Why evaluating the sound of a single speaker is essential

[aarons915]

This topic always seems to be a heated debate but I'm curious how many people have attempted their own test and have had speakers be preferred in mono but not stereo or vice versa? Years ago I was comparing 2 sets of speakers the way I always did, in stereo and switching between the 2 sets and they were both good speakers so it was extremely hard to choose which ones I preferred. This was early on in my readings of Dr. Toole's work and I decided to try to emulate the Harman style of comparison as close as I could, 1 speaker of each in mono, level-matched, able to instantly switch back and forth and I did the comparison blindfolded. I realize that doing the comparison blind also likely had an effect on my preferences but what I found is what was a stalemate was now an obvious preference for 1 speaker, every single song I played and within 10 seconds of listening, the differences were obvious and repeatable.

People are arguing that a speaker could win in mono but not in stereo are basically arguing that there is some sort of interaction between 2 speakers that doesn't exist in mono and I'm not sure what that could be. To me stereo just adds interference and gets busier which is why it makes it more difficult to spot differences.

 

[Audionaut]

basically arguing that there is some sort of interaction between 2 speakers that doesn't exist in mono and I'm not sure what that could be

With good recordings of a symphony, for example, you get a lot of information about the room in stereo that you don't have in mono.
You can assign the instruments to their location, which is impossible in mono

.

 

[aarons915]

With good recordings of a symphony, for example, you get a lot of information about the room in stereo that you don't have in mono.
You can assign the instruments to their location, which is impossible in mono

.View attachment 454387

Yes so stereo vs mono is much different in a recording, we're talking about why a speaker would sound different in mono vs stereo. Why would having 1, 2 or 5 speakers change how an individual speaker sounds?

 

[MattHooper]

This topic always seems to be a heated debate but I'm curious how many people have attempted their own test and have had speakers be preferred in mono but not stereo or vice versa?

I haven’t done any formal test, but long ago, I noticed that hard panned instruments and voices - essentially mono - were quite revealing of a speakers tonal characteristics.

Just as Doc Toole point out - there’s enough panning to be found even in many stereo record recordings, that can reveal inadequacies.

 

[Arindal]

People are arguing that a speaker could win in mono but not in stereo are basically arguing that there is some sort of interaction between 2 speakers that doesn't exist in mono and I'm not sure what that could be.

There is a bunch of potential aspects fitting your description, just a few examples:

- phantom localization stability
- phantom localization width
- phantom source tonality
- perceived depth-of-field
- perceived proximity of dry sound sources (or such with decor related reverb)
- perceived room reverb from the recording / ambience
- transparency of very complex sound mixtures, such as a 16+-voice choir or an orchestra piece containing lots of counterpoints

I'm curious how many people have attempted their own test and have had speakers be preferred in mono but not stereo or vice versa?

I have done that, even in a controlled room and with a group of experienced listeners. There are aspects which you simply cannot judge in mono, but more revealingly there are others flipping the test result. It was particularly those having to do with proximity. If in mono a speaker is preferred which with dry sources gives the impression of reduced proximity hence slightly more natural depth-of-field, you can be sure this one will be dismissed in stereo for the lack of proximity, overly distant imaging, exaggerated depth-of-field, dominating ambience or diffuse imaging.

To me stereo just adds interference and gets busier which is why it makes it more difficult to spot differences.

That is certainly true if you execute anything close to a discrimination test. For spotting imperfections like audible distortion, narrow banded FR issues, compression, judging the bass quality or alike, mono is the way to go IMHO.

 

[Jaxx1138]

Where have I written that loudspeakers are defying the laws of acoustics in order to achieve phantom imaging?

You know, it is pretty cheap to call others ´flat-earthers´. It is seemingly more difficult to understand how a loudspeaker interacts with the room and how the resulting soundfield comprised of direct sound, reverb contained in the direct sound plus indirect reflections, is creating a phantom image with two-channel stereo recordings. I do not see a single solid evidence why phantom localization, ambience and depth-of-field of stereo recordings could be judged easily in mono with a downmixed or channel-mapped recording.

