Speaker Testing: why mono is better

[MerlinGS]

Thank you for replying, Amir. I can't really argue against the point you make. In your opinion, how much weight should be given to spatial quality in mono versus spatial quality in stereo?

There seems to be an assumption in this question (I apologize if I'm wrong) that has been repeated by others. Despite the evaluation being mono vs stereo, the question is whether the results would have remained the same in a comparison of 2 speakers vs 1 with a mono signal. I have little reason to assume they would differ (i.e. I think 2 speakers playing a mono signal would exhibit the same advantages as when playing a stereo signal), and the clarification matters because some posters have noted the signal can sometimes be mono, as if to suggest the findings of 1 speaker vs 2 still applied with mono signal.

PS, Questioning Toole's conclusions is not tantamount to attacking his work or his findings, but rather raising doubts about any dogmatic perspectives based on those findings. It should be noted that other highly qualified (with PhDs) and well respected individuals have raised many of these questions and the answers have not been fully addressed. In a previous post I mentioned Dr. Geddes, a highly respected individual in audio research and he has raised many of these questions.

PPS Another poster mentioned that the type of recordings could affect preference. He stated that classical recordings [or recordings in a theatre/auditorium] would most likely sound better [all other things being equal] with speakers exhibiting constant directivity and wide dispersion, conversely, he also suggested that studio recording may be better served by speakers exhibiting controlled directivity with narrower dispersion [e.g. speakers using a waveguide to narrow the directivity]. The logic is recordings from an audiorium/theatre will create a sense of space that is more diffused, consistent with the qualities of the wide dispersion speaker, in contrast studio recordings produce spatial characteristics that are less diffused and highlight the placement of musicians within the space. These points were also raised by Dr. Geddes, and he further noted that over 90% of recordings were done in studios. This of course raised a question, how does the vast discrepancy in % between recording types (studio vs venue) affect one's conclusions about speaker design and priorities? I'm not trying to suggest Dr. Geddes is right, I'm just noting assumptions are being made that disregard many important variables that should also be considered.

PPPS Another area the research may be lacking is how the findings can be extrapolated to MCH. In a particular blind test the Revel Salon 2 performed better than the JBL 2 (once again showing speakers with wide dispersion may be subjectively preferable than better measuring speakers with narrower directivity), but the question not answered was which speaker would perform better in MCH in a properly treated room; i.e. could one extrapolate from the blind test that the Salon 2s would be preferred. I think a reasonable argument could be made that the M2s would be preferred. The M2's controlled directivity would give the speaker the advantage in imaging within the sense of space, and the use of MCH within the context of a properly treated room (and perhaps even an untreated room) would cancel any advantage the Salon 2's wide dispersion had in the earlier comparison.

 

[tuga]

You don't need long term listening to do all the things you say. In half hour I play tons of different content. Selection of proper content that is more revealing of speaker differences massively helps with this as well.

Not everyone is as gifted a listenr as you are.

 

[pozz]

I aspire to effectively present the spatial information on the recording while minimizing the "small room signature" inherent to the playback room. Ime doing so calls for some unorthodox manipulation of the reflections, hence my involvement with controlled-pattern polydirectionals.

Toole had a strange remark on this point. Or at least one I've not one I've noticed before or had the ability to test.

He said that the largest source of ambience will dominate the spatial presentation: if a room is more reverberant than the recording, the room's character will lead the listener impression. But if the recording featured stronger reverberance than the room, the recording will lead the listener impression. At least in this equation the speaker has only a contributing, not a dominant role.

There's nothing in the psychoacoustic research I've read that suggests imaging somehow takes on a different character under different reproduction formats. The perceptual mechanisms stay the same. I would argue that increasing the loudspeaker count from one to two doesn't represent a different set of listening conditions from a perceptual standpoint. The only difference is additional information, which does not necessarily make speaker easier to assess.

Let's say you massively increase channel and speaker count. What starts mattering there is the recording and matrixing of playback info. It does not call for a radically different loudspeaker design apart from in WFS systems.

