Why don't all speaker manufacturers design for flat on-axis and smooth off-axis?

[amirm]

Superior and socially-acceptable wide dispersion JBL LSR 308 vs inferior MartinLogan hybrid only suitable for hermits and other deplorables as measured at the listening position in my room without averaging and with spawn of satan flat equaliztion applied to both:

Frequency response measurements like this need to be divided into two regions:

1. Region 1 below transition frequency of a few hundred hertz. Here, we want to see 1/12 to 1/24 octave filtering to see modal response.

2. Region above transition frequencies. Due to reduction of our hearing resolution proportional to rise in frequency, the graph needs to be filtered to 1/6th octave. If this is one, all those comb filtering effects (spikes) go away, giving us a much clearer view of the overall tonal response.

Ultimately though, in-room measurements don't correlate as well to our preference as anechoic ones do.

 

[RayDunzl]

Due to reduction of our hearing resolution proportional to rise in frequency, the graph needs to be filtered to 1/6th octave. If this is one, all those comb filtering effects (spikes) go away, giving us a much clearer view of the overall tonal response.

Ok.

But I want to see some 1/6th octave averaging on those DAC measurements, just to be fair.

 

[Daverz]

Frequency response measurements like this need to be divided into two regions:

1. Region 1 below transition frequency of a few hundred hertz. Here, we want to see 1/12 to 1/24 octave filtering to see modal response.

2. Region above transition frequencies. Due to reduction of our hearing resolution proportional to rise in frequency, the graph needs to be filtered to 1/6th octave. If this is one, all those comb filtering effects (spikes) go away, giving us a much clearer view of the overall tonal response.

Ultimately though, in-room measurements don't correlate as well to our preference as anechoic ones do.

REW's "Var Smoothing" roughly corresponds to this:

"Variable smoothing applies 1/48 octave below 100 Hz, 1/3 octave above 10 kHz and varies between 1/48 and 1/3 octave from 100 Hz to 10 kHz, reaching 1/6 octave at 1 kHz. Variable smoothing is recommended for responses that are to be equalised"

 

[Thomas_A]

Well, the use of a center channel inherently corrects for that. Of course, there still are always "phantom images" between adjacent speakers in stereo and multichannel.

Yes but when you play with only two speakers. Thats the point.

 

[Blumlein 88]

And that is how I was until I attended the Harman blind test. It instantly sensitized me to this issue. Now I notice it in every demo I listen to.

BTW, the Quad was tested in the AES paper I quoted earlier. Here is what the bypass test showed with them:

View attachment 29638

These speakers have serious resonances and other problems that are ignored when one falls in love with their diffused sound.

Maybe an interesting addition to that test would have been a speaker in a real room with the microphone and not an anechoic one. The colorations would be increased for all, and not at all hard to tell from the bypass. But would a surprise result like the Quad performing better in this situation than those other speakers been the result?

I'm not looking at the Quad with rose colored glasses I don't think. I'm very aware of the ways they aren't fidelity to the input. But they seem to make many more people just stop and listen than other speakers for some reason. I've likely mentioned before about a lady taking up money for a charity at my side door. I had music playing inside. I excused myself to get a donation for her, when I returned she had stepped inside. Was looking at the speakers in something of a trance by the music. Was surprised and embarrassed when I returned. She sheepishly said, "I'm sorry, ......I shouldn't have....it sounded like someone was really playing music".

My experience of speakers designed by Harman is very good. They aren't captivating or mesmerizing. They do just simply play the music, and it is very right though not perfect. You soon don't worry about the speakers. Since everything ahead of them is of good fidelity you are left with the music and unfortunately the often poor recordings. It doesn't make them worse, and good recordings are very good. The speakers just don't draw attention to themselves. Exactly what a speaker with fidelity should do. I can of course apply a Quad curve to them as well.

 

[Blumlein 88]

Frequency response measurements like this need to be divided into two regions:

1. Region 1 below transition frequency of a few hundred hertz. Here, we want to see 1/12 to 1/24 octave filtering to see modal response.

2. Region above transition frequencies. Due to reduction of our hearing resolution proportional to rise in frequency, the graph needs to be filtered to 1/6th octave. If this is one, all those comb filtering effects (spikes) go away, giving us a much clearer view of the overall tonal response.

