Manufacturer's Specifications:
Type: Unidirectional, surface-mounted electret condenser.
Frequency Response: 20 Hz to 20 kHz, at 30° incidence to infinite surface.
Polar Pattern: Half-cardioid (cardioid in hemisphere above mounting surface).
Output Impedance: Nominal, 150 ohms; actual, 90 ohms.
Output Level: -69.0 dB (0.35 mV) at 1 kHz (0 dB = 1 V/µbar, measured with sound source at 30° incidence to infinite surface).
Maximum SPL: 144 dB at 1 kHz, sound source at 30° incidence, 800-ohm load.
Hum Pickup: 16 dB equivalent SPL in 1-mOe field (60 Hz).
Output Noise: 23 dB SPL, A-weighted.
Phasing: Positive voltage at pin 2 with positive pressure on diaphragm.
Power: Two 9-V alkaline batteries (NEMA 1604A) for 300 hours continuous use, or 11 to 52 V d.c. simplex voltage (operational to 9 V d.c.); current drain, 1.8 mA.
Filter Response: In low-cut position, -12 dB/octave below 80 Hz; in flat position, -6 dB/octave below 30 Hz.
Cable: 7.6 meters (25 ft.), two-conductor, shielded, small diameter, be tween microphone and preamp; extendable to 75 feet with accessory cables.
Cases: Microphone, matte black enameled die-cast base and perforated steel grille with replaceable or cleanable fine-mesh screen and foam-pad wind/dirt barrier; preamplifier, matte black enameled die cast aluminum.
Dimensions: Microphone, 3 3/4 in. W x 5 1/4 in. D x 3/8 in. H (9.5 cm x 13 cm x 1.6 cm) preamplifier, 2 3/8 in. W x 4 3/8 in. D x 1-1/16 in. H (6 cm x 11.1 cm x 2.7 cm).
Weight: Microphone, 9.3 oz. (263 grams), not including cable; preamplifier, 11.3 oz. (320 grams), including batteries.
Price: $300.
Company Address: 222 Hartrey Ave., Evanston, Ill. 60204.
This is my third review of a surface-mounted boundary microphone. The previous reviews (March 1983 and March 1985) described the Crown Pressure Zone Microphones (PZMs), which consist of a tiny, omnidirectional electret condenser capsule placed upside down, 1 mm from a metal plate. This assembly must be used on a large baffle surface for optimum results-at least 24 inches in diameter to obtain reasonably flat low-frequency response.
Since "PZM" is a trade name, competitors have used the terms "boundary" or "surface-mounted" to describe their own versions. It is not necessary for the transducer element of a boundary mike to be positioned upside down, as long as the input port or diaphragm is located within a few millimeters of the surface. The element should be small, about 10 mm in diameter. These dimensions pertain to music applications where uniform frequency response and polar pattern is required up to 15 kHz. For speech applications, the upper limit need be only 6 kHz, so these dimensions may be larger. Shure designed a boundary micro phone for speech reinforcement applications long before the introduction of the PZM. It consisted of a cardioid element enclosed in a foam block, which was placed on a desk or table near the talker.
The boundary microphone is most convenient to use in applications where it can be placed on a natural boundary, such as a wall, floor, ceiling, or piano lid. If you need a stand-mounted mike, it makes more sense to use a conventional unit than to mount a large baffle on a stand with a boundary mike. An exception, in my opinion, is a "bipolar pair" stereo array (described in the March 1985 Crown review), which is a low-cost alternative to an expensive stereo microphone.
The omnidirectional boundary microphone has uniform pickup over a hemisphere, and therefore does not discriminate against ambient noise, particularly noises originating behind the microphone, i.e., in the direction opposite to the sound source. To provide discrimination, Shure has developed a directional boundary microphone, the SM91. The sound pickup pattern is a half-cardioid. To achieve uniform characteristics to 15 kHz, the SM91 uses a new electret condenser capsule only 10 mm in diameter. This capsule is mounted on its side rather than upside down. It is positioned close to the backing plate and protected by a foam and stainless-steel screen. Since floor-mounted mikes may get dirty, Shure has thoughtfully provided for easy disassembly and cleaning of the screens.
The microphone capsule includes an FET impedance converter. Up to 75 feet of cable may be used between microphone and preamplifier, including the 25 feet furnished plus a 50-foot extension. These cables are terminated with Switchcraft TA3F miniature three-pin connectors, which are readily available from electronic parts suppliers.