My thoughts so far at least.

 

[Thunder22]

Mono testing is Audio Science Review's method of procedure. It is objective and nobody can argue that. Everything is explained and corroborated. Nothing is hidden. This thread is not about subjective listening even though Amirm's video uses it as an ends to the mean.

Subjective listening could mean almost anything, with unlimited variables and doesn't prove anything.

Maybe a title change is in order; Speaker testing: Why listening subjectively in stereo can or will mask high fidelity while selecting speakers. Disclaimer this does not mean that while you are at home jamming to some tunes, that you have to do it with a single speaker. Nor does it mean that you have to follow this advice, but if you are not here to learn; then what brings you?

 

[Blumlein 88]

Not everyone is as gifted a listenr as you are.

You are working under the wrong assumption. He may be a gifted or at least a highly trained listener. That doesn't mean he can quickly audition and get better results because of that. The situation is all listeners can get better results with short term listening. Though non-intuitive the longer listening sessions hampers one ability to discriminate among real differences.

 

[PeteL]

Another poster mentioned that the type of recordings could affect preference. He stated that classical recordings [or recordings in a theatre/auditorium] would most likely sound better [all other things being equal] with speakers exhibiting constant directivity and wide dispersion, conversely, he also suggested that studio recording may be better served by speakers exhibiting controlled directivity with narrower dispersion [e.g. speakers using a waveguide to narrow the directivity].

This is only a poster opinion, but it is key to this debate, let's accept the premise that a single speaker gives us all information about it's "spatial quality". Sure, it's hard to debate this, if you get great stereo imaging if you put two, it can only be because of attribute of one speaker, they are technically both the same, but the question is how and why? Before talking about music style, does science even know what attribute of a speaker should give us a better, more precise stereo image? Intuitively, we are led to think it's about directivity, but is it?

Amir said himself "the best you can hope from a reviewer is correct assessment of the tonality of the speaker."

So is the hypothesis here that maybe NO attribute of a speaker is linked to better stereo imaging, that if you hear a good stereo image, that would STRICKLY be because of parameters that are not about the physical speaker, or if linked to the speaker attributes, it's because these attributes simply work better in your environment, with this positioning, with your listening position, with your type of music, or simply that they sound better but not that they work great at creating an image Intrinsically. I'm fully willing to accept this hypothesis but so far science didn't explain me that. And to push the questioning further, what is "better" imaging, fundamentally? Is it a very small sweet spot but perfect 3d localisation of the sound element, or is it the hability for a pair of stereo speaker to maintain a coherent sound stage in a larger listening area? those two concepts would appeal to very different speakers attributes no?

Before even trying to think that it's possible to judge the hability of a speaker to recreate a stereo image by listening to just one, do we even know what is it that we want? Does "science" even know what it is we are looking for?

 

[Blumlein 88]

Toole had a strange remark on this point. Or at least one I've not one I've noticed before or had the ability to test.

He said that the largest source of ambience will dominate the spatial presentation: if a room is more reverberant than the recording, the room's character will lead the listener impression. But if the recording featured stronger reverberance than the room, the recording will lead the listener impression. At least in this equation the speaker has only a contributing, not a dominant role.

There's nothing in the psychoacoustic research I've read that suggests imaging somehow takes on a different character under different reproduction formats. The perceptual mechanisms stay the same. I would argue that increasing the loudspeaker count from one to two doesn't represent a different set of listening conditions from a perceptual standpoint. The only difference is additional information, which does not necessarily make speaker easier to assess.

Let's say you massively increase channel and speaker count. What starts mattering there is the recording and matrixing of playback info. It does not call for a radically different loudspeaker design apart from in WFS systems.

My thoughts so far at least.

