Genuinely curious (given the reputation for natural sounding timbre of some more boxy speakers). What specfically makes the timbre..better?
Please understand that I'm trying to extrapolate from the little that I know, so there may certainly be a number of inaccuracies or mischaracterizations, which I would certainly appreciate having corrected.
In discussing timbre here, I'm referring to the subjective "naturalness" of how non-closely-mic'd vocals and instruments sound when reproduced through loudspeakers. We can compare our memory, faulty though it may be, of actual listening experiences at concerts or recitals. Since we rarely put our ears where the microphones would be when dealing with closely mic'd vocals, many of us are unable to compare the experience of listening to a actual performing singer from fewer than a few feet away, and it may be more difficult for us to compare how well these are reproduced. Musical instruments can sound quite different to the performer and an audience member, so closely mic'd instrumental recordings may better replicate what a performer might hear themselves, but I think that's a different topic of discussion.
The frequency range of the human voice might be crudely approximated as 100-1000 Hz. The violin is a little higher, with much of the range being ~200-3.5 kHz. Perhaps there is a relatively critical range from a few hundred to a few thousand hertz where the ear is relatively sensitive in general, or more specifically to changes in directivity.
If you look at the Kef LS50 Meta (
https://www.audiosciencereview.com/forum/index.php?threads/kef-ls50-meta-review-speaker.25574/), Kef R3 (
https://www.audiosciencereview.com/forum/index.php?threads/kef-r3-speaker-review.12021/), or Genelec 8361A (
https://www.audiosciencereview.com/...s/genelec-8361a-review-powered-monitor.28039/), these all have steadily rising Sound Power DI from 100 to 1000 Hz, although they're nicely smooth (what I think of as "controlled" directivity). As you'd expect, the sound power increasingly diverges from the on-axis response (compare the relative difference at, say, 200-300 Hz with 800-1000 Hz).
Compare that with the JBL LRS6332 (
https://jblpro.com/en-US/site_elements/lsr6332-spec-sheet) or perhaps even the Revel F328Be (
https://www.audiosciencereview.com/forum/index.php?threads/revel-f328be-speaker-review.17443/), where there is less relative divergence between the on-axis and power responses in this range, and the SP DI is relatively flatter, implying relatively more "constant" directivity. I wasn't able to think of other wider baffle speakers with Spin-o-rama measurements. The JBL M2 (
https://www.erinsaudiocorner.com/loudspeakers/jbl_m2/) isn't a great example in many ways, since it's crossed over so low, hence the step in directivity approaching the 800 Hz crossover. The D&D (8C
https://www.erinsaudiocorner.com/loudspeakers/dutch_dutch_8c/) also has the rise in SP DI at the top end of this range as one approaches the crossover frequency, it does extends further down at the bottom of this range. If you look at
https://www.audiosciencereview.com/...cal-music-listening.33212/page-2#post-1162572, the reverberation times for the two concert halls from 125-1000 Hz and the red line in figure 3 of the Lokki paper are all reasonably flat in this frequency range.
These speakers have a second range from 1-2 kHz up to about 6-8 kHz where the SP DI is relatively flatter again, then DI rises again, and SP falls off above that frequency range. Many classic concert halls have relatively little response above 6-8 kHz or so (again:
https://www.audiosciencereview.com/...cal-music-listening.33212/page-2#post-1162572).
The difference between smoothly rising and relatively flatter directivities may not be significant in a relatively anechoic room, but in a more reflective listening environment, the sound power might be expected to play an increasingly important role.
@Floyd Toole has commented multiple times on the importance of preserving the relative spectral timbre of reflections (I hope I'm not misquoting), hence Harman's apparent embrace of wider-radiating conventional speakers for consumer use. Harman and Klippel take some reasonable shortcuts with things like their estimated in-room response curve, but consider the Floor Bounce curve in Klippel's measurements--floor bounce is dependent on multiple factors, specifically height of transducer and listener and relative distance between (
https://mehlau.net/audio/floorbounce/ and
http://dtmblabber.blogspot.com/2010/12/little-more-into-boundary-conditions.html). Toole, in his books and articles, has written things like "“The shape of the room curve is clearly signaled in the shapes of both the “early-reflections” curve and the inverted DI." Maybe the flatter room curve in the critical range may sound more "natural" in terms of timbre to certain listeners.
To sum up, perhaps it's because wider baffle speakers push up the frequency at which sound power rolls off to the upper end of the "critical range" that they may sound more natural in more reflective listening rooms to some listeners of classical music, and perhaps this derives from the relative frequency ranges of these instruments, as well as the relative frequency responses of the venues in which concert goers are accustomed to hear them.
Hope that makes sense. I wrote rather more than I originally intended, but that's pretty much all that I know.
Young-Ho