The preamp input circuitry appears to be balanced and designed to reject stray r.f. picked up in the cable. This is a desirable feature for use in urban locations.
The SM91 is powered by the batteries in the preamp box or remotely from equipment having phantom power avail able. Battery power will satisfy most audiophiles, but response to earlier reviews indicates that some readers will want to use remote powering. As previously explained, two kinds of remote power schemes, phantom and A-B, are in use today. The SM91 uses phantom power, though Shure's Technical Data Sheet refers to it throughout as "simplex" power ("phantom" is used parenthetically in one place).
Shure says this is a telephone-company term, but in communications usage (and in my dictionaries of electronic terms), "simplex" refers to radio transmission and reception on a single frequency. One term should not have two us ages within the field of electronics, I believe.
I have described electret condenser and boundary micro phones in the Journal of the Audio Engineering Society (July-August 1985) and have previously reviewed two Shure electret microphones in Audio (the SM81 in August 1980 and the SM85 in May 1982).
Measurements
Figure 1 shows the measured impedance characteristics. The impedance with power on, 94 ohms at 1 kHz, agrees with specifications. (The mike's 680-ohm impedance with power off is still a suitable termination for microphone input circuits.) The inductive rise at 20 kHz with power on raises some questions as to optimum load impedance for best transient response, lowest noise, and minimum sensitivity to stray r.f. pickup. For my simple tests, I used an "unloaded" 150-ohm input of a professional-grade preamp. Shure specifies a minimum load of 800 ohms, so some experimentation with loads might be in order for critical applications.
Fig. 1--Impedance vs. frequency, with battery power supply on (lower curve)
and off (upper curve). Mid-frequency impedance with power on is 94 ohms; with
power off, it is 680 ohms, still low enough to properly load preamp or mixer
inputs. Note that the impedance scale us logarithmic.
The acoustical testing geometry for boundary micro phones is not standardized, and as explained in previous reviews, I have devised my own methods which I think come close to conforming with EIA Standards for testing conventional mikes. The geometry I used for testing the frequency response of the SM91 with various directions of sound incidence is shown in Fig. 2. You must visualize the polar pattern of any microphone as a three-dimensional "figure of revolution." In the case of the SM91, imagine that the two dimensional cardioid shown in Fig. 2A is revolved 360° about its axis. Then, the plane surface of the wood baffle cuts the 3-D cardioid in half. I measure the responses of a conventional, cylindrical mike with a circular diaphragm in only one plane, the horizontal. Because of symmetry, the mike can presumably be rotated in its mounting without changing these responses. My acoustic tests consist of measuring frequency response at various angles of incidence. If these strip-charts were recorded at many angles, one could plot polar diagrams at selected frequencies. A boundary mike requires response tests in one or more half-planes orthogonal (perpendicular) to the boundary plane.
The omnidirectional mikes previously reviewed were tested in only one half-plane for angles from 0° to 90°. The 0° axis was perpendicular to the baffle. Because of symmetry, three curves (0°, 45°, and 90°) were adequate to describe the performance.
Figure 2 shows that the SM91 requires tests in two orthogonal vertical- planes. For clarity, I refer to one plane as "longitudinal" because it is aligned with the principal axis of pickup (Fig. 2A). The other (Fig. 2B) is called "transverse" because it is at right angles to the principal axis. Shure defines the angle or direction at which frequency response is measured as being theta and being 30°. If the SM91 is visualized as being on a stage apron, then an included theta angle of 0° to 90° would include all of the upstage-downstage sound sources. The audience noise and reverberant sound sources would lie at an included angle of 90° to 180°, and presumably would be attenuated. If the mike is connected to a sound-reinforcement system, these angles would include reflected amplified sound, which should be attenuated to minimize feedback.
The transverse plane includes sources to stage left and right. Note that this plane intersects the longitudinal plane where theta equals 90°. An inclined transverse plane intersecting where theta equaled 30° or 45° might have been a better choice for my tests, but this would have been too difficult to set up because the heavy-duty pipe stand required to support the massive baffle had no means to rotate the baffle about a horizontal axis, but only allowed rotation about a vertical axis. In all tests, the sound source remained in a fixed position while the baffle and mike were rotated.
