Speaker Testing: why mono is better

[TheTalbotHound]

Given the isolation between the cups, the interactions we worry about with stereo speaker testing is not there. Indeed, I find that tonality tests are much easier with headphones than speakers.

As I explained when this question was asked earlier , at some point I do like to test in mono to see if more insight can be had. For now, the system is working so I keep doing it.

With headphones, it is mandatory in my book to provide subjective commentary on spatial qualities so I have to listen stereo anyway.

Maybe with a few open backs that leak a lot might make it from one side of the head to the other. Doubt that will have much of an effect but that is yet to be seen.

 

[amirm]

Maybe with a few open backs that leak a lot might make it from one side of the head to the other. Doubt that will have much of an effect but that is yet to be seen.

I tested for that for a different project. I even put a box around both cups but could not cause appreciable impact.

 

[MerlinGS]

I disagree. Science shows that mono testing is better for comparing speakers objectively. I've hosted both mono and stereo blind tests and my personal experiences have been inline with this. Mono makes it much clearer and easier to decide which speaker you prefer.

I think @Inner Space's position on this matter is being misrepresented. To my understanding, he has acknowledged numeours times that mono is sufficient to determine the sonic qualities of a speaker. His contention is that if only one speaker is tested, there is no way to determine if both speakers share the same qualities; i.e. his argument is about quality control. There is no doubt that if two speakers differ sufficiently, their stereo performance will falter. However, Amirm only chooses to test 1 speaker, thus leading to @Inner Space's conclusion, that the only reason two speakers are not tested is "cost, time, and inconvenience, not about engineering".

@Inner Space's conclusion is both right and wrong, but not for the reason's assumed in the above noted post. Surely, we can all agree that it is important to ensure speaker manufacturers' QC don't lack the capabilities to run a proper production line; i.e. with sufficient QC to ensure all the speakers produce are of sufficient quality (in this context quality just means similar specs). However, @Inner Space's conclusion makes an assumption that appears invalid; i.e. Amirm considers part of his role to determine the QC of manufacturers. From Amirm's numerous posts it would seem his concerns relating to measurements (as they pertain to speakers and headphones) are generally focused on examing their relationship with perceived preference as established by the Harman studies, and to his own listening experience/preference.

@richard12511 , if I'm wrong about @Inner Space's argument, please accept my apoligies.

 

[amirm]

I think @Inner Space's position on this matter is being misrepresented. To my understanding, he has acknowledged numeours times that mono is sufficient to determine the sonic qualities of a speaker. His contention is that if only one speaker is tested, there is no way to determine if both speakers share the same qualities; i.e. his argument is about quality control. There is no doubt that if two speakers differ sufficiently, their stereo performance will falter. However, Amirm only chooses to test 1 speaker, thus leading to @Inner Space's conclusion, that the only reason two speakers are not tested is "cost, time, and inconvenience, not about engineering".

If there is variability, there is no way anything can be done about it. So what if I test two? How is that representative of the two someone else buys? The only solution would be buy dozens of speakers and testing and surely no reasonable person would ever suggest that. Nor that our data means nothing because there is variability between speakers. Such variability will be less noticed in stereo anyway.

 

[Inner Space]

Nor that our data means nothing because there is variability between speakers.

For the record, I think your data is meticulously done and extremely valuable. I'm delighted to have it, as I'm sure are many others.

The only solution would be buy dozens of speakers and testing and surely no reasonable person would ever suggest that.

Purely anecdotal and not germane, but - I'm retired now, but 3 years ago was hired to do the stereo mix for a streaming series made in Hollywood. I said I would only do it in my own suite at home. Their counter was, OK, but you have to use JBL 4367s, to maintain the "house sound" (yeah, rooms, variables, etc, etc, but these are producers, not brain surgeons). We ordered 12 units from Harman and had them drop-shipped to a studio I know for testing. I wanted a pair that matched within 0.5dB for sensitivity, FR and distortion.

Only 2 of the 12 made it. Errors from one extreme to the other reached 4dB. One was 2dB less sensitive than the other 11. Ironically the "good" pair shipped onward to me were tipped up hot above the crossover, a clear error, but at least two identical errors. I EQed them down by draping a cloth over the waveguide (what I call TSP - textile signal processing, a method I have seen and used all over - whoever you are, your favorite album was quite possibly mastered through a towel) and I thought they sounded great. The images they created were excellent. I never sent them back - I still have them, in my regular listening room, still sounding great. My version of Hollywood accounting.