I've done recordings in large churches with pure spaced omni pairs. You can listen to one channel and it is full range mono. Listening to two channels causes you to get a greatly increased reverberent sound from that space due to the de-correlation between channels which comes from the spacing. Those recordings do have so much 'room sound' from the recording site they always overwhelm the listening room. '

Now same technique in a smaller space say twice listening room size or smaller, and you get a big increase in room sound, but if the listening room isn't heavily damped you don't overwhelm the room's sound in the same sense. I suppose this is what Toole is talking about. The imaging is more precise on these latter recordings, but some of that is due to the closer distance being used to record. And in this sense is also fidelity to the original. You hear a more precise image in person like a small club than if you are listening in a large church to the same group.

 

[tuga]

You are working under the wrong assumption. He may be a gifted or at least a highly trained listener. That doesn't mean he can quickly audition and get better results because of that.

That was my point exactly.

The situation is all listeners can get better results with short term listening. Though non-intuitive the longer listening sessions hampers one ability to discriminate among real differences.

I can see the issues with long-term listening but short-term listening - which is itself not devoid of issues - is insuficient to characterise performance (and in my view inadequate to assess preference), just like the spinorama falls short of producing a complete picture of a speaker's performance.
It would be so nice and a lot easier, I know, but oversimplifying the assessment, whether listening or measured, only leads to incorrect data.
Ultimately SINAD ratings or a Spin are (almost ) just as misleading as some subjectivist reviewer's listening report...

In my view and experience short-term listening must be complemented with long-term listening, just like single-speaker evaluation needs to be complemented with stereo two-speaker evaluation. Let it be clear that I am not defending the idea that long-term listening should replace short-term.
I understand your anger for having been deceived in the past and led to fall for the high-end hoax; you are not alone there I am sure many others myself included feel the same. But it looks to me like the posture that you've now taken is too extreme and dogmatic, and at times preachy... The search for Science knowledge didn't end in Toole's compilation of his and other people's research.

 

[Purité Audio]

I don’t believe anyone would disagree with you, measurements simply allow you ( if you understand them) to weed out the complete lemons.
Ultimately the choice is subjective.
Keith

 

[tuga]

Toole had a strange remark on this point. Or at least one I've not one I've noticed before or had the ability to test.

He said that the largest source of ambience will dominate the spatial presentation: if a room is more reverberant than the recording, the room's character will lead the listener impression. But if the recording featured stronger reverberance than the room, the recording will lead the listener impression. At least in this equation the speaker has only a contributing, not a dominant role.

There's nothing in the psychoacoustic research I've read that suggests imaging somehow takes on a different character under different reproduction formats. The perceptual mechanisms stay the same. I would argue that increasing the loudspeaker count from one to two doesn't represent a different set of listening conditions from a perceptual standpoint. The only difference is additional information, which does not necessarily make speaker easier to assess.

Let's say you massively increase channel and speaker count. What starts mattering there is the recording and matrixing of playback info. It does not call for a radically different loudspeaker design apart from in WFS systems.

My thoughts so far at least.

I find that misleading.
The listening room acoustics will always impact playback recording in the same manner, what will happen is that when the recording has a lot of reverberance (f.e. church choir) the impact of the listening room will be less noticeable (in the case of a lively/reflective room).

Some people enjoy this euphonic effect which results from the overlapping of the listening room acoustic footprint over the recorded acoustic (a room in a room) but in my view and experience it distorts the recreation of the original venue acoustics and detracts from the realism that can potentially be achieve when the listening room is more "quiet/dead" and makes good controlled use of early vs. late reflections.
Toole, and Amir as well, enjoy this euphonic effect and thus may inadvertedly be generating some bias.

 

[tuga]

There seems to be an assumption in this question (I apologize if I'm wrong) that has been repeated by others. Despite the evaluation being mono vs stereo, the question is whether the results would have remained the same in a comparison of 2 speakers vs 1 with a mono signal. I have little reason to assume they would differ (i.e. I think 2 speakers playing a mono signal would exhibit the same advantages as when playing a stereo signal), and the clarification matters because some posters have noted the signal can sometimes be mono, as if to suggest the findings of 1 speaker vs 2 still applied with mono signal.