I used angular increments of 45° for theta and alpha. Values of theta were: 0°, 45°, 90°, 135°, and 180°. Values of alpha were: 0° (same as theta being equal to 90`), 45°, and 90°, in one quadrant only because of symmetry. Shure defines theta being equal to 30° as the principal axis for response testing. It is of little concern that I omitted 30°, because the 0° and 45° response curves which I measured are smooth, parallel, and not too different from Shure's 30° curve.
Fig. 2--Geometry of sound tests for a directional boundary microphone in longitudinal
(A) and transverse (B) vertical planes.
Fig. 3--Frequency response on pseudo-infinite baffle (0 dB equals -61 dBV/Pa).
Fig. 4--Front-hemisphere frequency response for three longitudinal vertical
angles, with microphone outdoors on 24-inch baffle at 6-foot distance (0 dB
equals -61 dBV/Pa).
Fig. 5--Rear-hemisphere frequency response for three longitudinal vertical
angles, under same conditions as Fig. 4.
Fig. 6--Same as Fig. 4, but with microphone on 48-inch baffle.
The previous reviews of boundary mikes showed that my sensitivity data were 6 dB higher than the manufacturer's because of a difference in calibration procedure. They calibrated the SPL at the baffle surface, but I calibrated free- space SPL prior to the introduction of the baffle. In addition to being easier to do, my procedure, I believe, follows EIA Standards.
The first response tests were conducted indoors, using a 2-inch-diameter precision sound source, with the mike and 24-inch square baffle only 6 inches away and oriented so theta was 90°. In previous reviews, I found that this produces the smoothest response curve for boundary mikes because it minimizes the wiggles due to reflections at the baffle edges. The reason is that, in the case of a 24-inch baffle, the reflected sound path is greater than 2 feet, compared to the 6-inch direct sound path. The results are similar to what one would expect from a large surface such as a floor or wall; hence the term "pseudo-infinite baffle response." The PZM microphone manufacturer conducted a similar test by using Time-Delay Spectrometry (TDS) to minimize reflections. The TDS method, unfortunately, had a lower limit frequency of 200 Hz in that test. My conventional test method, made possible by the very small sound source, works down to 30 to 40 Hz. In all tests, the microphone was mounted about 1 inch from the center of the baffle.
Figure 3 shows my pseudo-infinite baffle curve for the SM91. Sound is incident where theta equals 90°. If the mike were an ideal cardioid in free space, the 90° response would be 6 dB below the 0° response. However, this is not true for a boundary mike, as shall be seen. The SM91 shows an essentially flat curve to 2 kHz, then rises to +9.5 dB at 5 kHz. The manufacturer's curve for theta equaling 30° shows a rising response which, from 50 Hz to 5 kHz, increases 6 or 7 dB. Previous reviews showed that a rising response is desirable for boundary microphones. The astute reader will wonder why there is no bass boost due to proximity effect.
The response at 6 inches is much like the curves (to be discussed) at 6 feet. I can think of two reasons for this. First, Fig. 3 shows the response where theta equals 90°, and a cardioid has no proximity effect at 90°. Second, the particle velocity is zero at a boundary, which should cause proximity effect to vanish. In most applications, the microphone will be too far from the source for proximity effect to be a factor.
I tried to measure directional responses indoors, at 12 and 24 inches, with the small source and 24-inch baffle, but results were poor. Then I resorted to outdoor testing with my large, spherical source (an 8-inch cone speaker in an 18 inch sphere, as compared to my small source's 2-inch aluminum piston in an 8-inch sphere). This was calibrated with a laboratory ribbon velocity mike prior to introducing the baffle. (See Audio, September 1978 and March 1983).
The baffle/mike system was mounted 6 feet from the source so that a semblance of accuracy could be maintained with a 48-inch baffle rotated to the various angles of theta or alpha.
