

I think that, as we both more or less agree, on average the most neutral speaker is most likely to sound best with most recordings.

Of course, within "neutral" there are a range of possibilities when it comes to speakers, primarily to do with polar response, which as I mentioned earlier is the very thing I suspect this whole discussion may boil down to...



It's actually very simple. They are just mixing the two channels to one mono channel using either a simple analogue circuit or a DSP. Or in other cases, there are two small full range drivers on opposite ends of the device, playing a very narrow version of stereo, sometimes with additional DSP effects processing to try to increase the perceived stereo width.



Agree



I totally agree with this. The speaker can't reproduce an original acoustic event, it can only reproduce a recording.



Also an interesting question, but basically following on from my previous point: speakers were never intended to reproduce acoustic events; they were only ever intended (or indeed able) to reproduce recordings (in the broadest sense, i.e. including material that was not actually recorded because it was synthesised, etc).

Back then, all we had for speakers were recordings. Today, still all we have are recordings. So speakers are still being asked to do exactly the same thing. It's only the content of the recordings that is ever-changing.

I’ll think about the baffle issue.

My understanding of the main issue with baffle size is the “baffle step” problem. When the frequency gets down to the point where the wavelength is longer than the baffle width, and the sound becomes significantly less directional.

This results in a dilemma, because dealing with the discontinuity requires adding bass frequencies. But since the sound energy is still in the room, this adds overall low energy frequencies to the total sound energy, making the final frequency response heard by the listener hard to predict.

Something like that. Presumably it presents and issue for off axis response.

I know that the trend in speaker design has been toward narrow baffles, moving the frequency point where sound becomes more directional from the speaker higher. I’m not sure what benefit this brings. I guess it brings more of the curve into a range where there is less attenuation of off axis sound?

The other issue is edge diffraction which I don’t understand. The descriptions I’ve seen seem to describe almost as a reflection from the sound traveling across front of baffle, when it hits the edge, some bounces back causing comb filtering interference.

The modern studio monitors tend to have curved baffles, which I think is supposed to minimize the problems with diffraction.

I totally get what you’re saying: that my hypothesis proposed is only relevant if there actually is some kind of “family sound” for different approaches to speaker designs. Essentially that is my tentative hypothesis. That is my subjective experience.

It’s really bizarre, most music I listen to on studio monitors just sounds wrong.

I’m pretty confident I could point out the significant elements of this perception and another experienced listener would get it right away.

One thing about my subjective perception is that it’s very stable.

I’ve gone down the rabbit holes being convinced that an amp sounds “thin” only to be listening on another day and it sounds fine. This is not like that.

This archetypal experience of sound obsessives confusing themselves is what has given the “trust your ears” method a bad name.

Many audiophile claims of perceived sound differences fail miserably in AB or ABX testing. That’s a first level. Second level is that if you can distinguish between sounds, is your preference stable.

The Harman research is so interesting because it connected subjective preference testing with objective measurements, and demonstrated that these results apply across listener demographic in terms of ranking order, and that they are repeatable.

I’m confident I can easily distinguish speakers in a blind test. I’m not quite as confident my sighted preferences would correspond to blinded preferences.

What I need to do for at least sanity check is find a home speaker that is designed in accordance with the principles espoused by Toole, and see what I think.

Do folks think any of the older Revel models would qualify? I would like to find something $1000 or less, used.