Why evaluating the sound of a single speaker is essential

[JanL]

All good points, but as a counterpoint, if you have a system that adds a bump here or there, or sprinkles on a little distortion of one type or another, you will undoubtedly find some music that sounds better on this system than any other systems, but there will be other pieces of music that sound terrible on this system. A transparent system will give you the best playback for the broadest selection of music.

It's why I have multiple headphones and IEM's and my system where I listen to music 90% of the time is based around pro gear and pro plugin (Dsoniq Realphones). I can change the EQ and spatial character of the reproduction instantly. If something doesn't sound right, I will briefly switch to my reference playlist to make sure I'm not imagining things and then I will grab some controls to make my enjoyment of the music better. Some mastering jobs are very dull sounding to me. I guess they lose definition when pumping the levels.

 

[JanL]

Nothing wrong with having opinions. The problem has been claiming mine is better than yours.

It's kinda like sex. No one cares what you do in your own home. Trying to get everyone in town to do it your way is .... well, kind of a problem.

It's not a good analogy. I think people should do it their way, and not what the doctor prescribed as the best way to do it.

 

[RexrothPigeon]

I think

You have opinions, cool, but none of your posts have anything substantive to do with the actual thread ("why evaluating the sound of a single speaker is essential).

If your point is that measurements and empirical research aren't relevant to your personal listening preferences (or that they shouldn't be relevant to other people's personal listening preferences), that's fine, but this thread isn't about that. Maybe try the "are measurements everything or nothing" thread (but probably read it first, since all your views have come up many times already.

 

[MattHooper]

Why do you want your speakers...whose task is merely to convert electric signal into acoustic waves -- to add colors?

If it’s entertaining or pleasant… sure. Why not?

Do you want your DACs coloring?

Currently no (I use a benchmark DAC). But for many years I preferred a CDP in my system that seemed to color the sound in a way that I really loved (Meridian 508.20… also detected using blind testing). Again, if it gives me more pleasure, why not?

Your cables?

No.

Your amps?

YES

You can add all sorts of colors after the fact, by myriad means.

Sure. But some means are more interesting/entertaining/desirable for some people than others. You might prefer a car for getting between A and B finding it more practical, and I might prefer a bicycle.

Maybe what you really want to be is a *sound engineer*.

Well…. I manipulate sound for a living, so I guess extending that to my sound system isn’t a stretch.

 

[JanL]

You have opinions, cool, but none of your posts have anything substantive to do with the actual thread ("why evaluating the sound of a single speaker is essential).

If your point is that measurements and empirical research aren't relevant to your personal listening preferences (or that they shouldn't be relevant to other people's personal listening preferences), that's fine, but this thread isn't about that. Maybe try the "are measurements everything or nothing" thread (but probably read it first, since all your views have come up many times already.

Respectfully disagree. In response to one of the posts by the OP, I pointed out that there is a lot of distortion in the material, which makes judging speakers (single or multiple) more difficult these days. Yes it was a tangent, but it was very much topic adjacent. Then I replied to others quoting me.

I'm sure there is nothing new in the 900+ posts of this thread and all of it was talked about before anyway.... We don't need to be 100% productive at all times here. Do we?

 

[RexrothPigeon]

Respectfully disagree. In response to one of the posts by the OP, I pointed out that there is a lot of distortion in the material, which makes judging speakers (single or multiple) more difficult these days. Yes it was a tangent, but it was very much topic adjacent. Then I replied to others quoting me.

I'm sure there is nothing new in the 900+ posts of this thread and all of it was talked about before anyway.... We don't need to be 100% productive at all times here. Do we?