PS, Questioning Toole's conclusions is not tantamount to attacking his work or his findings, but rather raising doubts about any dogmatic perspectives based on those findings. It should be noted that other highly qualified (with PhDs) and well respected individuals have raised many of these questions and the answers have not been fully addressed. In a previous post I mentioned Dr. Geddes, a highly respected individual in audio research and he has raised many of these questions.

PPS Another poster mentioned that the type of recordings could affect preference. He stated that classical recordings [or recordings in a theatre/auditorium] would most likely sound better [all other things being equal] with speakers exhibiting constant directivity and wide dispersion, conversely, he also suggested that studio recording may be better served by speakers exhibiting controlled directivity with narrower dispersion [e.g. speakers using a waveguide to narrow the directivity]. The logic is recordings from an audiorium/theatre will create a sense of space that is more diffused, consistent with the qualities of the wide dispersion speaker, in contrast studio recordings produce spatial characteristics that are less diffused and highlight the placement of musicians within the space. These points were also raised by Dr. Geddes, and he further noted that over 90% of recordings were done in studios. This of course raised a question, how does the vast discrepancy in % between recording types (studio vs venue) affect one's conclusions about speaker design and priorities? I'm not trying to suggest Dr. Geddes is right, I'm just noting assumptions are being made that disregard many important variables that should also be considered.

PPPS Another area the research may be lacking is how the findings can be extrapolated to MCH. In a particular blind test the Revel Salon 2 performed better than the JBL 2 (once again showing speakers with wide dispersion may be subjectively preferable than better measuring speakers with narrower directivity), but the question not answered was which speaker would perform better in MCH in a properly treated room; i.e. could one extrapolate from the blind test that the Salon 2s would be preferred. I think a reasonable argument could be made that the M2s would be preferred. The M2's controlled directivity would give the speaker the advantage in imaging within the sense of space, and the use of MCH within the context of a properly treated room (and perhaps even an untreated room) would cancel any advantage the Salon 2's wide dispersion had in the earlier comparison.

Very eloquently put, thank you.

In regard to the hypothesis that you put forth in the PPPS, I am led to believe (after 15 years of frequenting audio forums) that many audiophiles are willing to trade sharper phantom image focus for a deeper sense of "immersiveness" or "envelopment" and a wider stage.
Also, for years I though that most classical music listeners would, like myself, prefer a "quieter" room and/or narrow directivity speakers both of which are able to better recreate the recorded ambience (by reducing listening-room interference) but I am now firmly convinced that this is not the case.

 

[richard12511]

Dr. Toole is an ASR member; perhaps you should question his credentials and integrity directly in a PM to him.

Personally I abhor the ongoing "audiophile" attacks on anything and anyone technical or "researchers". We've completely flipped from depending upon the science to ignoring and actively denigrating it.

In my personal system I do tend to try to take the room out of the equation, so agree with that premise, though for most of us it is unrealistic.

Also, saying it's "as simple as a Beolab 90" made me chuckle. That's about as far away from simple as loudspeaker design goes .

 

[richard12511]

Nearly 300 posts of people's feelings yet no one is willing to do the necessary research to prove otherwise.

Setting up and executing good blind tests with multiple people is a lot of work, so it doesn't surprise me that no one has done that work within the time frame of 300 posts on a forum with thousands of members . As I mentioned before, I'm planning a blind listening test GTG in College Station, TX, and anyone is welcome, just PM me. The more, the merrier . We can test mono vs stereo.

Personally, I still haven't made up my mind. I disagree with @amirm's conclusion that the linked Harman study is sufficient to prove that "mono preference will always be the same as stereo preference". Obviously, that statement can't be proven 100%, but I don't mean "proven" in that sense. I mean proven more along the lines of "beyond reasonable doubt". Given the limited data we have, I think @Duke's interpretation is just as valid as Toole's and Amir's. I just don't think there's enough data to make any reasonable definitive conclusion. Duke also rightfully points out that even with that limited data, we actually did have a preference order change from mono to stereo. Spatial quality of AA was preferred in mono, but E was preferred in stereo. The fact that there was no overall order change is just a product of the Harman weighting. Had they weighted the Tonal/Spatial scores differently, there would have been an overall preference change too.

