AN-1112

APPLICATION NOTE
One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106, U.S.A. • Tel: 781.329.4700 • Fax: 781.461.3113 • www.analog.com

Microphone Specifications Explained

by Jerad Lewis

INTRODUCTION where OutputREF is the full-scale digital output level (1.0).

A MEMS microphone IC is unique among Analog Devices, Inc., The 94 dB SPL acoustic input signal for the sensitivity speci-
products in that its input is an acoustic pressure wave. For this fication is an rms level; the peak level of this signal is 97 dB SPL.
reason, some specifications included in the data sheets for these Because the 0 dBFS reference used for a digital microphone’s
parts may not be familiar, or familiar specifications may be applied sensitivity is a peak level, the resulting sensitivity specification
in unfamiliar ways. This application note explains the specifica- is also a peak level. For example, the ADMP521’s −26 dBFS
tions and terms found in MEMS microphone data sheets so that sensitivity means that its output will be −26 dBFS peak with a
the microphone can be appropriately designed into a system. 94 dB SPL rms input signal. The rms output level for this input
signal will be −29 dBFS. This 3 dB difference between peak and
SENSITIVITY
rms levels needs to be considered when designing and tuning
The sensitivity of a microphone is the electrical response at its the signal processing that follows the microphone.
output to a given standard acoustic input. The standard refer-
Higher sensitivity does not always indicate a better microphone.
ence input signal for microphone sensitivity measurements is a
A microphone with higher sensitivity typically has less headroom
1 kHz sine wave at 94 dB sound pressure level (SPL), or 1 pascal
between the output level under typical conditions, such as conver-
(Pa, a measurement of pressure). A microphone with a higher
sational speech, and the maximum output level. In near-field
sensitivity value has a higher level output for a fixed acoustic
(close-talking) applications, a microphone with higher sensitivity
input than a microphone with a lower sensitivity value. Micro-
is more likely to cause distortion. This distortion often reduces
phone sensitivity in decibels (dB) is typically a negative number;
the overall dynamic range of the microphone.
therefore, a higher sensitivity is a smaller absolute value.
It is important to note the units presented with the sensitivity DIRECTIONALITY
specifications of the microphone. It is incorrect to directly Directionality describes the pattern in which the microphone’s
compare the sensitivity of two microphones if the sensitivity is sensitivity changes when the sound source changes position in
not specified with the same unit. For analog microphones, the space. All Analog Devices MEMS microphones are omnidirec-
sensitivity is typically specified in units of dBV, that is, decibels tional (or omni), which means that they are equally sensitive to
with reference to 1.0 V rms. For digital microphones, the sensitivity sound coming from all directions, regardless of the orientation
is typically specified in dBFS, that is, decibels with reference to a of the microphone. Figure 1 shows a 2-axis polar plot of the micro-
full-scale digital output (dBFS). For digital microphones, a full- phone’s response. This plot looks the same regardless of whether
scale signal is defined as the highest signal level that can be output the microphone’s port is oriented in the x-y, x-z, or y-z plane.
from the microphone; for Analog Devices MEMS microphones, 0°
0dB
this level is 120 dB SPL. See the Maximum Acoustic Input section 330° 30°

for a fuller description of this signal level. 500Hz

–10dB

1kHz
Sensitivity is the ratio of the input pressure to the electrical output 300°
5kHz
–20dB 60°

(voltage or digital word). For analog microphones, sensitivity is

–30dB
typically measured in mV/Pa. This measurement is converted to
a decibel value with the following equation: 270° 90°

 Sensitivity mV / Pa  –30dB

Sensitivity dBV = 20 × log 10  


 Output REF  –20db
240° 120°

where OutputREF is the 1 V/Pa (1000 mV/Pa) reference output ratio. –10dB

For digital microphones, sensitivity is measured as a percentage

09766-001

210° 0dB 150°

of the full-scale output that is generated by a 94 dB SPL input. For 180°

a digital microphone, the conversion equation is Figure 1. Omnidirectional Microphone Response Plot
 Sensitivity % FS 
Sensitivity dBFS = 20 × log 10  

 Output REF 
Rev. A | Page 1 of 4
AN-1112 Application Note
The directional response of a system may not be omnidirectional SPL and is not a design parameter that is adjustable by changing
when an omni microphone is designed into a larger enclosure, the gain of the microphone ASIC.
such as a cell phone. For the system designer, an omnidirectional dBV OUTPUT
0
microphone provides greater flexibility in the design of the system
response to an acoustic input than if the microphone itself has a dB SPL INPUT
–10
MAXIMUM ACOUSTIC INPUT 120
directional response.
–20
Multiple omnidirectional microphones can be arranged in arrays to 110

create various directivity patterns and for beamforming applications. 100

–30
SENSITIVITY (−38dBV)
SIGNAL-TO-NOISE RATIO (SNR) REFERENCE SPL (94dB)
90
–40

The signal-to-noise ratio (SNR) specifies the ratio of a reference

DYNAMIC RANGE
–50
80
signal to the noise level of the microphone output. This measure-

SIGNAL-TO-NOISE RATIO
ment includes noise contributed by both the microphone element 70
–60

and the ASIC incorporated into the MEMS microphone package. –70
60
The SNR is the difference in decibels between the noise level and
a standard 1 kHz, 94 dB SPL reference signal. 50
–80

