Speaker Testing: why mono is better

[Duke]

We have the science of what is good measurements and what is good listening protocols. Combined, they produce defensible, highly reliable conclusions. [emphasis Amir's]

IF spatial quality really does matter (see section on Klippel's study, page 185-186 in 3rd edition of Toole's book), and IF good directivity measurements and good mono listening produce defensible, highly reliable conclusions (which are more accurate than stereo listening), then WHY do your reviews not include your conclusions about spatial quality?

 

[richard12511]

Can the spatial performance of a time/intensity trading configuration be evaluated by controlled blind listening to a single loudspeaker? If so, how?

So, the scenario is:
Speaker A is preferred over speaker B in mono
Speaker A is preferred over speaker B in Harman style stereo(speakers pointed straight ahead)
Speaker B is preferred over speaker A in any toe in style stereo(by utilizing extreme toe in, which speaker B can't utilize).

The question is, is there any way to setup the mono test so that speaker B wins? I would think the answer is no, which proves your point I suppose.

IMO, the above scenario is definitely possible. Possible, but how likely? This is a situation that wouldn't come up in Harman blind tests, since they point the speakers straight ahead(at least afaik). Would be interesting to test this edge case many times with many speakers to see if it could break the "mono and stereo preference are always the same" theory. In my sample size of 1, it's not enough. Extreme toe in definitely closes the gap between my JTRs and Genelecs quite a bit, but I still prefer the Genelecs overall in stereo.

Thinking about edge cases that could break the rule.

Maybe a Danley SH50(50°) vs Ascend Sierra 2(160°)?

It's not hard to imagine that the Ascend would win in a Harman style stereo blind test(no toe in).
The Danley's are even more extreme than my JTRs, and even the JTRs do quite poorly with no toe in(unless you sit really far back). Imaging is weak, and what imaging is there completely collapses if you move your head and inch to the left. The Ascend will likely have the image clarity advantage, and will definitely have the image stability advantage here.

Now allow them to be toe in the way that works best for each speaker. With extreme toe in, the Danley will now definitely have the image clarity advantage, and possibly even the image stability advantage. The Ascend's sound doesn't really change all that much with toe in changes, so it doesn't really improve much at all, though it will still have the envelopment advantage.

Question is then, is it possible that changing those two spatial quality advantages from one speaker to the other is enough to switch the overall preference?

Rather extreme example, I know, but I think it's gonna take something extreme like this to find the exception to the rule. If Harman's been trying to find the exception now for 30 years, and still hasn't, then we need to do something different(ie allow free form toe in).

 

[Kal Rubinson]

Dolby. which is already a rather lenient recommendation.

The real 'standardized' positioning is agreed upon by the industry in the ITU-R BS.775-1

View attachment 120837

Equal distance between you and all the speakers.

Hmmm. Not seen that one but I have seen similar from Dolby and other with different recommendations but that's working from memory. In practice, as mentioned, having the surrounds at 90deg works well for quite music if it the surround signals are primarily ambiance.

 

[abdo123]

In practice, as mentioned, having the surrounds at 90deg works well for quite music if it the surround signals are primarily ambiance.

well looks like the multichannel effect is masking the imperfect placement of your speakers .

I personally think the angles must be respected, as for the distances between speakers time delay can help.

 

[Inner Space]

At the end of the day, this translates into kabuki theater for some of you. You worry that your favorite speaker won't measure well, or sound well in mono so make up theories about how some other measurement or listening test method would vindicate you. Well, no such thing exists. We have the science of what is good measurements and what is good listening protocols. Combined, they produce defensible, highly reliable conclusions. This, is our charter. Not wishful thinking to make everyone happy with their choice of speakers.

Honestly, I think that's a little patronizing. It's not ego on the line here - it's logic. The good measurements and listening protocols do a fine job of describing a single speaker's quality playing in mono. If you want to say that's automatically 100% meaningful for a pair of the same speakers playing in stereo ... maybe you're right, but then why not just test one channel of electronics and make the same assumption? But you don't, and occasionally you notice problematic mismatches, which you note as demerits.

I wouldn't ask for stereo listening tests. Some extra measurements are all that's needed. Pair matching, and localized cabinet disturbances. Then you could say, e.g., "This speaker passed the mono tests with flying colors, had no cabinet talk, and its mate measured exactly the same, so I confidently expect it to do well in stereo." Or conversely: "This speaker was great in mono, but it had a slight woody murmur, which was inoffensive or even agreeable in mono, and its mate measured very differently, both of which defects will tend to compromise its imaging stability, so I can't recommend it for stereo."