To change from longitudinal to transverse planes, I simply turned the mike 90°. First, I mounted the SM91 on a 24-inch baffle. Figure 4 shows the front-hemisphere responses for theta angles of 0°, 45°, and 90°. The lumps in the 90° response probably are due to edge reflection, because otherwise the curve should resemble Fig. 3. The 0° and 45° curves do not show these wiggles. The 45° curve is several dB above the 90° curve, as with a cardioid in free space. The 0° curve is 6 dB down, similar to the 90° curve, but for a different reason: At 0° incidence, the pressure doubling vanishes. I set the 0-dB reference at-61 dBV/Pa because it looked best for the 0° and 90° curves. A better measure of axial sensitivity is the 1-kHz level of the 45° response, which averages about-55 dBV/Pa. Rated output at 30° and 1 Pa (94 dB SPL) is-49 dBV. Thus, these measurements are 6 dB low compared to Shure's value, just opposite to the outcome or the PZM tests. I trust that this disparity is likewise attributable to calibration techniques. Alternatively, it may just be a worst-case sum of production tolerances, lab differences, and the difference between sensitivity at theta angles of 30° and 45°.
This caused a problem with noise-level calculation, which will be discussed later. The output level, -55 dBV/Pa, is acceptable for a 94-ohm mike; my JAES paper establishes -60 dBV/Pa as a minimum for a 250-ohm (passive) mike.
The frequency response curve for 45° rises about 12 dB, compared to 8 dB for the 0° measurement. Shure's value of 6 to 7 dB at 30° is significantly less than my 45° rise. The response curves are very smooth from 40 Hz to 20 kHz, so that if extra "brightness" is perceived, it may be equalized to suit the application.
Figure 5 shows responses for the rear hemisphere. If I had plotted the 45° curve as a reference, the discrimination at 135° at 1 kHz would be 10 dB, and at 180° would be 24 dB. This is excellent performance indeed.
Figures 6 and 7 show front- and rear-hemisphere responses with a 48-inch baffle. The rise of the 45° curve is the same as 0°: 10 dB, more in line with Shure's curve for 30°.
Again, these curves are smooth, but the 90° curve is lumpy at frequencies about one octave lower than the curve in Fig. 4. This corresponds to the difference in baffle sizes. It may be concluded that only the 90° response of the SM91 is sensitive to baffle-edge reflections. Discrimination at 135° and 180° compares favorably to Fig. 5.
It was mentioned in the PZM reviews that the pressure-doubling effect ( + 6 dB) vanishes at a frequency related to baffle size, thus causing a step-down in bass response.
(Frequencies are listed in the March 1983 review.) This effect is only very slightly seen in Fig. 6, as a 3-dB rise in the 45° bass response as compared to Fig. 4.
Figure 8 shows the transverse-plane response with the 48-inch baffle. The 0° curve is similar to the theta equals 90° curve. The 45° response picks up some edge reflections, but the 90° curve is smooth.
I conclude that the frequency response of the SM91, when the unit is placed on a large surface, is uniform for frontal angles of both theta and alpha equaling 0° and 90° and probably all angles in between. It is smooth from 40 Hz to 20 kHz, rising about 8 to 10 dB, which is appropriate for a boundary mike. Response where theta equals 180° is essentially a perfect null.
Fig. 7--Same as Fig. 5, but with microphone on 48-inch baffle.
Fig. 8--Frequency response for three transverse vertical angles, with microphone
outdoors on 48-inch baffle at 6-foot distance (0 dB equals-61 dBV/Pa).
Fig. 9--Noise spectrum (1/2-octave bands). Overall noise levels are -114 dBV,
unweighted, and -119 dBV, A-weighted.
Measured noise level versus frequency is shown in Fig. 9.
Battery power was used in this test. The measured levels in dBV are shown because the conversion to equivalent SPL is subject to interpretation. If I use my measured 45° value of acoustic sensitivity,-55 dBV/Pa at 1 kHz, the A-weighted overall noise level is 30 dB equivalent SPL. Rated noise level is 24 dB, but to obtain this value I would have to use the rated sensitivity figure. The noise level of the PZM I used for comparison measured -118 dBV A-weighted, which equaled 24 dB SPL, based on its -48 dBV/Pa measured sensitivity. (The PZM's 680-ohm impedance accounted for the higher output voltage level.) Phasing was as specified, and agrees with EIA Standards. Clipping level exceeded 0.4 V peak output on speech (-11 dBV rms). Conversion to SPL depends on the acoustic sensitivity figure: Using-55 dBV, I obtain 138 dB, and-49 dBV results in 132 dB. Remembering that clipping exceeds these figures, the results compare favorably to the 144 dB specified.
Use and Listening Tests
------ Top view of the Shure SM91.
Note the steel grille, useful protection for a mike that will often be used on floors.