That said, my brain keeps coming back to Toole's statement that they've been doing this for 30 years, and there have been 0 exceptions of mono to stereo preference changing. While I don't think that single study is sufficient, I would have to agree that 30 years worth of similar data probably is. If it really were possible for preference to change from mono to stereo, it almost surely would have happened by now, and Toole's opinion would have changed accordingly. The fact that Toole's opinion hasn't changed after repeat testing really lends credit to his current opinion(imo). @Blumlein 88 pointed this out in a previous post, and I completely agree. Harman has access to much more mono vs stereo data than we do, so I think we should be careful about extracting conclusions based on that one study.

Btw, if you want to read/hear where Toole says this, it's in the below video:

Unfortunately, I don't remember the specific time stamp where he says it, but I would highly recommend watching the video in its entirety anyway.

 

[Jdunk54nl]

Setting up and executing good blind tests with multiple people is a lot of work, so it doesn't surprise me that no one has done that work within the time frame of 300 posts on a forum with thousands of members . As I mentioned before, I'm planning a blind listening test GTG in College Station, TX, and anyone is welcome, just PM me. The more, the merrier . We can test mono vs stereo.

Personally, I still haven't made up my mind. I disagree with @amirm's conclusion that the linked Harman study is sufficient to prove that "mono preference will always be the same as stereo preference". Obviously, that statement can't be proven 100%, but I don't mean "proven" in that sense. I mean proven more along the lines of "beyond reasonable doubt". Given the limited data we have, I think @Duke's interpretation is just as valid as Toole's and Amir's. I just don't think there's enough data to make any reasonable definitive conclusion. Duke also rightfully points out that even with that limited data, we actually did have a preference order change from mono to stereo. Spatial quality of AA was preferred in mono, but E was preferred in stereo. The fact that there was no overall order change is just a product of the Harman weighting. Had they weighted the Tonal/Spatial scores differently, there would have been an overall preference change too.

That said, my brain keeps coming back to Toole's statement that they've been doing this for 30 years, and there have been 0 exceptions of mono to stereo preference changing. While I don't think that single study is sufficient, I would have to agree that 30 years worth of similar data probably is. If it really were possible for preference to change from mono to stereo, it almost surely would have happened by now, and Toole's opinion would have changed accordingly. The fact that Toole's opinion hasn't changed after repeat testing really lends credit to his current opinion(imo). @Blumlein 88 pointed this out in a previous post, and I completely agree. Harman has access to much more mono vs stereo data than we do, so I think we should be careful about extracting conclusions based on that one study.

Btw, if you want to read/hear where Toole says this, it's in the below video:

Unfortunately, I don't remember the specific time stamp where he says it, but I would highly recommend watching the video in its entirety anyway.

It’s not just 300 posts though, that is just on here. This research has been available from Toole for many years and hasn’t been refuted as far as I know or could find.

I appreciate you taking on setting up a blind test. It is a great start but, from what I read, would not be up to the standards of science to be able to refute Toole’s research, which I’m also sure you are aware. It is still better than 99% of people (including myself). I’m trying to get into some abx tests with a friend that designs speakers for a really big audio company, so hopefully soon. I don’t have the resources for an abx test nor do I want to take that on. I’ve done some education research that could have been published and peer reviewed. To officially publish that, the amount of extra work involved was not worth it to me and my peers doing it.

So I’m at the mercy of the available research, and the papers I’ve read on it seem to be good and unbiased as to what the results would turn out to be. It really seems that all parties involved have nothing to lose if the research turned out to say different.

So, like I said in the last post, if the research throughout the years, officially published or not, would say different than single speaker testing, what does any party have to lose by changing their approach? If it was better, they’d have plenty to gain by changing their approach! They could produce worse speakers and sell them for the same price!