My apologies. That was not my intention. Just that understanding acoustics can assist in why and how imaging is a function of the recording. The speaker only reproduces the signal. The better it does that , the better it sounds. The overall directivity will contribute to how wide the soundstage appears to be. But most of that can be controlled through speaker positioning and possibly some room treatment. The rest is all in the hands of the mixing and mastering people.

 

[Jaxx1138]

With good recordings of a symphony, for example, you get a lot of information about the room in stereo that you don't have in mono.
You can assign the instruments to their location, which is impossible in mono

.View attachment 454387

As what you have said is indeed true, I believe the primary discussion was concerning conducting a subjective evaluation of a loudspeakers overall fidelity. This is much easier for humans with a mono speaker (at least the research seems to indicate this statistically). So the application is a subjective review by a listener of a loudspeaker. This is separate from our binaural hearing (humans perception) and listening to a stereo reproduction of music. This will always be a more pleasurable experience. But in that case you are listening to the music /recording. The other is only for evaluating the loudspeakers performance in a subjective manner.

 

[Arindal]

The speaker only reproduces the signal. The better it does that , the better it sounds.

That is trivial, but does not say anything. A signal is just digital or analogue. It is not a three-dimensional soundfield. Nothing is defined about how the original signal should sound like and how ´better´ is defined in terms of localization, proximity, ambience, imaging, transparency, phantom source tonality and alike. Most of these aspects cannot even be evaluated as ´better´ or ´worse´.

There is some understanding that listening conditions in a studio should be considered more of an ´original´ representing the intension of mixing engineer and artists, but that is also not a clear definition as listening conditions in studios also vary. There is not standardization although the differences might be much smaller compared to home listening. Getting closer to what most of studios do, is actually a good idea, but not definable by precise criteria.

The overall directivity will contribute to how wide the soundstage appears to be.

Soundstage width stands in contradiction to proximity, depth-of-field and localization width/stability. Every loudspeaker reproduction system will have to present a balanced compromise between these.

But most of that can be controlled through speaker positioning and possibly some room treatment. The rest is all in the hands of the mixing and mastering people.

Actually, it is not. How a mix translates to an actual soundfield perception in terms of imaging, is mostly depending on the loudspeakers themselves, if there are not terrible mistakes at play regarding matching of loudspeakers and positioning.

conducting a subjective evaluation of a loudspeakers overall fidelity. This is much easier for humans with a mono speaker (at least the research seems to indicate this statistically).

As mentioned, I dispute that and tried to elaborate why. There is common understanding that mono listening tests are superior for discrimination tests, like for detecting audible distortion. But this in my understanding is rather important for loudspeaker development, particularly early stages thereof (like when choosing drivers, designing waveguides and baffle geometry).

Apologies accepted, I did not take it personally. The topic of stereo imaging is just way more complex than you might think, and requires in my understanding decades of listening experience with own recordings and setting up both microphone arrangements in a concert hall and speakers in a room to get a vague understanding how the results are correlating.

 

[aarons915]

There is a bunch of potential aspects fitting your description, just a few examples:

- phantom localization stability
- phantom localization width
- phantom source tonality
- perceived depth-of-field
- perceived proximity of dry sound sources (or such with decor related reverb)
- perceived room reverb from the recording / ambience
- transparency of very complex sound mixtures, such as a 16+-voice choir or an orchestra piece containing lots of counterpoints



I have done that, even in a controlled room and with a group of experienced listeners. There are aspects which you simply cannot judge in mono, but more revealingly there are others flipping the test result. It was particularly those having to do with proximity. If in mono a speaker is preferred which with dry sources gives the impression of reduced proximity hence slightly more natural depth-of-field, you can be sure this one will be dismissed in stereo for the lack of proximity, overly distant imaging, exaggerated depth-of-field, dominating ambience or diffuse imaging.



That is certainly true if you execute anything close to a discrimination test. For spotting imperfections like audible distortion, narrow banded FR issues, compression, judging the bass quality or alike, mono is the way to go IMHO.