We don't, but this thread suffered a bad derail the last couple days that had to be cleaned up by mods and the actual useful info in the thread is getting buried by drive-by personal preference posts and other tangents

 

[MattHooper]

In it this subject is, I would like to think, exhaustively examined and explained, using scientific evidence. Long story short, human listeners are increasingly less sensitive to sound quality degradations in loudspeakers as the channel count is increased from one to two and two to five. The overall result of adding channels is more spatial and directional information, which can be highly entertaining, but the end result is that the binaural hearing system has difficulty separating the spatial cues in the recordings, from the spatial cues in the listening room. Increasing channel count increases the persuasion of the recorded space. As a result listeners are unable to discern timbral errors caused by resonances in loudspeakers with the same sensitivity as in mono/single-loudspeaker comparisons.

About this… and I’m sure this information is somewhere in this thread, but….

To what degree do room reflections come into this when evaluating sound from a single speaker?

One of the effects of a resonance in a speaker can be a slightly boxy tone and/or a “stuck to the speaker” presentation.

Do I presume correctly that room reflections can also affect the ability to discriminate resonances in a single speaker? And if so are room reflections just as polluting of this discrimination as adding surround speakers?
This would seem to be where some of the “hearing through the room” phenomenon (above 2-300Hz?) comes into play?

I was thinking of MY MBL OMNIS and how they effortlessly floated hard panned sounds free of sounding stuck in the loudspeaker. Wondering about untangling how much of that would be due to room interaction and reflections versus low colouration in the speaker itself. (I would add that since I have lots of control over my room reflections, in this case the floating free effect was there whether I was cutting down lots of reflections or not).

 

[Newman]

I think people should do it their way, and not what the doctor prescribed as the best way to do it.

The key point lies behind your phrase, "best way to do it", meaning do what? In your case, you are extending music production into the home listening phase. And it's easy to see why:-

...I have been doing audio professionally for many years and I have opinions. I also know that the music I listen to was made by people with their own opinions, which I might or might not agree with.

Okay, so you are an audio pro, and you enjoy doing your own post-production at home as an overlay on the production work embedded in the recordings, which you don't rate. Got it.

But that is of scant interest to us, we the general audiophile, who are rarely interested in becoming post-producers, and are more usually keen on sound reproduction in the home, as a way to maximally appreciate good recorded productions. And there are plenty of good productions out there, it seems, in the opinion of the consumers, if not you. The audiophile subgroup of all consumers, generally, want that. They/we want to hear that excellence, as produced.

There is no reason for an audiophile to blindly trust the music producers. As Floyd Toole pointed out repeatedly, they are a part of the circle of confusion. They make decisions based on the gear and rooms they have and more importantly their own taste.

Nobody "blindly trusts the music producers". Audiophiles have been critical of mixing and especially mastering for, well, forever. Part of the 'audiophile quest' has always been to find and share good and great productions of recordings of music that we love. And they are out there. And, although much of what you say about pop music productions has been an issue and recognised by audiophiles, that doesn't stop us from wanting to hear the good ones via excellent reproduction. And many, many audiophiles have quite wide tastes in music into other genres, where production issues have been far less concerning and the music is often quite acoustic (vs electronic) and we don't want to hear that via some fake colouration or imbalance added by our home reproduction gear.

I agree with @RexrothPigeon that your recent posts here are way off topic to this thread, even though you perhaps thought it might be relevant to point out that recordings so often contain distortion that it's a waste of time to evaluate any objective attribute of our home speakers, including singly. You are almost arguing that audio science has no place in our choice of gear. Well, that argument has been much argued, over-argued in fact, and isn't convincing.

PS have you actually read any edition of Toole's book? I'm guessing not. Because he discusses at length things like psychoacoustics and widespread listener preference for specific sonic aspects, also the sound engineers as a group and their tendencies, and also their predilection for sonically coloured speakers in the studio, especially in the past. This predilection can change: the Circle of Confusion can become less confused, and progress is being made, but it is a work in progress.

cheers

 

[Floyd Toole]

Do I presume correctly that room reflections can also affect the ability to discriminate resonances in a single speaker? And if so are room reflections just as polluting of this discrimination as adding surround speakers?