 

[Thomas savage]

If you would be interested, I can describe what I believe to be a stereo setup where something desirable happens to the spatial quality which cannot be evaluated by testing a single speaker. No magic, but some unorthodox thinking from the mind of Earl Geddes. I would be describing a situation which is applicable to a relatively small number of loudspeakers, which are designed to be used a certain way. Let me know if you'd like for me to proceed.

That would be brilliant, I'm assuming this is a judicial change in a known parameter of the speakers behaviour that can be observed under testing . So can be seen in a single speaker but not experienced subjectively while being restricted to that single speaker arrangement.

New thread!

While one can see certain fringe designs might benefit from stereo playback as they are designed exclusively for that I'm still waiting for evidence this cant be seen in the behaviour of a single speaker once one knows what they are looking for.

Personally I don't see the point of such designs, if you have a reliably performant speaker you can with software create expansive and mind boggling spatial trickery via digital single processing through software.

The job of software engineers who are involved in such developments would be made easier if there were universal standards for speaker design. Fortunately for the most part there is concensus on what makes a good speaker and @amirm is following that consensus in his testing, Hooray!

The future is in advancement in DSP and perceptual spatial based software.

Hardware is yesterdays news .

 

[Thomas savage]

You are working under the wrong assumption. He may be a gifted or at least a highly trained listener. That doesn't mean he can quickly audition and get better results because of that. The situation is all listeners can get better results with short term listening. Though non-intuitive the longer listening sessions hampers one ability to discriminate among real differences.

More time for the imagination to slip in its influence... , insidious haha

 

[richard12511]

It’s not just 300 posts though, that is just on here. This research has been available from Toole for many years and hasn’t been refuted as far as I know or could find.

The first part of my post to you was in jest, which is why I used the "". I actually agree with you here. IMO there's a lot of "Toole's research is obviously wrong" with usually no data at all to back up those opinions. I think more data is needed to make a conclusion either way, and I think Harman has this data. I'm sure they do, and this is where's Toole's opinion is coming from.

I appreciate you taking on setting up a blind test. It is a great start but, from what I read, would not be up to the standards of science to be able to refute Toole’s research, which I’m also sure you are aware.

Not a big fan of this. What you say is true(my test won't be as valid as the Harman tests), but it's the wrong attitude to have(imo). Not really sure how to say this, but it seems unfair(?) to complain(?) about no one bringing any data of their own to the table, and then when do, telling them that their data can never be valid, anyway. Unfair and complain are probably the wrong words, but I couldn't really properly express myself there. I think we should be encouraging more blind tests. They may not be as useful as the most carefully controlled tests with huge sample sizes, but that doesn't mean they have no use. I just see it as a shame. I think it's the same reason @echopraxia ended up giving up on his Salon2 vs 8351b blind test. My test no doubt won't be perfect, but I certainly intend to do the best I can(which is why I just bought a $1,100 speaker switcher).

Biggest weakness with my test will no doubt be sample size. No way am I getting near the 300 people that Harman had. Other biggest weakness of my test will be that I don't have trained listeners, which is (imo) is a huge negative. I do think there are some items I can improve upon what Harman did, the main one being taking the time to setup each speaker in its best position and toe in. The Harman tests put every speaker in the same position and toe in, which advantages wider dispersion(Toole says this himself).

So I’m at the mercy of the available research, and the papers I’ve read on it seem to be good and unbiased as to what the results would turn out to be. It really seems that all parties involved have nothing to lose if the research turned out to say different.

Agreed, and the fact that Harman is still using mono testing as it's primary method 30 years later makes me think that they really do believe it, and it makes me believe they're probably right. They're ultimately in this to make money. If stereo testing had any merit, Harman would be using it. The fact that they don't speaks volumes.


I still think it would be interesting to blind test something like the Beolab 90 in mono, wide, and narrow settings in both mono and stereo. If those tests gave the exact same results with a large enough sample size, I bet that could convince a lot of the skeptics.

