Can we measure:
-Sound stage depth
-Sound stage width
-Harshness
I think most/many people who have listened to speakers will have observed differences in the above characteristics between speaker systems... can they be measured though?
The correlations to measurements are there, but not as black/white or 0/1 results, but rather as statistical probabilities with sometimes large individual scatter and interaction of several causes for an observed effect.
a) Sound stage width
The perceived stage width is related to the horizontal directivity and radiation (and the listening room, respectively early reflections) of the loudspeaker - but not only, also the low bass plays a role.
This is easily measurable, the problem is the interpretation of the measurements, especially if the loudspeakers measure similarly.
b) Sound stage depth
The vertical radiation of the loudspeaker should play a role here (especially the ceiling reflection) but not independent of other influences (horizontal radiation, room reflections).
The
directional bands introduced by J. Blauert can explain effects such as perceived stage depth and height. But also here large individual differences can occur.
Blauert [1] demonstrated that the direction of a sound image for a 1/3 octave band noise is a
function of the center frequency only and does not depend on the source elevation angle. He
referred to the frequency band by which the direction of sound image is determined as the
directional band.
Sources:
Individual differences in directional bands
Sound_Localization_in_the_Median_Plane
Spatial Perception
For stereo listening and further simplification, you can "mix" vertical and horizontal listening:
With
loudspeaker stereophony (i.e. in the
horizontal plane) the original direction-determining bands v (front), h (back) and o (top) of the median plane can be easily reinterpreted as two directionless hearing sensations “
present ” and “
diffuse ”. This is also noteworthy for the
equalizer settings for the sound processing in stereophony and the
surround sound of broadband signals. Where:
- front v = present in the sound , close, direct, superficial - this can be achieved by increasing the frequencies 300 to 400 Hz and 3 to 4 kHz as well as lowering frequencies by 1 kHz.
- back h (and above o) = diffuse , distant and spatial in the sound - this can be achieved by increasing the frequencies by 1 kHz
Source:
directional bands introduced by J. Blauert
Again, the measurements are less the problem than the complex interplay of different effects and their interpretation.
c) Harshness
Harshness in loudspeakers can have many different causes (mechanical causes, distortion, radiation errors, tuning errors).
The most common cause of harshness I have found in loudspeaker tuning is the critical frequency range around the ear canal resonance (2.2-3kHz, 2.7kHz is often cited as the average). Even small humps in the sound power (or horiz. ER, PIR) in this frequency range very often leads to the sound being perceived as harsh, especially if the cause is the horizontal radiation of the speaker.
But if this frequency range is cut back too much, sound coloration occurs and it affects the perceived stage depth (e.g. "voices" are perceived as less present, see above) - it's a balancing act.
I keep repeating myself
IMO, it is not so much the measurements that are a problem, but the complexity and mass of the measurement data and their interpretation
.