This has been discussed in other postings (somewhere in ASR) and is directly addressed in Section 5.8 in the 4th edition. Early refections in small rooms improve the detection of resonances by presenting the ears with repetitions of the direct sound. It is the recorded large-room reflections in stereo and multichannel that detract from the detectability of resonances. Figure 5.13 explains it.

 

[Floyd Toole]

Dynamic range compression is a given for 99% of music released in the last 50 years. Even if the original recording was not compressed in any way, the mastering used compression and limiting to adapt to the medium, be it tape, vinyl or digital.

Now, not many audiophiles might realize this, but distortion is the spice of music production. It is generously sprinkled on individual tracks and mixes. It takes the form of clipping the components in the signal path, as in "driving the preamps", hitting the tape", "slamming the mix bus" or using their digital equivalents in the form of various plugins for computer based recorders. The same happens during mastering stage. Everyone wants a full sounding, loud record and mastering engineers are giving their customers what they want.
A pop music mastering signal chain might have 5 to 10 processors in series, each adding a bit of harmonic distortion to the final product. That way they get it to sound full and loud without resolving to hard limiting, which was popular in the 90's and early 2000's but produced harsh sounding masters.

So now we can have well-engineered speakers or headphones, and super clean amplifiers, and we listen to sounds that are enhanced through added harmonics in amounts that would disqualify any part of the reproduction chain in a typical review on this website. We are talking 10-20% THD on every track at all frequencies. This is why the best reference tracks come from the time before DAWs and plugins were ubiquitous. Back then they used compression to make things sit in the mix correctly, but they didn't use deliberate distortion as much. It was there but sparse. Now it's everywhere.

Thanks. Yes, it is well known that distortions of various kinds are used to give recordings and components within them specific timbres. No problem, it is art. However, your observation that this practice is almost a norm in pop music helps to explain why complaints of distortion in competently designed full-sized loudspeakers are not common - small powered units are a different population. With rare exceptions, loudspeakers are perceptually differentiated on the basis of their linear performance, not their non-linear performance, and it is magnitude response, not phase response that conveys the perceptual information.

 

[MattHooper]

This has been discussed in other postings (somewhere in ASR) and is directly addressed in Section 5.8 in the 4th edition. Early refections in small rooms improve the detection of resonances by presenting the ears with repetitions of the direct sound. It is the recorded large-room reflections in stereo and multichannel that detract from the detectability of resonances. Figure 5.13 explains it.

Ah! Right, thanks, that’s ringing a bell now about the early reflections.

But I hadn’t remembered that it was the actual reverb/acoustics in the recordings that reduced discrimination. I thought there was something about adding speakers in a room per se that reduced discrimination of speaker resonances, but I guess this is more about overwhelming the ear with the recorded acoustics versus the room acoustics?
All of which suggest that “dry” recordings would be preferable for use in discriminating resonances in loudspeakers?

 

[Floyd Toole]

Ah! Right, thanks, that’s ringing a bell now about the early reflections.

But I hadn’t remembered that it was the actual reverb/acoustics in the recordings that reduced discrimination. I thought there was something about adding speakers in a room per se that reduced discrimination of speaker resonances, but I guess this is more about overwhelming the ear with the recorded acoustics versus the room acoustics?
All of which suggest that “dry” recordings would be preferable for use in discriminating resonances in loudspeakers?

Sorry, not so. Reflections/reverberation in recordings also increase the audibility of resonances (repetitions again) but only when delivered through a single loudspeaker. It is the arrival of binaural large room spatial information that makes the difference. Only then do we perceive large room spatial envelopment, which is what appears to attract our attention away from the small room reflections. It is an interesting "room within room" perceptual process.

When delivered through a single loudspeaker dry and wet recordings, and technical signals are all treated the same way. There is no spatial perception other than that of the listening room.

Do you have the 4th edition? It is all there, and simplified in the website slide shows.

 

[MattHooper]

Sorry, not so. Reflections/reverberation in recordings also increase the audibility of resonances (repetitions again) but only when delivered through a single loudspeaker. It is the arrival of binaural large room spatial information that makes the difference. Only then do we perceive large room spatial envelopment, which is what appears to attract our attention away from the small room reflections. It is an interesting "room within room" perceptual process.

