I've come to believe that the tone of a loudspeaker in a normally reverberant space is disproportionately determined by its anechoic frequency response. By tone I mean the perception of the frequency response or transfer function -- the thing that you can affect with an equalizer. By anechoic frequency response I mean the free-field pressure response measured within the nominal coverage, as opposed to the integrated power response, as you would measure in a reverberation chamber.
This is equivalent to saying that the frequency response of the direct sound matters most, and the reflections, beyond some time interval, not so much.
I got spanked on the other forum for making this assertion and that got me to question why I even think this or what I'm confusing here.
McCarthy, in "Sound Systems: Design and Optimization", divides the perceptual effect of reflections into tonal, spatial, and echo, depending on their arrival time relative to the direct sound. He draws the line between tonal and spatial perception at the arrival time that corresponds to a spacing of the zeros of the resulting comb filter that is equal to the critical bandwidth (CB), with earlier arrivals being perceived as a change in tone.
This is a provocative observation. My understanding of CB is that it concerns our inability to discern pure tones as distinct when they are close in frequency, and is a result of the neuro-physics of the basilar membrane. This is not exactly the same as perceiving a change in timbre of complex tones only when the spacing of notches in the response exceeds the CB, but does not seem inconsistent either.
Let's take the foregoing as fact. At low frequencies, say below 500Hz, CB is generally accepted to be about 100Hz. At higher frequencies it is 1/6 to 1/3 octave (McCarthy uses 1/6), but, in any case, greater than 100Hz. So a strong reflection arriving 10ms after the direct sound (other than causing a suckout -- I love that word -- at 50Hz) straddles the fence between tone and space and we would expect a reflection at 20ms or 30ms to be firmly in the space camp. 20ms is the round-trip time for the reflection off a surface only 11 ft away.
So it would seem that the anechoic response of the loudspeaker is, at the very least, a good starting point in any venue larger than a control room. And this is not to say that one should never "equalize the room", say, in cases where there is an annoyingly long decay in a certain range or to compensate for boundary effects at low frequencies or path loss at high frequencies.
I'd like to know how people think about all this and would love some references on the subject (especially peer reviewed). I also wonder how this relates, if at all, to the observation that narrow peaks in the frequency response are highly audible, while narrow notches are not.
--Frank
This is equivalent to saying that the frequency response of the direct sound matters most, and the reflections, beyond some time interval, not so much.
I got spanked on the other forum for making this assertion and that got me to question why I even think this or what I'm confusing here.
McCarthy, in "Sound Systems: Design and Optimization", divides the perceptual effect of reflections into tonal, spatial, and echo, depending on their arrival time relative to the direct sound. He draws the line between tonal and spatial perception at the arrival time that corresponds to a spacing of the zeros of the resulting comb filter that is equal to the critical bandwidth (CB), with earlier arrivals being perceived as a change in tone.
This is a provocative observation. My understanding of CB is that it concerns our inability to discern pure tones as distinct when they are close in frequency, and is a result of the neuro-physics of the basilar membrane. This is not exactly the same as perceiving a change in timbre of complex tones only when the spacing of notches in the response exceeds the CB, but does not seem inconsistent either.
Let's take the foregoing as fact. At low frequencies, say below 500Hz, CB is generally accepted to be about 100Hz. At higher frequencies it is 1/6 to 1/3 octave (McCarthy uses 1/6), but, in any case, greater than 100Hz. So a strong reflection arriving 10ms after the direct sound (other than causing a suckout -- I love that word -- at 50Hz) straddles the fence between tone and space and we would expect a reflection at 20ms or 30ms to be firmly in the space camp. 20ms is the round-trip time for the reflection off a surface only 11 ft away.
So it would seem that the anechoic response of the loudspeaker is, at the very least, a good starting point in any venue larger than a control room. And this is not to say that one should never "equalize the room", say, in cases where there is an annoyingly long decay in a certain range or to compensate for boundary effects at low frequencies or path loss at high frequencies.
I'd like to know how people think about all this and would love some references on the subject (especially peer reviewed). I also wonder how this relates, if at all, to the observation that narrow peaks in the frequency response are highly audible, while narrow notches are not.
--Frank
Last edited:
).