Speaker dispersion patterns, and soundstage depth

[Keith_W]

Before I start - i'll say this. I am asking a question, not making an assertion. My question is: what influence does the speaker dispersion pattern have on perceived soundstage depth? 

 

For years, I have had the understanding that speakers with a narrow dispersion with little rear dispersion are better, because of less room interaction. I suppose this makes them more predictable from a designer's point of view. From what I have been told: speakers that beam the high frequencies more tend to project the soundstage forwards. And according to this theory, speakers that project rearwards should have all sorts of early reflections, issues with time smearing (i.e. reflections arriving before the Haas threshold), and thus sound horrible. 

 

For those who don't know what the Haas threshold is (a.k.a. The Law of the First Wavefront), it states that sounds that arrive at our ears within the first 20ms are perceived as direct sound (in reality the Haas threshold varies for different types of sounds - typically between 5ms to 40ms). Any sound that arrives later than this is perceived as an echo. It takes about 3ms for sound to travel one meter. Therefore, if you position an omnidirectional speaker 1m from the rear wall, the reflection will arrive at your ears within 6ms of the direct sound - well within the Haas threshold. Therefore it should result in smearing. 

 

These are my observations. I have listened to close to a couple of hundred systems in my journey. From what I have heard: 

 

 

1. Nearly every speaker I have heard that projects most of its high frequencies forward tends to have a soundstage that is at the same plane or forward of the speaker. This includes every horn I have listened to, with few exceptions (including: my own Acapellas, Spharon, Avantgarde Solo, Uno, Duo, Trio, Zingali, Red Spade PSE-144, various JBL's, Altecs, Azura, waveguides, and I may have forgotten a few). Also includes wide baffle speakers like most Osborn designs. The only exceptions are some Tannoys, like the Westminister. 

 

 

2. Nearly every speaker I have heard that projects high frequencies rearwards have a more open and spacious sound. The soundstage seems to be to the rear of the speaker plane, with better layering. This includes - ESL dipoles (I have heard Magnepans, Quads, Accustats, Martin-Logans), conventional dipoles (Kyron), and omnidirectional speakers (MBL radialstrahler, German Physiks, B&O). The Kyron even has a rear firing tweeter. When I asked the designer why he did this, he said that it makes the speaker sound more spacious. 

 

 

3. Conventional boxed speakers with a narrow baffle, or a pinched waist around the tweeter also do a great job of image layering and rear soundstage projection. I presume this is because less of the high frequencies are reflected off the baffle and are allowed to creep rearwards. Examples I have heard - nearly all bookshelf speakers, B&W Nautilus 800, 801, 802, SGR Octagons, Vivid Giyas. 

 

In other words, my perception seems to contradict the theory. 

 

My questions: 

 

- has this also been your experience? 

- does anybody have a psychoacoustic explanation as to why omnidirectional speakers seem to sound more spacious? 

- what measurement correlates with soundstage depth, anyway? 

Edited September 25, 2016 by Keith_W

[THOMO]

You left out rear reflecting speakers like Bose 901s.

[Happy]

4 minutes ago, THOMO said:

You left out rear reflecting speakers like Bose 901s.

 

But we're talking about real speakers I assume

[blybo]

I'm probably receiving my horn loaded open baffle speakers tomorrow which I'm hoping will give the best of both worlds... we shall see.

 

The designer says having an 80 degree dispersion helps to alleviate room issues, and I did really like the sound of Maggies in my room, just wanted more punch. 

[svenr]

12 hours ago, Keith_W said:

My questions: 

- has this also been your experience? 

- does anybody have a psychoacoustic explanation as to why omnidirectional speakers seem to sound more spacious? 

- what measurement correlates with soundstage depth, anyway? 

 

There exists a lot of research into spacious perception.  A good starting point is Floyd Toole's "Sound Reproduction".  Apparent stage width (ASW), interaural cross-correlation (IACC) or lateral fraction of energy (LFE) are just some of the physical measures that indirectly relate to space perception.