 

[Thomas_A]

My point of view, as a few others also have pointed to, is that mono listening IS the most sensitive way to evaluate frequency response deviations. As Amir and Toole among others points out. However, going from mono to stereo impose some differences in timbre that are audible. Thus, I don't believe that the "perfect" mono speaker is the perfect stereo speaker, and certainly not for all. The compensation needed may however be small enough to grade the speaker as "perfect" (e.g. within +/- 1 dB or so), so the question may be academic. Toole also admits that stereo is flawed in that way, but instead to deal with the "insoluble" problem, he propose to use multichannel. It is then a matter of the algorithm to decode stereo recordings to three-channel, including the comb filtering effect that may or may not have occurred during the recording process.

 

[sergeauckland]

If there is variability, there is no way anything can be done about it. So what if I test two? How is that representative of the two someone else buys? The only solution would be buy dozens of speakers and testing and surely no reasonable person would ever suggest that. Nor that our data means nothing because there is variability between speakers. Such variability will be less noticed in stereo anyway.

Whilst I appreciate and understand the reasons for only testing one example of each loudspeaker, is there any evidence for the statement highlighted? Precise Stereo imaging demands well matched loudspeakers, so I fail to see how this statement can be true.

S.

 

[MerlinGS]

If there is variability, there is no way anything can be done about it. So what if I test two? How is that representative of the two someone else buys?

Agreed

Such variability will be less noticed in stereo anyway.

Just for clarity. @Inner Space was not requesting that speakers be tested by playing them in stereo, but rather that each individual speaker be tested to determine variability. But as you noted, and as @Inner Space's subsequent post illustrated (where he had 12 JBL 4367 speakers tested for variability and only 2 were closely matched), this could be a difficult endeavour, and I assume its purpose would be incongruent with this website/forum, or your own interests.

 

[amirm]

Whilst I appreciate and understand the reasons for only testing one example of each loudspeaker, is there any evidence for the statement highlighted? Precise Stereo imaging demands well matched loudspeakers, so I fail to see how this statement can be true.

S.

Only level and delay are important there, not slight tonality differences. On the latter, that is what the research is about which I have referenced.

 

[amirm]

Just for clarity. @Inner Space was not requesting that speakers be tested by playing them in stereo, but rather that each individual speaker be tested to determine variability.

I don't know how that has anything to do with the research presented, or the work I have been doing in this forum. You all can do such testing yourself anyway.

 

[Duke]

@amirm, I very much enjoyed your conversation with @joentell a couple of days ago. I agree with you almost all the time.

One of the things I noticed was that you said tonality "is the most important thing a speaker does. Everything else is secondary." And I agree with your proirity there! As Floyd Toole has said, "Frequency response is the most important parameter in any audio component. If it is wrong, nothing else matters."

My understanding is that you claim mono testing is better than stereo testing, and I agree as far as sound quality goes. But IF that is a blanket claim with no qualifiers, it implies mono listening (and/or measurements) to be better than stereo listening for evaluating stereo spatial quality. Have I misunderstood you? Anyway, THAT is what I'm arguing about.

Let me start by claiming that, once we have good sound quality (which includes tonality), spatial quality matters a lot. Quoting from page 183 of the third edition of Floyd Toole's book:

"Sound and spatial qualities both contribute to our musical pleasure, but to what extent? Here they are shown to be of comparable importance." [emphasis mine]

And on page 186:

"Thus sensations of sound quality and spaciousness contribute equally to impressions of "naturalness", and spaciousness dominated the impression of "pleasantness". [emphasis mine]

My intention with these citations is to make a case for the importance of spatial quality - BOTH sound quality and spatial quality matter a lot. (And spatial quality apparently matters more to some than to others... for instance about his current listening room Toole reports that “stereo reproduction is very satisfying, but I still employ tasteful upmixing for many recordings to embellish the sense of space.”)

When I asked whether you could predict spatial quality in stereo from your measurements and/or mono listening, you replied, "I cannot."

To recap: Spatial qualtity in stereo apparently matters a lot, but is not something you can predict from your measurements or from your mono listening tests. Is that correct?

For the sake of clarity, and at the risk of annoyingly repeating myself, regardless of your thoughts about any of the preceding, could you please reply to this question:

In your opinion, is mono testing (listening or measuring or both) a more accurate predictor of stereo spatial quality than stereo listening is?

 

[Blumlein 88]

@amirmsnip....
In your opinion, is mono testing (listening or measuring or both) a more accurate predictor of stereo spatial quality than stereo listening is?

