|
This article will go into the
complex interaction between tweeter dispersion patterns, dipole radiation
patterns, crossovers and driver positions and how these affect what you hear.
This will be done by comparing real world measured frequency and time domain
data from the Ronin as well as a test sample that is representative of many two
and three way speakers. All of this matters for one reason: we listen to our
speakers in rooms, we don't just hear the speaker, we hear some version of its
response off every surface in our room. The Sample: The test sample speaker is a boxed, two way speaker, using a Scan-Speak 15cm midwoofer and a Seas 27mm Silk dome tweeter. Both of these drivers or very close variants by the same manufacturers are all over in the world of Hi-End sound reproduction. They are mounted on a box 9 inches wide, 14 inches tall and 12 inches deep. The speaker has 4dB of baffle step compensation. It is made to be placed well into a room, as any high end speaker would be to minimize interaction with nearby surfaces. The crossover is acoustically 3rd order at 2200Hz. The box edges do not have round-overs and the tweeter is not offset. For those of you that have designed crossovers before, it is slightly underlapped, and the drivers sum almost completely in-phase at 2200Hz due to the non-time aligned baffle, this eliminates the off-axis 3dB lump at the crossover point, and overall reduces the power response at the crossover point, much like a linkwitz-riley 4th order crossover. This speaker was designed to be listened to at 20degrees off axis, frequency response (unsmoothed, rectangular window shown below).
|
(Figure 1) Test Sample Frequency Response on Design Axis: |
|
Keep in mind that any speaker with somewhere around 3rd or 4th
order crossovers with similar driver size, spacing, and similar on-axis
frequency response will perform in much the same manner. At the end of this
article I will make some generalizations about how other types of crossovers,
drivers sizes, and spacing will affect what you see in these samples. Horizontal Off Axis Response: First I will look at the horizontal off-axis frequency response. This places weight on the off-axis response of each driver, and the diffraction characteristics of the cabinet, but places very little weight on the crossover. Only if gross errors were made in selecting the crossover frequency would this figure show much of a difference. |
 |
|
Observations: There are three main things to note here: 1. Broad dispersion in lower midrange. Due to the fact that as frequency goes down, all drivers in boxes become omni-directional. Depending on the width and height of the baffle, systems with only forward firing drivers become almost completely omni-directional somewhere right around 500Hz for very small speakers, or down around 350Hz for fairly large size speakers. Generally, from that point up to about 1500Hz they are becoming more directional in a gradual manner. This is easy to spot in the off-axis plot above, as the lines slowly separate as frequency increase from 350Hz up to 1500Hz. 2. Above 1500Hz a general ‘tightening’ of the lines can be observed up to about 3500Hz, where all but the on-axis actually seem to intersect. At 3500Hz, the sample speaker is playing at equal volumes all across the a 120 degree horizontal plane. However at just above 1000Hz, the sample has 7dB less output at 60 degrees off axis. This means that midrange driver is actually starting to beam as low as 1000Hz. It should be mentioned that this is a 15cm driver and it is made of paper. Paper drivers generally have better off-axis response due to the slight flexing of the cone that starts in this region. 3. This broad dispersion continues right up into about 5500Hz, where the tweeter then starts to beam. With smaller diameter tweeters, this point is just pushed higher. Vertical Off Axis Response: First will be the vertical response taken in 10 degree increments above the speaker. Since the speaker is a non-symmetric design (not an MTM), the response above the speaker is different than that ob the response below the speaker. In order to keep things somewhat comprehendible, they will be looked at separately. Vertical off-axis is mostly dictated by the crossover frequency, slope and driver spacing. |
 |
|
Observations: Other than the similar wide dispersion that we saw between 3500Hz and 5500Hz in the horizontal plots, there is one apparent difference. Just above 2000Hz, we see a large suck-out at 30 degrees off axis (magenta). This is due to the fact that the speaker was designed to be listened to with the front baffle vertical and the listening location at an equal height to the speaker. Above and below this listening point, the drivers become different distances from the listener, and cause cancellations to occur, this one as deep as –15dB. If I searched for a few minutes longer, there will probably be a point where this would be –30dB or so. Even at 20 degrees off axis it is –7dB (blue ). |
 |
|
Observations: Much the same as above, a bit worse with regard to off-axis such-outs. Ronin Off-Axis Response: The following measurements were taken at the exact same location, under the same conditions. They apply to both the horizontal and vertical off-axis responses due to the coaxial nature of the midrange and tweeter. |
 |
|
Observations: We see a slight narrowed dispersion from as low as 300Hz on up through 3Khz, where there is a slight broadening of response from 3500Hz to 5000Hz. All of this is due to the coaxial, waveguide mounting of the tweeter and the open baffle dipole nature of the midrange. No matter where the measurement is taken, the tweeter and midrange are at much the same distance as they were on-axis. The dipole radiation keeps a similar dispersion characteristic through the lower midrange all of the way down to 300Hz. Keep in mind that the magic of the music happens between 300 and 3000Hz, and absolute balance in this region is paramount. Also notice that the 25mm tweeter has far better off-axis response at 60 degrees off axis than the 28mm tweeter does in the test sample. Off Axis Impulse Response Measurements: Below you will see impulse response measurements taken on axis and at 50 degrees off axis. On the test sample note the delay in the midrange output, due to the fact that it is farther away in the off-axis measurement. This is similar to the response that would bounce off of the ceiling or floor and arrive at your ears. |
Conclusion
|
Info