[davewantsmoore]

19 hours ago, Keith_W said:

Before I start - i'll say this. I am asking a question, not making an assertion. My question is: what influence does the speaker dispersion pattern have on perceived soundstage depth? 

 

It isn't possible to just generalise this by dispersion pattern alone....  depends also on the room size, and listener position.

 

Thing which affect imaging:

 

Reflected sound which has a different spectral balance to the direct sound

Level of reflected sound (ratio to direct)

Delay of reflected sound

 

Changing the speaker dispersion pattern (without changing the room size or listener location) affects all these things.

[Guest Doppler Effect]

I'm away and very much in the middle of nowhere at the moment, but just a couple of quick comments.

 

Firstly in regard to perception of reflected or delayed sound.  Sounds reaching the ear within the first 5 to 6 ms are indistinguishable and therefore perceived as a single event, potentially with the 'smear' you are speaking of.  From 5 to 6 ms up to 40 ms the reflected sound is perceived as spatial information.  Above that is echo.

 

if you think of an average dispersion monopole speaker positioned not so distant from a side wall, draw a line from the speaker to first reflection point along the side wall and to the listener.  Now do it for a dipole speaker positioned a meter from the rear wall, only for the dipole speaker, with its rearward sound and null to the sides will require you to draw this line from the speaker to the rear wall and then to the listener.  Which line is longer?  

 

There is more to it than that, and spatial perception does include directionality of the sound, but you get the idea.

 

oh, and your calculations are out.  You calculated the 1 m distance from the speaker to the rear wall, but forgot that it must travel at least this distance back again to be in line with its original starting point, which is what it is delayed from.  That makes it a 2 m round trip. 

[davewantsmoore]

20 hours ago, Keith_W said:

My questions: 

 

- has this also been your experience? 

- does anybody have a psychoacoustic explanation as to why omnidirectional speakers seem to sound more spacious? 

- what measurement correlates with soundstage depth, anyway? 

 

Somewhat.    Such a well produced illusion of layering and depth is rare.     Most speakers have some sort of error in them which prevents it.

 

... but I think you are thinking about the concept "backward".    Rather than trying to identify "what creates soundstage layering".     Consider that the soundstage "layering" is simply an illusion, which is encoded into the input signal.     Only reproducing the input signal well, can recreate the illusion.     So what you are looking to identify, is what do the speaker which have trouble, do that harms the reproduction of the input signal  (and thus the illusion).    ie.   it sound sound "spacious" because it lets sound go to the rear.....   Rather it doesn't sound "spacious" because it does <something bad>.

 

 

Dipoles are narrow dispersion, and when well designed can have low diffraction, and without errors in their dispersion pattern.    Completely omni-directional speaker (rare) are also typically free of diffraction, or changing dispersion with frequency.

 

Box speakers can easily suffer from diffraction, and can also suffer from drastically changing dispersion with frequency.   (some of the examples you list, are probably fairly free from both though).

 

Horns in a small room can easily fall foul of these type of problems too.   Depends on the design (they come in all shapes and sizes)

Edited September 26, 2016 by davewantsmoore

[davewantsmoore]

8 hours ago, blybo said:

I'm probably receiving my horn loaded open baffle speakers tomorrow which I'm hoping will give the best of both worlds.

 

 

An open-baffle is a good way to limit the coverage angle at low-mid frequencies .....   but ideally you want the sound radiating backwards to be an exact copy of the sound radiating forwards.    ie. you would "need" a rear facing horn....  otherwise the delayed sound from the open-baffle component, will be missing all the high frequencies (which were fired form the horn, but only forwards).

 

BUT.   Most speakers have large compromises somewhere .... and if they're the speaker I"m thinking of, they are pretty good :thumbup:

 

[blybo]

 

 

An open-baffle is a good way to limit the coverage angle at low-mid frequencies .....   but ideally you want the sound radiating backwards to be an exact copy of the sound radiating forwards.    ie. you would "need" a rear facing horn....  otherwise the delayed sound from the open-baffle component, will be missing all the high frequencies (which were fired form the horn, but only forwards).

