What is the ideal directivity pattern for stereo speakers?

[keyser]

What is the ideal directivity pattern for stereo speakers?

I think most industry professionals would agree about many criteria that constitute a good or maybe 'perfect' loudspeaker. On another forum speaker dave (who is also a diyaudio member) posted the following list (for the complete list, click on the link):

The Goals for an "Ideal Loudspeaker" - The Classic Speaker Pages Discussion Forums - Page 8

An Ideal speaker has:

1) Very flat on-axis response
2) Very smooth on axis response
3) Very flat and smooth response through any likely listening window
4) Smooth and resonance free power response, but of no particular curve
5) Holes in the power response are acceptable but peaks are not
6) Generally rising directivity (non flat d.i.)
7) Generally wide dispersion
8) Wide bandwidth with a -10dB cutoff below 35 Hz
9) Interfaces well with the room, gives a smooth in-room curve below 200Hz
10) Adequately low distortion. Low AM distortion with high woofer excursion

Most criteria will probably be on the list of most other pro's as well, but I think not everybody would agree about point 7. Opinions seem to vary quite a lot when it comes to what is the ideal directivity.

What would the dispersion pattern of a theoretically ideal speaker look like?

Please, psycho-acoustics first! The practical implementation can be saved for later.

[Saturnus]

In principle there are 2 ideal directivity patterns: mono-directional and omni-directional. All other states are involves a acceptable level of compromise.

Mono-directional is not achievable by loudspeakers as such but are with headphones where your ears will only recieve the direct radiation of the loudspeakers.

Omni-directional speakers do exist but most are poor implementations that doesn't deliver an acceptable level of sound quality.

Bi-polar and to a lesser extent di-polar speakers delivers an directivity pattern in the far field that closely resemble a true omni-directional speaker, and that's the reason behind the DIY popularity of di-polar speakers. Strangely, bi-polar speakers aren't nearly as popular although it has some very significant advantages over di-polar speakers but that is probably due to lack of knowledge to some extent, and that it incurs twice the cost.

[LineArray]

I think especially for 4) and 7) it is a bit difficult to choose independently
from the listening room.

Having a "good" listening room with high diffusivity, relatively low
decay time, no spectral anomalies in decay time (say smoothly falling
with frequency or constant above say 300Hz) a speaker designed
according to that criteria will surely perform very well, if listening
distance is not too far.

When i was visiting an audio fair some years ago using empty
hotel rooms for presentation the "Amphion" models from Finland
presented in that year had a clear advantage over the "wide"
but also often "discontinouus directivity" conventional multiway
speakers presented under similar conditions.

Amphion uses a resistance box for the midrange in some of their Models,
achieving a cardioid like radiaton pattern, and waveguides for the tweeter
to match that pattern to some extent at the crossover frequency.

But that pattern could still be considered as "fairly" wide, as there seems
to be no big change for small off axis angles ... as long as you have that and
tonality is consistent even for somewhat larger angles even such
a pattern may work well, and surely better under poor conditions.

I was referring to horizontal dispersion, if the vertical dispersion
can be made fairly consistent it may be considerably narrower than
the horizontal dispersion without any disadvantage IMO.


Kind Regards

[boris81]

Talking about the dispersion pattern might still be too implementation specific. We might discover that "one CD pattern fits all" doesn't necessarily exist.

I think that first and foremost we need to map out the arrival of the ideal signal to the listener. I've heard that early reflections are good and that late reflections are bad but the exact time range, intensity and frequency are subject of speculations. I'm not even convinced that we need to know what is ideal, just what is not desired.

I would emphasize that once we have a solid understanding of the requirements, the technical details of the delivery method will be more or less trivial.

[MisterTwister]

Does directivity pattern really matter in home environment?

[LineArray]

Hi MisterTwister,

i would say especially in home environment the radiation pattern
matters, just because home environment can be expected to
be even less uniform in speaker-listener placement and reverberant
conditions than e.g. control rooms ...

[gainphile]

In a home environment the goal is to have delayed reflections as replicas of direct sound. Hence these would qualify:

- Omni-directional
- Dipole
- Cardioid

Considering that the lower the frequency, the less sensitive the perception is as to direct vs. reverberant sound, a speaker with narrow directivity (aka waveguides) may qualify too.

Each would have their own limitations and challenges, ie. there is no perfect dipole speakers or perfect Omni speakers.

[MisterTwister]

Linearray, then how would you explain positive reviews of all kinds of speakers, from omni to horn radiation patterns. It appears to be a matter of taste, not something that has correct way of doing. How many people can tell a difference between omni and waveguide speakers in a blind test?
remember this long thread:
Linkwitz Orions beaten by Behringer.... what!!?
speakers had different radiation patterns, none of the speakers had decisive victory.

p.s. I understand importance of directivity in PA applications( night clubs, stadiums etc.)

[kevinkr]

My recipe these days is for mid and hf horns with tightly controlled directivity oriented relatively on axis at my listening position. (60 degrees or less of dispersion in the horizontal plane, and 40 - 60 degrees in the vertical plane.. This works well in my room - good image and good uniform measured frequency response at my listening position. Probably not optimum for larger rooms or where a very wide listening area is desired, but it works for me..

Everyone needs to find their own path of course, for a long time I was quite enthralled with the MG1.6QRs I owned - these days and in this room, unlikely...

Good subject..

[GM]

Originally Posted by keyser

What would the dispersion pattern of a theoretically ideal speaker look like?

No early reflections, i.e. none off any boundary until behind the listening position, ergo would be different in virtually every multi-use room.

GM