Ultimately though, in-room measurements don't correlate as well to our preference as anechoic ones do.

@RayDunzl could use the ERB curve in REW. That will fix it all up nicely. The one I prefer for loudspeakers myself. But it won't switch to 1/24th in the low end for modal viewing. He could show full range ERB and also 200 hz and below 1/24th smoothing.

 

[RayDunzl]

@RayDunzl could use the ERB curve in REW. That will fix it all up nicely. The one I prefer for loudspeakers myself.

Doesn't change what I hear.

 

[STUDIO51]

Are you sure that is the only reason? You may want to check this link..

Yes, Dr. Sean Olive said he prefers M2 but he uses salons mainly because of WAF . There are some reasons for preferring M2, It may be because M2 has better objective performance including spinorama. Only He know the truth.



I have been interested in that link for a long time. There are some parts that I personally disagree with.

The first reason is that the test is a mono speaker test. I know that a mono speaker test is more reasonable for a quick switch. But we do not usually use the speakers in front. HRTF (Head-related transfer function) of the person is 0 degree 10 degrees 20 degrees 30 degrees all different. As a result, the same speaker is recognized differently when listening at 30 degrees and at 0 degrees.

The second reason is general listeners can also make incorrect choices because they are not familiar with mono sound sources. I sometimes listen to a mono speaker, which is more damaging to the sound image and darker than when it is stereo, so I prefer speakers with low bass and high treble in mono test.

 

[Juhazi]

Amirm critisized me saying that B&W speakers get compensation from the room. This Sterophile test says the same.. But several tested models are different, not all do it as well as this one. The '80s Matrix models were far better - it was before they idolized their too wide FST-midrange driver and egg-tweeter.

Ray shows unsmoothed responses above, the blue line is obviously ML panel. I have done the same comparison several times between my dipoles and normal 2- or 3-ways. in my wide room the difference is not that big. My setup has speakers on the long wall, almost 2m from sidewalls. The dipoles are at 30deg angle, beams crossing just in front of the listener. This way backwave is projected further sideways to get a long path before reaching ears. Recommended!

 

[amirm]

The first reason is that the test is a mono speaker test. I know that a mono speaker test is more reasonable for a quick switch. But we do not usually use the speakers in front.

Mono is used because listeners are more picky about performance of speakers that way. The more speakers, the less picky they become, ostensibly blinded by the spatial effects of more than one channel.

The above graph is from their Room EQ article. With no EQ, ratings for mono (blue) were much harsher than with surround. Indeed with surround there was little difference between EQ systems (or not).

Harman does have a multi-channel speaker shuffler room but they realized it is not necessary per above:

 

[amirm]

The second reason is general listeners can also make incorrect choices because they are not familiar with mono sound sources.

None of us who took the tests at Harman had difficulty assigning scores to mono speakers.

Same analogy exists in video. Professional monitors have a button that turns off color so that black and white images (luminance) can be examined better. Color makes the display pretty but makes it harder to see artifacts.

 

[STUDIO51]

None of us who took the tests at Harman had difficulty assigning scores to mono speakers.

Same analogy exists in video. Professional monitors have a button that turns off color so that black and white images (luminance) can be examined better. Color makes the display pretty but makes it harder to see artifacts.

Mono speaker test may be effective in fast finding defective speakers. However, since the two speakers used in the shoot out are defect-free speakers, I think stereo listening is more appropriate.

 

[Cosmik]

Yes, I read that part of the Grimm paper with some puzzlement. It says of an omni speaker: "The sound stage may be quite con-vincing but the brass section sounds ghostly, as if it’s pointing in all directions. Well, it is. Not in the re-cording, but during replay. "

I'm not sure exactly what they are talking about there.

I guess they're saying that if you have two sources, one very directional and one quite omni, and record them dry (as might be typical) and replay them over an audio system, they will both be reproduced with whatever dispersion pattern the speaker imposes on them. There will be no difference in the level of ambience that the listener's room applies.

As I have noticed, many audiophiles listen in rooms with fashionable wooden floors (except for a vestigial token rug and maybe some token absorber thingies on the wall) so this level of ambience might be high compared to those of us who favour a carpeted room. In fact we may often be talking about completely different levels of self-imposed acoustic problems.

