Good speakers less efficient?

dipolar radiators have efficiencies in the 85 to 88db range.

i happen to prefer this type of speaker. depending upon the room and the impedance curve, they can pose a challenge fot a tube amplifier.

if one prefers this design to other designs... well, draw your own conclusion.

Sharorne,

My experience with tonal balance shifting as the volume level changes has generally been the opposite of yours, though I don't doubt your observation. My experience with high efficiency speakers has mostly been with horn systems (though usually with a direct-radiator prosound-type woofer).

I have generally attributed tonal balance shift to differing power compression characteristics of the different drivers. Relatively few drivers give a full 3 dB increase in loudness for a doubling of input power - usually it's more like 2.6-2.8 dB. If the woofer compresses more than the tweeter, then the tonal balance will shift towards the tweeter's end of the spectrum at high volume levels. If the tweeter compresses more, then the tonal balance will shift towards the woofer's end of the spectrum at high volume levels.

I recall a couple of years ago listening to a very well known and generally well respected three-way, which sounded recessed and distant at low volume levels, well balanced at medium to borderline high volume levels, and decidedly forward at high volume levels. I think that the midrange (which was an especially high-tech unit) had less power compression than the woofer and tweeter, so the speaker had been "voiced" to sound balanced in the medium to borderline-high volume range. At higher volume levels the midrange left the woofer and tweeter in the dust, and at lower volume levels it withdrew into the background (okay I'm exaggerating a bit with my imagery here, but to my ears that was the trend).

Getting back to your observation, this is just speculation on my part but it's possible that very high efficiency conventional tweeters use ultrathin voice coil wire that heats up quite easily and rapidly, so that in practice such tweeters have more thermal compression than their medium efficiency cousins.

There can be other mechanisms at play, such as the ear's level-dependent perception of certain types of distortion, that may be dominant in some situations. This would tend to shift the perceived tonal balance upward as the volume level increased.

I don't think I've given an satisfactory answer (at least I'm not satisfied with it). I think there's quite a bit more to be known on the subject than what I now know.

Duke

That's a good point Bartok.

Most mid-range to large models from manufacturers such as Merlin, Altec, Klipsch, Zu, Avantgarde, Cain and Cain, Silverline, Acoustic Zen, Gallo, Wharfdale, Coincident, Tannoy, Fostex, Oris, Duevel, Green Mountain Audio, Maxxhorn, Devore, etc. really do suck most amps dry.

Thanks.

Duke,

Thanks, I am sure you are right about the causes or shifting tonal balance (from different driver sensitivities). My experience is probably not representative as it does not involve more than a few dozen designs.

However there is another angle. The Fletcher-Munson equal loudness curves. These also have impact on the perception of sound at various levels;

130 db at 20Hz sounds as loud as 100 db at 2 Khz.
110 db at 20Hz sounds as loud as 60 db at 2 Khz.

Notice that this implies that our ears compress the mid range relative to the bass as you go to higher SPL's (a 20 db increase in ultra LF requires a massive 40 db increase in mid range to maintain equal terms of loudness perception)

I guess this explains why rock mixes that are designed to play loud are usually mixed thin in the bass. It implies that to maintain consistency to our ears at various sound levels then the bass should be compressed as you increase SPL in order not to have too much shift in the balance.

Does this imply that relative compression between drivers might actually be desirable in a speaker?

Does this imply that there is fundamentally a SPL sweetspot for particular forms of music that is independent of the speaker used?

Thanks for your insights.

Shadorne says.... "Does this imply that there is fundamentally a SPL sweetspot for particular forms of music that is independent of the speaker used?"

Of course. Just yesterday I listened to my CD of Kings College Chior Christmas Carols, and with all those voices, and the organ thundering away, I got to thinking about why I have six large subwoofer drivers. No way that music would sound right coming out of a 6 inch bookshelf speaker. On the other hand, I have violin recordings that must the played at a much lower SPL. A real violin can play only so loud, and if you crank it up louder than that it sounds lousy.