When delivered through a single loudspeaker dry and wet recordings, and technical signals are all treated the same way. There is no spatial perception other than that of the listening room.

Do you have the 4th edition? It is all there, and simplified in the website slide shows.

Ah so I was well off track there.

I don’t have the book, but I will look for the website. thanks.

 

[ChrisG]

A thought:

Part of the enjoyment of a good playback system is that a stereo image can be presented. Sources appear to be in-between the speakers (or occasionally beyond), or at least, that's the intent.

In this regard, a "poor" playback system will still obviously sound like two discrete sound sources, and it's trivial to point at them with our eyes closed.
By contrast, a "good" playback system will make it difficult to discern where the sources of the sound are. Instead, we perceive instruments spread in front of us.

I haven't seen anyone seriously attempt to measure the "disappearing" act that good playback systems sometimes pull off, and yet I trust we've all heard such a thing at least once or twice.


I've noted that the PDF attached to the first post of this thread states that loudspeakers which may rate poorly in mono-testing can sometimes rate well in stereo/multi-channel formats. The conclusion drawn is that mono-testing must be more stringent, but I'm unconvinced that's the whole story: as far as I can tell, mono-testing cannot tell us about the stereo image, the "disappearing act", that a loudspeaker might achieve. A single speaker will, unsurprisingly, sound like all the instruments stacked on top of each other, being emitted from that box over there.

Is it possible that some poor-in-mono speakers provide a good enough stereo image that their subjective ratings improve specifically because they're being used in pairs? Would this help to explain the mono vs stereo rating discrepancies?

Apologies if this post is rambly or difficult to follow. It's been re-written a couple of times.

 

[MattHooper]

A thought:

Part of the enjoyment of a good playback system is that a stereo image can be presented. Sources appear to be in-between the speakers (or occasionally beyond), or at least, that's the intent.

In this regard, a "poor" playback system will still obviously sound like two discrete sound sources, and it's trivial to point at them with our eyes closed.
By contrast, a "good" playback system will make it difficult to discern where the sources of the sound are. Instead, we perceive instruments spread in front of us.

I haven't seen anyone seriously attempt to measure the "disappearing" act that good playback systems sometimes pull off, and yet I trust we've all heard such a thing at least once or twice.


I've noted that the PDF attached to the first post of this thread states that loudspeakers which may rate poorly in mono-testing can sometimes rate well in stereo/multi-channel formats. The conclusion drawn is that mono-testing must be more stringent, but I'm unconvinced that's the whole story: as far as I can tell, mono-testing cannot tell us about the stereo image, the "disappearing act", that a loudspeaker might achieve. A single speaker will, unsurprisingly, sound like all the instruments stacked on top of each other, being emitted from that box over there.

Is it possible that some poor-in-mono speakers provide a good enough stereo image that their subjective ratings improve specifically because they're being used in pairs? Would this help to explain the mono vs stereo rating discrepancies?

Apologies if this post is rambly or difficult to follow. It's been re-written a couple of times.

I believe that Dr Toole has pointed out (and many of us have experienced) the fact that sound played through a single speaker with low or no audible resonances will, along with what has learned about tonal balance, will also exhibit better spatial qualities.

In other words: play a recording of an instrument on a single poor speaker with lots of resonances and the resonant quality will tend to make it sound “ stuck in the speaker” or very, obviously coming from the speaker.

Whereas on a well designed speaker, it will seem to float more free of the speaker cabinet. I described it as: instead of the sound feeling like it’s stuck in the speaker, there’s a sensation that the speaker is an object around which the sound of the instrument is simply occurring. And also, in good speaker designs, even mono playback, can’t have more obvious spatial qualities and layering than on a poor speaker design. All of which helps predict those qualities in stereo as well.