Maybe the Grimm people were listening to their speakers in a factory unit at the time...

 

[amirm]

Mono speaker test may be effective in fast finding defective speakers. However, since the two speakers used in the shoot out are defect-free speakers, I think stereo listening is more appropriate.

Well, research shows that to NOT be the case. Here is Dr. Toole from is book:

So even in spatial qualities, mono testing works better. We are not talking about anything being broken either.

 

[STUDIO51]

Well, research shows that to NOT be the case. Here is Dr. Toole from is book:

View attachment 29666

So even in spatial qualities, mono testing works better. We are not talking about anything being broken either.

I also own and read Sound Reproduction. I also know about these results and his vast research is very respectful. but my personal thoughts and experiences do not match. I would like to have an in-depth conversation with Dr. Floyd Toole later.

 

[March Audio]

And may I add: This is not a critique against @Floyd Toole , Olive and Harman/Samsung for promoting their research. @Floyd Toole is AES’s most merited member. I am talking here about their followers that do the idolising which means curiosity and an open mind may get lost. To me, curiosity and open mindedness are big drivers of fun too

....mmm...you keep making this point. No-one is idolising, Toole et al are just being used as the go to reference because it is very accessible and there is very little , if any, evidence to contradict the research that has been performed. If you think the research is wrong say so and point out why with your supporting evidence. Otherwise you have no point. We are not all blindly following the individuals. The science speaks for itself.

 

[Krunok]

Yes, Dr. Sean Olive said he prefers M2 but he uses salons mainly because of WAF . There are some reasons for preferring M2, It may be because M2 has better objective performance including spinorama. Only He know the truth.



I have been interested in that link for a long time. There are some parts that I personally disagree with.

The first reason is that the test is a mono speaker test. I know that a mono speaker test is more reasonable for a quick switch. But we do not usually use the speakers in front. HRTF (Head-related transfer function) of the person is 0 degree 10 degrees 20 degrees 30 degrees all different. As a result, the same speaker is recognized differently when listening at 30 degrees and at 0 degrees.

The second reason is general listeners can also make incorrect choices because they are not familiar with mono sound sources. I sometimes listen to a mono speaker, which is more damaging to the sound image and darker than when it is stereo, so I prefer speakers with low bass and high treble in mono test.

I was referring to the results of the blind test, not his preference, and in that test vast majority preferred Salon2.

Saturday: Revel Salon2s preferred by about an 80% - 20% margin.

Sunday - Revel Salon2s preferred by about an 65% to 35% margin.

 

[Krunok]

My understanding is that both @Floyd Toole and Olive, like many others, fully acknowledge that room correction is beneficial at low frequencies, to eliminate room modes at a single listening position. This is quite uncontroversial and Toole describes such techniques in his book. As far as I know, both of them are much more sceptical of the benefit of room correction products at higher frequencies, especially if the loudspeaker is already good. I don't think there's much inconsistency there.

This is absolutely true and I will post a few quotes here from one of his lectures "Making a good loudspeaker - Imaging, space and great sound in rooms".

I believe this one says it very clearly:

Here is the one where dr. Tool distinguishes resonanes from acoustical interference. This is inportant because resonances can be equilased while acoustical interferences cannot:

Very important slide - don't make EQ based on a single sweep made at one point, what you need to make good equalisation are spatial measurements and aim for resonances:

Here dr. Tool is addresing another myth that filters are bad because they add phase shift:

Let me conslude wtih this slide which leaves no doubt about dr. Tools attitude toward room EQ:

 

[Bjorn]

Indeed, the position AFAIK is that minimum phase issues can be addressed, non mp cannot.

Which would imply a lot of the room modes cannot be fixed either, since a great deal of them are not minimum phase behavior. A general statement that room modes can be corrected by room correction is simply wrong.

 

[Krunok]

Which would imply a lot of the room modes cannot be fixed either, since a great deal of them are not minimum phase behavior. A general statement that room modes can be corrected by room correction is simply wrong.

In my limited experience pretty much all peaks can be fixed, dips not that much.

Here is a page from Dr. Tool's "Maximizing loudspeaker performance in rooms":