But according to Dr. Toole the speaker that wins in mono "always" wins in stereo so until someone shows this isn't always the case I'm not seeing the point of hashing out theories about why it might not be the case. Most of your examples seem related to dispersion anyway, it seems obvious to me that the reason why mono is easier to discern differences is simply because the dispersion differences are much more pronounced when you only have 1 speaker responsible for all of the reflections, once 2 or more are in the system they excite these reflections much better but it makes everything "busier" and is now harder to spot sound quality differences. The devil's advocate would say if differences are hard to spot with 2 or more speakers why bother buying great speakers in the first place.

 

[MarkS]

But according to Dr. Toole the speaker that wins in mono "always" wins in stereo

Data given in Toole's own book contradicts this statement.

Note that the average "spatial quality" rating of speaker E is higher than that of speaker AA in stereo, lower in mono:

 

[Arindal]

But according to Dr. Toole the speaker that wins in mono "always" wins in stereo

With all due respect, but I do not see a way to do a proper listening test in mono to back up this claim. Criteria of sound quality heavily relying on phantom localization, as the ones I have listed, cannot be judged in complete absence of this underlying psychoacoustical principle. In mono, there is no phantom localization for obvious reasons.

the reason why mono is easier to discern differences is simply because the dispersion differences are much more pronounced when you only have 1 speaker responsible for all of the reflections, once 2 or more are in the system they excite these reflections much better but it makes everything "busier" and is now harder to spot sound quality differences

You are describing an experiment of a discrimination test, and I surely agree for this part of a loudspeaker test, when it comes to identifying flaws like distortion or audible resonances.

I was referring to preference tests, particularly when it comes to judging the spatial qualities and describing which imaging listeners would prefer. This is not only heavily dependent on the listeners in the experiment being familiar with the recordings, it is also a test if the soundfield in the listening room produces a coherent picture of the recording. For example localization stability is heavily relying on the existence of several sound event including reflections, and in the end of the day a product of how well our brain perceives this complex mixture as one coherent picture. This includes the wanted and unwanted reflections, and if you take away half of them, like in mono, you cannot judge anything.

 

[MattHooper]

I’m not sure how the following could be evaluated in mono:

One of the qualities that I’ve always cared about and looked for in loud loudspeakers was a sense of sonic density, palpability, focus, and solidity to the Sonic images.

Stereo images by nature have a sort of phasey, ghostly see-through quality relative to a real sound source or relative even to the sound coming directly from a single speaker.

So to me the more dense, solid and corporal the sonic images with an a stereo field the more convincing and satisfying I find the experience.

That’s one of the reasons why, having auditioned a huge number of speakers and owned many as well, I have kept return returning to Thiel speakers, which in comparison with everything else I’ve had, presents and extremely well organized sonic picture. It’s like all the sonic information related to any particular sound - a saxophone, voice, drumstick on a drum rim - it all seems to be lined up and coming from the same place, making for a very solid palpable, precise images. After listening to some other speakers, switching to the Thiels can feel a bit like when you are at the optometrist being tested for a new eyeglass prescription, and the letters on the eye test snap into more focus with the proper prescription.

Now I’m quite familiar with identifying tonal and timbral qualities in mono. And even some spatial qualities - for instance, a speaker with low resonances will allow the sound even in mono to seem to float more free of the cabinet.

But I’m not sure how listening to a single speaker would tell me about the specific type of qualities I’ve described in terms of stereo imaging performance.

Any thoughts on this?

 

[Hayabusa]

I’m not sure how the following could be evaluated in mono:

One of the qualities that I’ve always cared about and looked for in loud loudspeakers was a sense of sonic density, palpability, focus, and solidity to the Sonic images.

Stereo images by nature have a sort of phase ghostly see-through quality relative to a real sound source or relative even to the sound coming directly from a single speaker.

So to me the more dense, solid and corporal the sonic images with an a stereo field the more convincing and satisfying I find the experience.

That’s one of the reasons why, having auditioned a huge number of speakers and owned many as well, I have kept return returning to Thiel speakers, which in comparison with everything else I’ve had, presents and extremely well organized sonic picture. It’s like all the sonic information related to any particular sound - a saxophone, voice, drumstick on a drum rim, it all seems to be lined up and coming from the same place, making for a very solid palpable, precise images. After listening to some other speakers, switching to the teals can feel a bit like when you are at the optometrist being tested for a new eyeglass prescription, and the letters on the eye test snap into more focus with the proper prescription.