 

BUT.   Most speakers have large compromises somewhere .... and if they're the speaker I"m thinking of, they are pretty good :thumbup:

 

Spatial Hologram M3 Turbo S and TNT tracking says "delivering" so looks like I picked a good day to look after the kids


Sent from my iPhone using Tapatalk

[davewantsmoore]

Yep, I thought it might be a Spatial. 

 

What an ace name, heh.

 

"Spatial Hologram M3 Turbo S"

[eltech]

23 hours ago, Keith_W said:

does anybody have a psychoacoustic explanation as to why omnidirectional speakers seem to sound more spacious? 

 

I think the logical reason for this is that for example if someone plays an instrument in a room, the sound basically radiates omnidirectionally (apart from a slight reduction in output where the person playing the instrument slightly blocks the sound) Like behind a guitarist has less output than in front. So a single omnidirectional speaker in a room more closely represents a single person in a room playing an instrument. Two omnidirectional speakers in a room (stereo) is not natural because then you have two omnidirectional sources of sound which is not natural, particularly if the music is of a solo instrument.

23 hours ago, Keith_W said:

Nearly every speaker I have heard that projects most of its high frequencies forward tends to have a soundstage that is at the same plane or forward of the speaker

 

Yes fair comment.

but is that a problem?

What do you want your speakers to do?

 

23 hours ago, Keith_W said:

- what measurement correlates with soundstage depth, anyway? 

I dont know about measurements, but will take a guess and say that it comes down to the quality of the recording, and the venue it is recorded at, and the microphone technique.

Generally with regard to a classical recording, I find that if the recording is good, and there is some natural reverberation on the stage, and the microphones capture that reverberation it allows our ears to place the instruments at some point in space.  Without the reverberation there is little to guide the brain to localise the instrument.  With really good recordings I've been able to position all of the instruments on stage. With bad classical recordings I can't make out the position of any instrument.

http://www.dpamicrophones.com/mic-university/classical-orchestra-a-b-stereo

Edited September 26, 2016 by eltech

[Newman]

On 9/25/2016 at 11:51 PM, Keith_W said:

Before I start - i'll say this. I am asking a question, not making an assertion. My question is: what influence does the speaker dispersion pattern have on perceived soundstage depth? 

 

 

First, a question of my own: are you talking about the sense of the size/depth of the venue/performance space? Or do you mean the sense that the performers (sonic sources) are at different distances from the listener?

 

Quote

 

My questions: 

 

- has this also been your experience? 

 

 

 

Sporadically and inconsistently -- so I think that means no, I don't experience a consistent pattern like you describe with loudspeaker type.
 

Quote

 

- does anybody have a psychoacoustic explanation as to why omnidirectional speakers seem to sound more spacious? 

 

Well, yes, the 'elephant in the room' psychoacoustic explanation that we really should stop overlooking: confirmation bias. In this case, confirmation of the generic expectation that a speaker that directs sound waves with more depth creates a result with more depth; and a speaker with a tiny front baffle that the listener thinks (rightly or not) 'allows' HF sound to play in more directions also creates a result with more apparent depth.

 

Quote

 

- what measurement correlates with soundstage depth, anyway? 

 

 

Delay. That is what a playback sound field generator uses to differentiate perceived distance in any of the three dimensions. There is nothing special about depth that has a unique measure.

[THOMO]

A large single omni -direcctional  speaker playing mono recordings could be very interesting to hear.

[eltech]

I think so.

[Keith_W]

svenr, the IACC in my system is 84.9%. Very good by most standards. 

 

KARRI, I did take the reflection time into account. 3ms for 1m, therefore 6ms for 2m. 

 

Sorry, I would like to type a more detailed reply, but don't have time today. Maybe later tonight or tomorrow. Thanks for all the replies everyone.

[Guest Doppler Effect]

3 hours ago, Keith_W said:

KARRI, I did take the reflection time into account. 3ms for 1m, therefore 6ms for 2m. 