 

[Floyd Toole]

A thought:

Part of the enjoyment of a good playback system is that a stereo image can be presented. Sources appear to be in-between the speakers (or occasionally beyond), or at least, that's the intent.

In this regard, a "poor" playback system will still obviously sound like two discrete sound sources, and it's trivial to point at them with our eyes closed.
By contrast, a "good" playback system will make it difficult to discern where the sources of the sound are. Instead, we perceive instruments spread in front of us.

I haven't seen anyone seriously attempt to measure the "disappearing" act that good playback systems sometimes pull off, and yet I trust we've all heard such a thing at least once or twice.


I've noted that the PDF attached to the first post of this thread states that loudspeakers which may rate poorly in mono-testing can sometimes rate well in stereo/multi-channel formats. The conclusion drawn is that mono-testing must be more stringent, but I'm unconvinced that's the whole story: as far as I can tell, mono-testing cannot tell us about the stereo image, the "disappearing act", that a loudspeaker might achieve. A single speaker will, unsurprisingly, sound like all the instruments stacked on top of each other, being emitted from that box over there.

Is it possible that some poor-in-mono speakers provide a good enough stereo image that their subjective ratings improve specifically because they're being used in pairs? Would this help to explain the mono vs stereo rating discrepancies?

Apologies if this post is rambly or difficult to follow. It's been re-written a couple of times

The original reason for testing in mono, the one that started this entire discussion, has to do with identifying resonances in loudspeakers. These are technical faults that should not exist if we have any interest in hearing the correct timbres of recorded voices and instruments. It turns out that in the experimental method that is most revealing - double-blind equal-loudness comparisons of 3 or 4 loudspeakers - listeners could recognize resonances more readily than when listening in stereo or multichannel. It was simply a subjective way to identify audible resonances so that loudspeaker design engineers could attempt to eliminate them. These tests were time consuming and very expensive to conduct and it was fortunate that the NRCC and Harman could afford the expense of doing them. As a result of decades of these tests, we have been able to develop a measurement method - the spinorama - that visually reveals the resonances, and that has the further advantage of showing impressively high correlations with subjective evaluations of overall sound quality. As a result, the tests no longer need to be done, and they have ceased.

Stereo imaging is an entirely separate consideration. While evaluating hundreds of loudspeakers in the tests just described listeners often commented that the highest rated ones - those with the fewest audible resonances - had a tendency to "disappear", allowing a sense of distance and space that was associated with the recording. Poor loudspeakers drew attention to themselves. You may be experiencing something like this.

But, as described in some detail in the 4th edition of my book, stereo itself is a problem. It is not an encode/decode process. It is merely a two-channel delivery system, and what is delivered is what someone in a recording studio created. There are no standards. Stereo recordings that employ hard-panned left and right images/instruments are delivering essentially mono signals that will be strongly localized to the loudspeaker - it is the source. These are combined with phantom (double mono) images that have very different timbres and spatial characteristics. Other recordings avoid extreme hard pans, and these present a spatially "softer" imaging. When images appear outside the left-right span of the loudspeakers it is very likely accidental, the result of binaural processing In the mix, or you have multidirectional loudspeakers energizing strong side-wall reflections. Some classical recordings present a wide soundstage because of the high proportion of reflected (decorrelated) sounds in the mix; others have been known to deliver spatially compressed, in my view, simply bad recordings. There are no standards.

So, the listener normally has no way to separate the variables. This is a reason to start with timbrally neutral, technically accurate, loudspeakers as they will provide the most advantageous platform from which to hear what was recorded whether listening in stereo or multichannel. Of course, you may not always like what was recorded - it is art.

 

[krabapple]

I am mystified by anyone who expects forward firing loudspeakers to 'disappear' all the time.

IME experience speakers 'disappear' in a content-dependent fashion -- even within a track -- thanks to occasional hard-panned content.

Haven't we all played early Beatles 'stereo' songs? Good luck getting most speakers to 'disappear' during those.