My bottom line is when I worked in big studios with big staffs, every single thing bought in, even from known and trusted suppliers, was immediately tested for spec - and any technical manager who assumed just because one channel worked OK, the other probably would too, would have been fired on the spot. For stereo, both speakers should be assessed - again, not necessarily listened to, but in some way tested for its fitness, which is always and explicitly as one half of a pair.

 

[richard12511]

Honestly, I think that's a little patronizing. It's not ego on the line here - it's logic. The good measurements and listening protocols do a fine job of describing a single speaker's quality playing in mono. If you want to say that's automatically 100% meaningful for a pair of the same speakers playing in stereo ... maybe you're right, but then why not just test one channel of electronics and make the same assumption? But you don't, and occasionally you notice problematic mismatches, which you note as demerits.

I wouldn't ask for stereo listening tests. Some extra measurements are all that's needed. Pair matching, and localized cabinet disturbances. Then you could say, e.g., "This speaker passed the mono tests with flying colors, had no cabinet talk, and its mate measured exactly the same, so I confidently expect it to do well in stereo." Or conversely: "This speaker was great in mono, but it had a slight woody murmur, which was inoffensive or even agreeable in mono, and its mate measured very differently, both of which defects will tend to compromise its imaging stability, so I can't recommend it for stereo."

My bottom line is when I worked in big studios with big staffs, every single thing bought in, even from known and trusted suppliers, was immediately tested for spec - and any technical manager who assumed just because one channel worked OK, the other probably would too, would have been fired on the spot. For stereo, both speakers should be assessed - again, not necessarily listened to, but in some way tested for its fitness, which is always and explicitly as one half of a pair.

He only has one speaker to test, though. Asking people to send in 2 samples will double shipping costs and reduce throughput.

 

[abdo123]

For stereo, both speakers should be assessed - again, not necessarily listened to, but in some way tested for its fitness, which is always and explicitly as one half of a pair.

Quality control is a completely different thing.

the reason behind the measures you mentioned is accuracy, not because stereo is this physics rules breaking phenomenon that should be carefully monitored and studied.

 

[amirm]

IF spatial quality really does matter (see section on Klippel's study, page 185-186 in 3rd edition of Toole's book), and IF good directivity measurements and good mono listening produce defensible, highly reliable conclusions (which are more accurate than stereo listening), then WHY do your reviews not include your conclusions about spatial quality?

Because that is not why I listen to speakers. I listen to speakers to see if the measurements are correct. When frequency response shows aberrations, I attempt to verify those using EQ and report on experience with that. I explained all of this in my recent video on headphone testing:

Spatial qualities are presented as pure measurements and comments are made on them. I have no reason to use my ears to second guess dispersion, etc.

 

[Inner Space]

He only has one speaker to test, though. Asking people to send in 2 samples will double shipping costs and reduce throughput.

Yeah, no question it would be more difficult, but is that a good enough reason? ASR ... we dodge the hard stuff here?

 

[amirm]

P.S. I do provide feedback on spatial quality for headphones as externalization is so important there.

 

[amirm]

Yeah, no question it would be more difficult, but is that a good enough reason? ASR ... we dodge the hard stuff here?

Excuse me? Buying a $100K system, spending an entire day measuring, listening and writing reviews is not "the hard stuff?" Where do you get such notions? I could stick a mic in front of the speaker and give you a gated response. That, would have been the easy route. Everything else I do is hugely above easily level. Testing in stereo doesn't even have precedence in research and you expect me to invent protocols and spend the time doing it?

What is wrong with you all by the way? Why don't you do some evaluations in stereo and get back to us? Too hard for you?

 

[Inner Space]

Quality control is a completely different thing.

the reason behind the measures you mentioned is accuracy, not because stereo is this physics rules breaking phenomenon that should be carefully monitored and studied.

Yeah, we live in a two-channel ecosystem, so let's measure more than half of it.

 

[amirm]

Honestly, I think that's a little patronizing. It's not ego on the line here - it's logic. The good measurements and listening protocols do a fine job of describing a single speaker's quality playing in mono.

Huh? You are another person who didn't watch the video, nor read or understood the research??? How many times do we have start over and explain this concept?

No, no, no. Mono testing shows the true characteristics of the speaker including its most important aspect: tonality. Everything else is vastly secondary when it comes to listener preference. This matters just as well in mono as it does in stereo. If you don't believe this, then you should not care what we do here. Buy any speaker that looks pretty to you and be done with it.