 

 

Quite right!

 

My apologies, I was on my phone in the middle of the NT and read your post as 3ms being short of the 6ms.  

 

Around 5 to 6 ms and up is perceived 'spatially'.  Under 5 is perceived as a singular sound.  At 1 metre you are adequately differentiating them.  

[svenr]

@Keith_W - doesn't IACC = 1.0 mean the left / right ear signals recorded are identical?  A perfectly diffuse SF would give IACC = 0 I thought?

 

In any case, Ando's "Concert Hall Acoustics" gives a well-written introduction to the subject.  He's got a lot to say about localisation of sources located on the median plane.  And it is the spatial reproduction of the median plane that determines the stage depth perception.

[Keith_W]

svenr, Acourate reports IACC as 0-100%. 100% means that the left and right speaker are perfectly identical. Prior to correction, it is already very good - 79%. But after correction, it goes up to 84.9%. 

[Nada]

With that great looking design of the  Spatial Hologram M3 Turbo S speakers I wonder how does the dual concentric system  work from the point of view of

 

a. The construction. Do dual concentric's work with independent mounting of the two drivers or does the tweeter ride on the bigger cone? With the  Hologram is the inner compression driver mounted so the 15" driver slides over it? I imagine it would be deleterious to add any weight to the 15" driver itself?

 

Below is how Emerald Physics CS 2.3 do it:

 

 

 

 

 

2. Dipole effect - the upper range might only project forward so what happens to the touted dipole depth effect?

 

Id really like to hear this speaker. It looks fabulous

 

 

Edited September 27, 2016 by Nada

[eltech]

1 hour ago, Nada said:

the upper range might only project forward

 

From the Hologram website " Driver compliment:  Two 12 inch mid/woofers, one wide bandwidth compression driver "

 

The compression driver will send all the sound forwards.

Edited September 27, 2016 by eltech

[blybo]

1 hour ago, Nada said:

With that great looking design of the  Spatial Hologram M3 Turbo S speakers I wonder how does the dual concentric system  work from the point of view of

 

a. The construction. Do dual concentric's work with independent mounting of the two drivers or does the tweeter ride on the bigger cone? With the  Hologram is the inner compression driver mounted so the 15" driver slides over it? I imagine it would be deleterious to add any weight to the 15" driver itself?



 

Below is how Emerald Physics CS 2.3 do it:

 

 



 



 



 

2. Dipole effect - the upper range might only project forward so what happens to the touted dipole depth effect?

 

Id really like to hear this speaker. It looks fabulous

 

 
 

I got mine yesterday as quoted in my showcase thread.



They work in the same way as the Emerald Physics, no wonder, as same designer. The compression driver does not touch the bass driver but sits in the throat of the big sucker. The M3 uses dual 15" drivers, the M4 dual 12".

 

Edited September 27, 2016 by blybo

[eltech]

On 26/09/2016 at 0:21 AM, Keith_W said:

Nearly every speaker I have heard that projects most of its high frequencies forward tends to have a soundstage that is at the same plane or forward of the speaker

 

Are you saying you have never experienced the illusion of sound depth in a classical recording? I have certainly percieved sound coming from behind the plane of the speakers with good classical recordings. One of the most fun things about a good recording is its ability to do that.

[AudioGeek]

55 minutes ago, eltech said:

 

Are you saying you have never experienced the illusion of sound depth in a classical recording? I have certainly percieved sound coming from behind the plane of the speakers with good classical recordings. One of the most fun things about a good recording is its ability to do that.

My previous Nakamich dragons do this very well. Also great width and depth. They have a reflecting column at the back which does this.

[Keith_W]

1 hour ago, eltech said:

 

Are you saying you have never experienced the illusion of sound depth in a classical recording? I have certainly percieved sound coming from behind the plane of the speakers with good classical recordings. One of the most fun things about a good recording is its ability to do that.

 

I didn't say that. What I said is that some speakers are better at conjuring this illusion than others.