 

[MattHooper]

I am mystified by anyone who expects forward firing loudspeakers to 'disappear' all the time.

IME experience speakers 'disappear' in a content-dependent fashion -- even within a track -- thanks to occasional hard-panned content.

Haven't we all played early Beatles 'stereo' songs? Good luck getting most speakers to 'disappear' during those.

I get what you mean - obviously certain placement of objects in the mix sound field are going to be more distant and attached from the speakers than others.

And you seem to be suggesting that tracks with hard panning are unlikely to aid the impression of any loudspeaker “ disappearing.”

I would disagree.

In my experience tracks that include hard panning tell you even MORE about how well a speaker seems to disappear as the apparent sound source.

Virtually any loudspeaker including those with obvious resonances can float images free of the speaker when those images have been panned away from the speakers themselves.
Any crappy speaker can float a centralized vocal between the speakers.

It’s really when sonic images are panned closer to the loudspeakers that good speakers separate themselves all the more in STILL seeming to disappear.

In my experience, and many others have observed this, when listening to tracks on speakers that have resonances it tends to manifest and announce their presence the more a sound is panned to the sides of the sound stage toward the speakers. And the more wide panned instruments will start to pull toward and glom into the speakers.

So for instance: you have a mix with a sound stage in which instruments off to the far right and left but with a sense of distance - like an orchestral, recording with first violins, and maybe harp and piano to the left, and maybe double bass, cellos, low brass or whatever to the right, but the whole recording has them in the distance, not up close to the speakers.

On a good speaker that disappears, the sound stage will maintain a sense of depth across the entire width, closer to that of a real Orchestra. As if you can stare past the loudspeakers to those instruments occurring in the distance in the corners behind them.

On a poor speaker with resonances, as instruments are placed more and more to the sides, they will tend to pull in more towards the loudspeakers, some gloming into the speakers, so you start to get a bit more of a consistent U-shape to sound stages that otherwise would have been more consistently expansive and deep.

The same principle applies to pretty much any mix whether the instruments are mixed to sound closer and upfront to the sides or not. In fact, the more hard panned and close instruments are in a mix, the more telling they can be in terms of a speaker disappearing or not. It’s the hardest test.

That’s why when I audition loudspeakers I have a number of tracks that have hard pan and double panned instruments, some right upfront to the speakers others distant, and the best speakers give a better sense of disappearing on even the most difficult tests.
The hard panned instruments or voices may be coming from the direction of the left and right speaker, but they still seem to be floating free of that speaker.

I’m surprised that you have not noticed this yourself or not had this experience in comparing how different loudspeakers perform with hard panned sounds.

 

[krabapple]

As a wise man once said: yadda yadda yadda

 

[LevityProject]

But, as described in some detail in the 4th edition of my book, stereo itself is a problem. It is not an encode/decode process. It is merely a two-channel delivery system, and what is delivered is what someone in a recording studio created. There are no standards. Stereo recordings that employ hard-panned left and right images/instruments are delivering essentially mono signals that will be strongly localized to the loudspeaker - it is the source. These are combined with phantom (double mono) images that have very different timbres and spatial characteristics. Other recordings avoid extreme hard pans, and these present a spatially "softer" imaging. When images appear outside the left-right span of the loudspeakers it is very likely accidental, the result of binaural processing In the mix, or you have multidirectional loudspeakers energizing strong side-wall reflections. Some classical recordings present a wide soundstage because of the high proportion of reflected (decorrelated) sounds in the mix; others have been known to deliver spatially compressed, in my view, simply bad recordings. There are no standards.

I hope this summarizes well:

• For the purpose of knowing if a speaker has neutral tonality, mono testing removes as many variables as possible and allows the tester to control for the variables that remain.
• Speakers at their core only translate what already exists in an analogue or digital signal into something we can hear and experience. Encoded stereo music played back over two loudspeakers (a "stereo set-up") produces so many acoustic and psychoacoustic variables that it's unreasonable to use this to make a final judgement on the tonality of a speaker.