Now I’m quite familiar with identifying tonal and timbral qualities in mono. And even some spatial qualities - for instance, a speaker with low resonances will allow the sound even in mono to seem to float more free of the cabinet.

But I’m not sure how listening to a single speaker would tell me about the specific type of qualities I’ve described in terms of stereo imaging performance.

Any thoughts on this?

Indeed, let's just test both : mono and stereo...

Now I am here: I always wondered about the 'obvious' question here:

How to practically listen in mono: listen to one channel or both channels summed?

@amirm @Floyd Toole

 

[Arindal]

How to practically listen in mono: listen to one channel or both channels summed?

From mixing and mastering point of view, both variants derived from a stereo recording are flawed and lead to misjudgments when it comes to tonal balance and other criteria of sound quality. One might think of the way a central voice would be panned and mixed for stereo and how this translates to tonal imbalances the moment you do a downmix or ditch one channel.

There is a second implication to this topic which cannot be solved: Listen to mono derived from stereo recordings at 0deg frontal or the sound from one speaker coming in at 30deg horizontally? Both variants produce different errors and might lead to different results what is preferred in terms of tonality.

In my understanding, there is only one way of doing a proper monaural listening test not prone to misjudging: Listening to a dedicated monaural mix on one speaker at 0deg.

 

[Jaxx1138]

That is trivial, but does not say anything. A signal is just digital or analogue. It is not a three-dimensional soundfield. Nothing is defined about how the original signal should sound like and how ´better´ is defined in terms of localization, proximity, ambience, imaging, transparency, phantom source tonality and alike. Most of these aspects cannot even be evaluated as ´better´ or ´worse´.

There is some understanding that listening conditions in a studio should be considered more of an ´original´ representing the intension of mixing engineer and artists, but that is also not a clear definition as listening conditions in studios also vary. There is not standardization although the differences might be much smaller compared to home listening. Getting closer to what most of studios do, is actually a good idea, but not definable by precise criteria.



Soundstage width stands in contradiction to proximity, depth-of-field and localization width/stability. Every loudspeaker reproduction system will have to present a balanced compromise between these.



Actually, it is not. How a mix translates to an actual soundfield perception in terms of imaging, is mostly depending on the loudspeakers themselves, if there are not terrible mistakes at play regarding matching of loudspeakers and positioning.



As mentioned, I dispute that and tried to elaborate why. There is common understanding that mono listening tests are superior for discrimination tests, like for detecting audible distortion. But this in my understanding is rather important for loudspeaker development, particularly early stages thereof (like when choosing drivers, designing waveguides and baffle geometry).

Apologies accepted, I did not take it personally. The topic of stereo imaging is just way more complex than you might think, and requires in my understanding decades of listening experience with own recordings and setting up both microphone arrangements in a concert hall and speakers in a room to get a vague understanding how the results are correlating.

Psychoacoustics are creations of perception and interfering waveforms, combining in an acoustic environment resulting in a convolution. Spaciousness is not a function of an individual loudspeaker, it is the combination of direct energy combined with reflections and potentially other loudspeakers. That is the physics of acoustics. The sense of space and ambience is the delayed secondary wave fronts arriving at your ear canals from reflections or other loudspeakers that creates this sense of space. Individual loudspeakers cannot do this without at the very least a reverberant acoustic space and or other independent sound sources. I am an acoustic engineer by trade.

 

[MarkS]

How to practically listen in mono: listen to one channel or both channels summed?

Harman data shows that the best signal for judging speaker quality in mono is pink noise.

In stereo, I like out-of-phase pink noise as a test signal. (That is, a mono signal, but with reversed polarity on one channel.) I want to hear it seeming to come from far outside the speaker locations.

 

[Hayabusa]

Harman data shows that the best signal for judging speaker quality in mono is pink noise.

In stereo, I like out-of-phase pink noise as a test signal. (That is, a mono signal, but with reversed polarity on one channel.) I want to hear it seeming to come from far outside the speaker locations.

Good one.

But Amir clearly listens to songs when he does its evaluation..
How do you make the mono signal Amir?