How is stereo listening a predictor of stereo spatial quality? Seems to me once you are listening you are evaluating the stereo spatial quality you are experiencing. I think this is why Amir can't answer your question the way you want it answered.

 

[Duke]

How is stereo listening a predictor of stereo spatial quality? Seems to me once you are listening you are evaluating the stereo spatial quality you are experiencing. I think this is why Amir can't answer your question the way you want it answered.

Well maybe "predictor" was the wrong word. How about "indicator."

The strange wording of my question arises as a consequence of the claim that "mono tests are better than stereo tests". If that is indeed the claim, with no limitations or qualifiers, it implies that mono tests of spatial quality are better than stereo tests of spatial quality. (I am not disputing the claim that mono tests are better than stereo tests as predictors or indicators of sound quality in stereo. Note that as used by Floyd Toole in this context, "sound quality" and "spatial quality" are two different categories, neither being a subset of the other.)

And since we listen to music in stereo, spatial quality in stereo is what really matters. (Spatial quality in mono matters only for its ability to predict or indicate spatial quality in stereo.)

Therefore, IF "spatial quality" really only matters in stereo, and IF mono tests are better at predicting (or indicating) it, THEN mono tests are better than stereo tests for predicting (or indicating) stereo spatial quality.

IF that sounds illogical to you, well that makes two of us.

And if it IS illogical, then the claim that "mono tests are better than stereo tests" may not apply to spatial quality.

If you see a logical error in my thinking, let me know. It's possible.

 

[Blumlein 88]

http://personal.ee.surrey.ac.uk/Personal/P.Jackson/pub/jaes14/ConettaEtAl_JAES14a_Preprint.pdf

I've not finished reading this yet. It may be more useful for the sources in the article. A few of them refer to examination of predictors for spatial audio quality and attempt to create an artificial listener to predict spatial attributes. Most of this article is oriented toward multi-channel or the spatial results of mix downs of mch to two or three channels, but some of the source articles appear to deal more with stereo. Particularly source 20,21, and 22.

 

[Duke]

Amir, I THINK I FOUND MY MISTAKE!!!

I had not listened carefully to what Floyd Toole was saying in that lecture he gave at McGill University in Montreal. I now think he was very precise with his choice of words and I MISSED IT. And I had carried my misunderstanding forward into this thread.

Here is what he said:

"So we do the test in mono. And we occasionally have done comparisons in stereo as well just to prove the point to, ah, questioning, ah, persons, and every time the ones that win the mono tests win the stereo tests. There has been no exception to that rule. There is nothing special about stereophony or multichannel insofar as sound quality is concerned." [emphasis in Toole's voice]

The thing I was missing was his use of the term "sound quality". I think Floyd Toole uses that term very specifically and consistently, and it I think it does not include the characteristics which comprise "spatial quality." In other words, I was unwittingly putting words into his mouth (implicitly if not explicitly), and likewise proceeded to put words into Amir's mouth. My apologies, @amirm and @Floyd Toole!

Here is that video clip, hopefully cued up to 9:25:

 

[Blumlein 88]

https://www.audiosciencereview.com/...-spatial-quality-in-audio-reproduction.21946/

Started a spatiality thread so this one can stay about mono testing.

It looks like there are test signals which can be used to assess spatial quality. And directional mapping which means you need only map one speaker's result and apply to two or more channels to have some predictive ability.

 

[tuga]

Amir, I THINK I FOUND MY MISTAKE!!!

I had not listened carefully to what Floyd Toole was saying in that lecture he gave at McGill University in Montreal. I now think he was very precise with his choice of words and I MISSED IT. And I had carried my misunderstanding forward into this thread.

Here is what he said:

"So we do the test in mono. And we occasionally have done comparisons in stereo as well just to prove the point to, ah, questioning, ah, persons, and every time the ones that win the mono tests win the stereo tests. There has been no exception to that rule. There is nothing special about stereophony or multichannel insofar as sound quality is concerned." [emphasis in Toole's voice]

The thing I was missing was his use of the term "sound quality". I think Floyd Toole uses that term very specifically and consistently, and it I think it does not include the characteristics which comprise "spatial quality." In other words, I was unwittingly putting words into his mouth (implicitly if not explicitly), and likewise proceeded to put words into Amir's mouth. My apologies, @amirm and @Floyd Toole!

Here is that video clip, hopefully cued up to 9:25:

Well spotted.

Now we need a quote about single-speaker testing and spatial quality.

