So Change to horns or stay with cones?
I guess this is the right answer not in my words:
Your argument about home audio listening is valid if you listen at fairly low average spl levels of lets say, around 75db. Because for a good sense of dynamics a loudspeaker must be able to easily handle 20db peaks above the average level with no compression. That is the minimum requirement and unfortunately the Scan-dinavian favorites (the drivers used in most high end dynamic speakers) fall short of this target at anything above mid 90's db. So if you want to listen at realistic average levels of say 95 -100 db you will need the speaker to handle a not-so-unrealistic 120db peak levels and that's at the listening position no less.
For audiophile drivers core size and voice coil size are of secondary importance. Contrast this with the pro drivers' big voice coils and oversized and vented magnetic cores that can sustain prolonged periods of abusive power (read 400-600 watts) with just maybe 2db of thermal compression while playing at around 120db average level. There really is no comparison! Thermal compression is real and one of the most important as well as overlooked parameters in loudspeaker performance. It's perhaps not surprising that this is the case for the audiophile speakers as the driver core is the most expensive part of the assembly and the designers using the same logic consciously chose this set of compromises.
But make no mistake, as excessive as these db levels might seem at first, for the person that wants realistic reproduction at the home this is what will be required of the system. (Manga)
Is there a need for an audio system to produce 120 db peak?
Symphony orchestra is playing one flute, but in the next second the orchestra barks with the whole power.
Most of the brass instruments can alone produce 120 db. There are 120 or so different instruments in the orchestra.
You are sitting at row 20, a flute is about 40-45 db. The orchestra BARKS. It is about 120 db at the row 20.
The difference is 80 db.
CD can record 90 db of a difference (called dynamic range). Recording engineer has to compress the sound. Some engineers can hide compression better though, but all sounds are too BIG to fit on CD.
No matter what it is, Jazz you name it. I bet to record a girl with the guitar some 6 db of compression is still needed. (Yurmac)
In my opinion the single most important benefit of a good horn is not increased dynamic contrast, but improved radiation pattern control (though it's nice to have both). The radiation pattern of most loudspeakers narrows and blooms and narrows again very significantly across the spectrum. The result is that the reverberant energy - mostly composed of off-axis radiation - has a different tonal balance from the on-axis sound. Since the ear/brain system is constantly analyzing incoming sounds as either first-arrivals or reflections, and using spectral constant to do so, a large discrepancy in the spectral balance of the first-arrival and reverberant sound makes correct classification more difficult for the ear/brain system; in effect, CPU usage goes up. Often the result over a half-hour or so is listening fatigue - literally, a head-ache because the ear/brain system having to work harder to correctly classify the reverberant energy whose spectral balance is unnatural.
But, don't get the idea that reflections are bad - early ones often are, but late-arriving ones are usually beneficial. A dense, late-arriving, highly diffuse, slowly decaying, spectrally correct reverberant field is what makes a good concert or recital hall sound so delicious.
Horns don't ordinarily give a more diffuse reverberant field than direct-radiator dynamic speakers. If anything, their typically narrower pattern results in a less-diffuse reverberant field; but that narrower pattern often makes it easier to "aim" the horns to minimize early-arrival reflections.
You see, reflections arriving before 10 milliseconds (corresponding to a path length of about 11 feet) are usually detrimental, whereas reflections arriving later than that are usually beneficial, assuming good spectral balance. (Audiokinesis)
Wonderfully put by this fellow audiogoners....