 

[Arindal]

The sense of space and ambience is the delayed secondary wave fronts arriving at your ear canals from reflections or other loudspeakers that creates this sense of space. Individual loudspeakers cannot do this without at the very least a reverberant acoustic space and or other independent sound sources.

That was my point: You do not have such mechanism at play when listening to a mono loudspeaker, so you cannot judge these criteria. With stereo recordings, a huge fraction of such properties relies on phantom localization and ambience, so you need a second loudspeaker to get the feeling.

Btw it is not only the second wavefront coming in to define the sense of space, it is basically a series of reflections over time, tonality and different angles, which our brain can prescribe to the initial sound event, the first wavefront.

Harman data shows that the best signal for judging speaker quality in mono is pink noise.

I wonder how one could judge depth-of-field, proximity, ambience, localization stability and other factors like reverb tonality solely with pink noise. AFAIK our brain has no mechanism to clearly distinguish direct from reflected pink noise, and the spectral composition of noise usually does not change suddenly over time (as it is the case with speech or singing).

Maybe this ´best signal for judging´ is related to discrimination tests as well? Would make sense.

 

[Jaxx1138]

That was my point: You do not have such mechanism at play when listening to a mono loudspeaker, so you cannot judge these criteria. With stereo recordings, a huge fraction of such properties relies on phantom localization and ambience, so you need a second loudspeaker to get the feeling.

Btw it is not only the second wavefront coming in to define the sense of space, it is basically a series of reflections over time, tonality and different angles, which our brain can prescribe to the initial sound event, the first wavefront.



I wonder how one could judge depth-of-field, proximity, ambience, localization stability and other factors like reverb tonality solely with pink noise. AFAIK our brain has no mechanism to clearly distinguish direct from reflected pink noise, and the spectral composition of noise usually does not change suddenly over time (as it is the case with speech or singing).

Maybe this ´best signal for judging´ is related to discrimination tests as well? Would make sense.

The problem with your evaluation is that you are conflating two separate phenomena. Evaluating the overall linearity and fidelity of a loudspeaker is not part of the ambience or spaciousness of the unit since spaciousness and ambience are not part of an individual loudspeaker. The subjective test done at Harman were primarily about timbre, tonal balance and frequency linearity. The spaciousness effect requires things separate from the individual loudspeaker. Such as a secondary loudspeaker with a stereo recording containing separate discrete signals with both frequency and phase differences in the signal. This is a psychoacoustic effect. It is created through the recording process. If the signal were to be mono played through two separate speakers equidistant from the listening position, there would be no spatial effect creating sounds between the speakers or in front or behind them. Rather than trust me , you can use one of the AI algorithms and have it reference Toole, Everest, Beranek and Kinsler.

 

[Arindal]

Evaluating the overall linearity and fidelity of a loudspeaker is not part of the ambience or spaciousness of the unit since spaciousness and ambience are not part of an individual loudspeaker.

Indirectly they are, as qualities of a speaker despite from just reproducing a tonally balanced and undistorted signal, greatly influence the ambience, localization and imaging aspect. Think of its following properties:

- pair tolerances (influence localization width and stability)
- dimensions and geometry of localization-relevant sound sources (localization stability and how a phantom source would be perceived within the ambience). This is particularly an issue under nearfield conditions or in a setup with pronounced early reflections
- evenness of frequency response over the potential listening window with the listener slightly moving or turning his/her head (again a typical nearfield issue)
- edge diffraction effects causing additional mirrored phantom localization or blurring the existing ones
- directivity index over frequency (greatly affects depth-of-field, perception of reverb tonality and direction, proximity and phantom source integration)

If the signal were to be mono played through two separate speakers equidistant from the listening position, there would be no spatial effect creating sounds between the speakers or in front or behind them.

Under anechoic conditions, certainly not. Nevertheless this test would be useful to separately judge localization stability, particularly when executed with a complex signal such as speech.

The subjective test done at Harman were primarily about timbre, tonal balance and frequency linearity.

While such criteria are judgeable in mono, this requires a designated mono mix standardized for the intended angle at which the speaker is placed, to come to a valid conclusion. I would not trust any verdict on tonal balance derived from a downmixed or channel-mapped stereo recording as this introduces a potential root for serious misjudgment.