The PZM or boundary has now been established as a legitimate high-quality microphone for both audiophile and professional applications, so it is a suitable reference for comparison to the SM91. I did not have any Crown units on hand, but I did have a Realistic, which was purchased from Radio Shack for $39.95. I found that the frequency response and noise level of this mike compare very favorably to the Crown units, so I elected to use it as a reference.
The outdoor tests of the SM91 were frequently disrupted by wind gusts, which produced large excursions of the chart recorder pen. The gusts did not exceed 10 knots, which is an acceptable value for outdoor sound testing. I could not use the integral low-frequency filter while I was making my tests.
Indoors, I found that I could "pop" the SM91 with "Peter Piper picked ..." spoken at 12 inches, and the mike was silent for a second following each pop. The Realistic would pop at 6 inches (with windscreen) and would not go silent after the pop impulse. The low-frequency filter in the SM91's preamp reduces but does not eliminate this effect. Popping should not be a problem, however, because the SM91 would rarely be used for close speech pickup.
Vibration noise caused by tapping my desk produced slightly more noise from the SM91 than from the Realistic. The SM91 mike and preamp have negligible magnetic hum pickup compared to the Realistic. When testing with speech at 12 inches from the mikes on my desk, I found that the SM91 had more high and low frequencies than the Realistic and less room-noise pickup. The SM91 made a superior speech tape in a room with a noisy air conditioner.
The SM91 and the Realistic were placed side by side on the carpeted center aisle of a church (which seats 900 and has a ceiling 40 feet high). They were used to record a concert by a chorus of 25 young persons about 15 feet distant, with a piano between them and the mikes, plus a set of drums to the right of the piano. A two-track cassette tape using the two boundary mikes was made on an inexpensive Sanyo recorder, while an Aiwa AD-F990B recorder (with Dolby HX Pro) was recording the concert simultaneously using a permanently flown AKG C422 stereo mike 20 feet above the boundary mikes. Comparisons of the boundary-mike tracks on a monaural NB basis showed that the SM91 had more vocal presence than the Realistic and more cymbal sound. Bass drum and piano sounded similar on both tracks. Overall, the SM91 sounded clearer and more pleasing. The audience was quiet, so no difference in ambient noise level could be detected. Reverberation was not heard in either mike, because the sources were relatively close and the room acoustics were not very live. Comparison to the Aiwa tape on a stereo basis is, of course, unfair. The HX Pro feature of the Aiwa, plus the very flat response of the AKG C422, means that high frequencies were very accurately recorded. I thought that the SM91's highs, by comparison, were a little excessive in relation to its low-frequency response. Surprisingly, the stereo perspective from the pair of dissimilar boundary microphones was not grossly different from that of the C422, and rather pleasing, I thought. I suspect that a pair of SM91s placed side by side and perhaps angled outwards might be excellent for stereo recordings.
I did not hear any noise (hiss) from either the Realistic or the SM91, but the sound level was consistently high at the mikes. No trouble was encountered with wind noise, of course, in this indoor application. The cable of the SM91 is too stiff and kinky. It would not lie flat on the floor, which is essential for a surface-mounted microphone.
The recorder gain settings for the Realistic and SM91 were much different. The gain setting for the Realistic was normal, but the SM91 required a much higher setting because of its low impedance and output voltage. I think that a 150-to-600 ohm transformer should be used if the SM91 is connected to the mike input of a cassette recorder. (Jensen makes excellent transformers for this purpose.) The minor problems evidenced in the measurements did not show up in the listening tests. These tests were limited in scope, due to lack of another concert to record and the fact that the single mike supplied for review precluded stereo recording. I conclude from these tests that the SM91 can make a cleaner speech tape in a noisy room, and a livelier sounding choral recording, than an omnidirectional boundary mike. If I were permitted only one pair of boundary mikes in my recording suitcase, I would choose the SM91.
--Jon R. Sank
(Source: Audio magazine, Jun. 1986)
Also see:
Shure SM85 Unidirectional Condenser Microphone (May 1982)
Shure SM89 Shotgun Condenser Mike (Dec. 1990)
Sennheiser MKH 40 Microphone (Equip. Profile, Jan. 1988)
Sennheiser HD-540 Reference II headphones (AURICLE, Jun. 1992)
Sennheiser Electronic Corporation (ad, Nov. 1984)
= = = =