 

[JustIntonation]

This topic of "why Amir tests speakers in mono" keeps coming up. I must have explained that a hundred times in text. Most of you probably know why. But I thought I do a video that covers the research and explains it all. Here it is:

References:
https://www.aes.org/e-lib/online/browse.cfm?elib=11740
Subjective Measurements of Loudspeakers: A Comparison of Stereo and Mono Listening

https://www.aes.org/e-lib/browse.cfm?elib=14622
Comparison of Loudspeaker-Room Equalization Preference for Multichannel, Stereo, and Mono Reproductions: Are Listeners More Discriminating in Mono?

https://www.aes.org/e-lib/browse.cfm?elib=8338
A New Laboratory for Evaluating Multichannel Audio Components and Systems

It is true that mono testing will allow you to hear things that are less easy to hear in stereo testing.
But.. tonality IS CHANGED to a large degree!
I think it's important to be aware of this.
And it is something that is easy to test. Simply hold a measurement mic near one's left ear and then one's right ear when listening to a single mono speaker. Then do the same when listening only to the right speaker of a stereo setup but with your head facing the virtual center, and then do the same when listening to both speakers with your head again facing the virtual center.
If you do the above and compare the recordings you'll notice first of all the difference in the HRTF when facing the speaker and not facing the speaker directly. And further for the stereo measurements you'll notice that up till a certain frequency (from memory somewhere between 1-2kHz, was some time ago I did this test) the two speakers combine constructively and above this they combine in a comb filter way (and this also depends on the spread / angle of the speakers).
This is how we'll always hear stereo speakers. And it changes tonality.
So in my opinion you cannot judge tonality when listening in mono correctly. Yes you can hear errors more easily in mono including fr deviations, but a speaker will have a tendency to sound bright (too much upper mids/lower treble etc) in mono much more quickly than in stereo (simply because music has been mixed in stereo / recordings judged and mastered in stereo).

One more sidenote. It does matter if one combines the left and right tracks of stereo music to make a mono track, or if one simply takes only the left or right audio track. The is because the bass is usually center and will combine constructively while a larger proportion of higher sound are panned left or right or even have left-right out of phase and either combine randomly or even destructively.
This is again something that'll change tonality for a lot of music relative to the stereo track.

 

[amirm]

This is how we'll always hear stereo speakers. And it changes tonality.

It does so more the reason to test in mono. The variability in stereo is content dependent and that is not what we are interested in.

One more sidenote. It does matter if one combines the left and right tracks of stereo music to make a mono track, or if one simply takes only the left or right audio track. The is because the bass is usually center and will combine constructively while a larger proportion of higher sound are panned left or right or even have left-right out of phase and either combine randomly or even destructively.
This is again something that'll change tonality for a lot of music relative to the stereo track.

It doesn't matter what the transformation does as long as you use the same content/setup for all speakers.

 

[JustIntonation]

It does so more the reason to test in mono. The variability in stereo is content dependent and that is not what we are interested in.


It doesn't matter what the transformation does as long as you use the same content/setup for all speakers.

But when we use actual music and start talking about subjective qualities of this music through a certain speaker, then it seems to me that listening in mono may well push music slightly to the 'bright' side while this would not be so when listening to that same music through a stereo setup?
That is my logic reasoning at least. I'm not seeing a good counter argument against this, am I missing something?

Edit: to explain my logic further.
Because stereo changes tonality and the music was mixed and mastered in stereo, the mixing and mastering engineers made tonality decisions based on the tonality stereo gives. And while different stereo angle etc setups give slightly different tonality changes, the overall change is an overall reduction in treble due to the two effects I described in my earlier post. So if we then play back this music in mono over a single speaker and then use this music to assess if the resulting sound is overall 'bright' then this may well be because of the music not the speaker it seems to me.
Different would be if we're playing pink noise or some other objective signal to subjectively assess tonality (though I don't see a practical way of doing this, would probably hardly learn anything in practice by actually using pink noise).
And just to be clear, I understand and agree from both theoretical standpoints and personal experience that listening in mono will tell you many things about a speaker that are harder to hear in stereo listening. And even some things about tonality (can't exactly quantify what), but for an overall musical experience and assessing if music is subjectively 'bright' I personally find I must listen in stereo.

Overall this has led me to believe that (in combination with your personal room acoustics) certain fairly flat measuring speakers which have a wide but smooth polar pattern including treble (no waveguide) were likely subjectively rated as 'bright' sounding by you partly because of this.
It might be an idea to for instance take a few good flat measuring speaker you found to sound bright in mono to stereo setups and see if you still find them as bright sounding in stereo? It seems to me the must be some subjective change in perception of those speakers tonality at the least?