 

[Jdunk54nl]

Not a big fan of this. What you say is true(my test won't be as valid as the Harman tests), but it's the wrong attitude to have(imo). Not really sure how to say this, but it seems unfair(?) to complain(?) about no one bringing any data of their own to the table, and then when do, telling them that their data can never be valid, anyway. Unfair and complain are probably the wrong words, but I couldn't really properly express myself there. I think we should be encouraging more blind tests. They may not be as useful as the most carefully controlled tests with huge sample sizes, but that doesn't mean they have no use. I just see it as a shame. I think it's the same reason @echopraxia ended up giving up on his Salon2 vs 8351b blind test. My test no doubt won't be perfect, but I certainly intend to do the best I can(which is why I just bought a $1,100 speaker switcher).

Biggest weakness with my test will no doubt be sample size. No way am I getting near the 300 people that Harman had. Other biggest weakness of my test will be that I don't have trained listeners, which is (imo) is a huge negative. I do think there are some items I can improve upon what Harman did, the main one being taking the time to setup each speaker in its best position and toe in. The Harman tests put every speaker in the same position and toe in, which advantages wider dispersion(Toole says this himself).

The reading of my words may have come off wrong or too strong "against" what you are doing, I am by no means discounting what you are doing. I think it is awesome and if college station wasn't so far away, I would be there in a heart beat. Like I said, it is more than what 99% are willing to put forth. It is more than what I am willing to do on my own. Phoenix does have some great resources and people to put something together. Maybe I can get @SIY to help put something together with our FNA (Food and Audio) monthly meetup group.

At the same time though, we "normal" people can only do so much (unfortunate but true). We don't have the budget, time, or space to properly do this. What data we can gather can be eye opening to those involved and WELL WORTH that outcome. But ultimately, we have too look at the research provided from others, if we have no evidence to provide of the contrary and see no biases being presented in the research or major issues then the conclusion should be that the research is sound and should be listened to.

The research that I have done, the time commitment is ridiculous just to gather the data in an "appropriate" manner that would be able to be published. As in, there was not holes in our thinking as far as we could tell (there probably still was) and it was all legally appropriate (we were conducting research on minors which has a BUNCH of hoops to jump through). Analyzing the data was another huge time suck. Writing the initial paper was again a huge time suck, and this wasn't even close to being able to be ready to publish yet. In the end, our research showed no statistical benefits, we would need more data collected to make any conclusion based on any of the objective data. Subjective data was promising, but we know how that goes.

 

[tuga]

Agreed, and the fact that Harman is still using mono testing as it's primary method 30 years later makes me think that they really do believe it, and it makes me believe they're probably right. They're ultimately in this to make money. If stereo testing had any merit, Harman would be using it. The fact that they don't speaks volumes.

What if Harman's interpretation of the data is in fact wrong?
The problem is that a whole listening assessment methodology which includes a purposely-built listening room is based in that sole assumption.

The logistical or practical advantages of performing single-speaker listening tests are indisputable though. Imagine having to set up pairs of speakers, adequately positioned for a sweet-spot listener and to conceive and build a shuffler that would take optimal setup into consideration.
Also if you are a proponent of wide directivity (even though it is a matter of preference, thus subjective) then it may make sense to conclude that the assessment method (single vs. pair) which better descriminates this parameter is the better method for assessing "spatial quality".

 

[tuga]

Dali speakers are designed to be listened to with no toe-in, which means that the axis should be pointing 30º away from the listening spot.
From my understanding this requirement would not be taken into consideration by the Harman testing method.
Yet I suspect that even though it would perform poorly in a single-speaker assessment due to the freakish tonal balance on axis these are speakers which when adequately set up (equilateral triangle, no toe-in) would probably rate very highly amongst early-reflection lovers due to their very wide horizontal directivity.

source: https://www.stereophile.com/content/dali-rubicon-8-loudspeaker-measurements