Thiel CS2.3 Loudspeaker Measurements part 2

In his excellent Testing Loudspeakers book, Joe D'Appolito examines the effects of measuring microphone distance on drive-unit integration with multiway speakers (p.72). One conclusion he draws is that at close microphone distances, speakers with odd-order crossovers—first-order, third-order—can appear to have notches in their on-axis response due to the 90 degree phase shift from the crossover filters. As the microphone distance is increased, these notches disappear. Of course, as I point out in the concluding article of my three-part "Measuring Loudspeakers" series elsewhere in this issue, with a time-windowed system like the DRA Labs MLSSA I use to make my measurements, the farther away you place the microphone, the lower the resolution, due to the earlier arrival of reflections from the room boundaries. The 50" mike distance I use is a compromise between the need for correct drive-unit integration and the opposed need for midrange resolution in the resulting graph. But it is possible that the lack of presence-region energy in fig.3 is actually due to the 50" mike distance and is not real, in that it disappears at the farther distances at which a speaker like the Thiel will be listened to.

Accordingly, I repeated the measurement at an 80" mike distance, about as far away as I can get in Stereophile's new measurement/listening room. The results are shown in fig.4. The top trace is the 50" measurement; the bottom trace is the one taken at 80". Yes, the lack of energy starts to fill in. But note the much larger effect of microphone distance on the absolute level of the tweeter. At the farther distance, the treble is better matched to the speaker's midrange. I have not experienced this kind of inverse proximity effect before—usually, the bass starts to tilt up as the microphone gets closer to the speaker. Perhaps this is a function of the CS2.3's unique coaxial, mechanically crossed-over midrange/HF unit. Whatever, fig.4 suggests that, at a normal listening distance in a room, the CS2.3's overall balance will be much flatter than is implied by fig.3. (When I visited the Thiel facility a couple of years back, I noted that Jim Thiel does measure his speakers much farther away than I can.)

Fig.4 Thiel CS2.3, 1/3-octave-smoothed anechoic response on tweeter axis at 50" (top) and 80" (bottom).

In a room, a speaker's perceived balanced will also be affected by its off-axis behavior. Fig.5 shows the changes in the Thiel's response as the microphone is moved at 5° intervals to the side. The high-frequency dispersion is excellent, correlating with the accurate, stable imaging. And in the presence region there is more energy to the speaker's sides, which will ameliorate the effect of the on-axis suckout. But note that an off-axis depression develops in the lower crossover region.

Fig.5 Thiel CS2.3, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90 degrees-5 degrees off-axis; reference response; differences in response 5 degrees-90 degrees off-axis.

I note that BD actually listened with his ears several inches above the tweeter axis, which is a low 32" from the floor. Fig.6 shows the actual responses of the CS2.3, plotted from 10° below the HF axis (front of graph) to 20° above it. The flattest response overall is actually the lowest, at just 20" from the floor. Moving above the tweeter axis at this 50" microphone distance fills in the presence-region suckout, but introduces another suckout at the lower crossover frequency. The effect of this in a room will be hard to predict; lobing in this frequency region is hard to perceive in-room, and the speaker's total power output is still smooth through this region.

Fig.6 Thiel CS2.3, vertical response family at 50", from back to front: responses 20 degrees-5 degrees above HF axis; reference response; responses 5 degrees-10 degrees below HF axis.

1026 Nandino Boulevard
Lexington, KY 40511
(606) 254-9427