If you want to say that's automatically 100% meaningful for a pair of the same speakers playing in stereo ... maybe you're right, but then why not just test one channel of electronics and make the same assumption? But you don't, and occasionally you notice problematic mismatches, which you note as demerits.

You very well could. You need to work more on your analogies. Other than headphone and power amps where there may be power sharing between channels, you absolutely could measure one channel and be done with it. I provide the other channel because it comes for free given that my analyzer is stereo. Otherwise, there would not be much need. And indeed audio analyzers used to be mono for this reason.

 

[Duke]

Because that is not why I listen to speakers. I listen to speakers to see if the measurements are correct.

Okay, let's assume the measurements are confirmed to be correct.

IF you wanted to, could you "produce defensible, highly reliable conclusions" (your words in post number 367) about the loudspeaker's spatial quality?

 

[amirm]

I cannot. I have experience, controlled setup and content, testing tonality (and distortion) for 150 speakers. I have nothing like that for spatial qualities.

 

[Inner Space]

Excuse me? Buying a $100K system, spending an entire day measuring, listening and writing reviews is not "the hard stuff?" Where do you get such notions? I could stick a mic in front of the speaker and give you a gated response. That, would have been the easy route. Everything else I do is hugely above easily level. Testing in stereo doesn't even have precedence in research and you expect me to invent protocols and spend the time doing it?

What is wrong with you all by the way? Why don't you do some evaluations in stereo and get back to us? Too hard for you?

I lived with minute-by-minute stereo evaluation for nearly 20 years, 200m album sales, numerous awards - and I'm getting back to you. Testing half of a transducer system is inadequate. If you were testing a stereo microphone, would you stop after one channel?

I get that it's all expensive and tiring and time-consuming, but if you're going to do it, then do it properly. Don't dig yourself into a defensive position.

 

[amirm]

I lived with minute-by-minute stereo evaluation for nearly 20 years, 200m album sales, numerous awards - and I'm getting back to you. Testing half of a transducer system is inadequate.

It is clear you are confusing content creation with evaluation of playback gear. Tell us about experience and research of evaluating speaker systems in controlled, defensible manner.

If you were testing a stereo microphone, would you stop after one channel?

I would test measure each channel separately to make sure I know of variations between them. What that has to do with the topic at hand, I don't know.

I get that it's all expensive and tiring and time-consuming, but if you're going to do it, then do it properly. Don't dig yourself into a defensive position.

It has nothing to do with expense. I already spent the big dollars. It is about wishful thinking and personal opinion not backed by any research. Do some work yourself and then you realize how hard it is to evaluate speakers in mono and how impossible it can get if you double to two channels. Don't confuse this work with producing albums. I have no clues about that field so would not remote presume to give you advice on that. I do however know this topic and ask you to at least bother to get educated on the topic before throwing in strong commentary.

 

[Duke]

I cannot. I have experience, controlled setup and content, testing tonality (and distortion) for 150 speakers. I have nothing like that for spatial qualities.

Just to make sure I understand correctly, are you saying that you cannot "produce defensible, highly reliable conclusions" about spatial quality from your directivity measurements and/or mono listening?

In post #127 you wrote, "Research shows that mono testing is more accurate than stereo testing."

Is that statement true for spatial quality?

(For the record: I am NOT suggesting that you do stereo testing! I am however skeptical about mono testing being a more accurate predictor of spatial quality than stereo testing.)

 

[amirm]

So it is clear: we all know everything in your bones tells you that speaker testing should be done in stereo. Please, please, don't keep repeating this. The purpose of the video was to show that this notion is not correct and that the most proper method of testing speakers is in mono. Some performance aspects may be left on the table and that is fine. Subjective testing is not about perfection. Measurements are for that.

 

[abdo123]

Yeah, we live in a two-channel ecosystem, so let's measure more than half of it.

When stereo enters the speaker design process maybe, otherwise it's pointless.

I lived with minute-by-minute stereo evaluation for nearly 20 years, 200m album sales, numerous awards - and I'm getting back to you. Testing half of a transducer system is inadequate. If you were testing a stereo microphone, would you stop after one channel?

I get that it's all expensive and tiring and time-consuming, but if you're going to do it, then do it properly. Don't dig yourself into a defensive position.

you do understand that Amir is not the manufacturer of any speaker that is measured on this forum?

and that it is not his responsibility to ensure that any two speakers in a production batch match?

he can only critique the unit he has, and he also has on multiple occasions measured other samples when other members sent them.