SNR is calculated by measuring the noise output of the micro- 40

–90

phone in a quiet, anechoic environment. This specification is

EQUIVALENT INPUT NOISE (32dB SPL), –100
typically presented over a 20 kHz bandwidth as an A-weighted NOISE FLOOR OF MICROPHONE 30
WITH 62dB SNR
value (dBA), which means that it includes a correction factor that 20
–110

corresponds to the human ear’s sensitivity to sound at different

–120
frequencies. When comparing SNR measurements of different 10

09766-002
microphones, it is important to make sure that the specifications 0
are presented using the same weighting and bandwidth; a reduced
Figure 2. Relationship Between dB SPL Input and dBV Output
bandwidth measurement makes the SNR specification better than for Analog Microphones
it is with a full 20 kHz bandwidth measurement.
DYNAMIC RANGE dB SPL INPUT dBFS OUTPUT
MAXIMUM ACOUSTIC INPUT 120 0
The dynamic range of a microphone is a measure of the difference
between the loudest and quietest SPLs to which the microphone 110 –10
responds linearly. The dynamic range of a microphone is not the
–20
same as its SNR (in contrast with audio ADCs or DACs, where 100
REFERENCE SPL (94dB) SENSITIVITY (−26dBFS)
the dynamic range and SNR are typically equivalent). 90 –30
DYNAMIC RANGE

The SNR of the microphone measures the difference between the 80 –40
noise floor and a 94 dB SPL reference, but the microphone still
SIGNAL-TO-NOISE RATIO

has a great deal of useful signal response above this reference 70 –50

level. The microphone responds linearly to acoustic signals from 60 –60

94 dB SPL up to the maximum acoustic input of 120 dB SPL.
Therefore, the dynamic range of a MEMS microphone is its 50 –70

SNR + 26 dB, where 26 dB = 120 dB − 94 dB. For example, the

40 –80
ADMP404 with an SNR of 62 dB has a dynamic range of 88 dB. EQUIVALENT INPUT NOISE (33dB SPL),
NOISE FLOOR OF MICROPHONE
WITH 61dB SNR 30
–90
Figure 2 shows the relationship between an acoustic input
measured in dB SPL and a microphone’s voltage output in dBV. 20 –100
The dynamic range and SNR measurements are shown between
these two scales for reference. Figure 2 uses an ADMP404 with a 10 –110
09766-003

−38 dBV sensitivity and 62 dB SNR to show these relationships. 0 –120

A similar relationship between dB SPL input and dBFS output Figure 3. Relationship Between dB SPL Input and dBFS Output
for Digital Microphones
for digital microphones is shown in Figure 3. Note that in this
figure, the maximum acoustic input of 120 dB SPL is mapped
directly to a 0 dBFS output signal. As long as the maximum
acoustic input corresponds to 0 dBFS and is set at 120 dB SPL, a
digital microphone always has a sensitivity of −26 dB. This is a
function of the definition of sensitivity being measured at 94 dB

Rev. A | Page 2 of 4
Application Note AN-1112
EQUIVALENT INPUT NOISE (EIN) 10

Equivalent input noise (EIN) is the output noise level of the 8

microphone, expressed in SPL, as a theoretical external noise 6

source placed at the microphone’s input. Input SPLs below the 4

SENSITIVITY (dB)
EIN level are under the noise floor of the microphone and outside 2
the dynamic range of signals for which the microphone produces
0
an output. EIN can be derived from either the dynamic range or
–2
SNR specification as follows:
–4
EIN = maximum acoustic input − dynamic range
–6
EIN = 94 dB − SNR
–8
The EIN of a microphone with 62 dB SNR and 120 dB maximum –10

09766-004
acoustic input is 32 dB SPL—approximately the SPL that would 100 1k 10k

be generated by a soft whisper in a quiet library at a distance of FREQUENCY (Hz)

5 meters. Figure 2 and Figure 3 show the EIN of the microphones. Figure 4. Frequency Response Mask

FREQUENCY RESPONSE
10
The frequency response of a microphone describes its output
level across the frequency spectrum. The high and low frequency

NORMALIZED SENSITIVITY (dB)

limits are described as the points at which the microphone
response is 3 dB below the reference output level at 1 kHz. The 0
reference level at 1 kHz is customarily normalized to 0 dB. Using
these conditions, Analog Devices MEMS microphones typically
have a uniform frequency response from 100 Hz to 15 kHz.
The frequency response specification also includes the deviation –10

limits from a flat response within the pass band. These values,
expressed as units of ±x dB, show the maximum deviation of
the output signal from a nominal 0 dB level between the −3 dB
points. –20

09766-005
100 1k 10k

For Analog Devices MEMS microphones, the low frequency FREQUENCY (Hz)

roll-off below the lower −3 dB point is first order (6 dB/octave, Figure 5. Typical Frequency Response (Measured)
or 20 dB/decade). The high frequency roll-off above the upper
−3 dB point is second order (−12 dB/octave, or −40 dB/decade). TOTAL HARMONIC DISTORTION (THD)
The data sheets for MEMS microphones show this frequency Total harmonic distortion (THD) is a measurement of the level of
response in two figures: one figure shows the frequency response distortion on the output signal for a given pure tone input signal.
mask, and the other figure shows the typical measured frequency This measurement is presented as a percentage. This percentage
response. The frequency response mask figure shows the upper is the ratio of the sum of the powers of all harmonic frequencies
and lower limits of the microphone output across frequency; the above the fundamental frequency to the power of the tone at the
microphone output is guaranteed to be within this mask. The fundamental frequency.
typical frequency response figure shows an actual microphone’s 5

response across the frequency band. Figure 4 and Figure 5 show ∑ Power ( f harmonic_x )
x =1
examples of these two plots taken from the ADMP404 data sheet. THD =
Power(f fundamental )
A microphone with a wide, flat frequency response helps in
the design of systems that require natural sound and high A higher THD measurement indicates a higher level of harmonics
intelligibility. present at the output of the microphone. The THD of the MEMS
microphones is calculated from the first five harmonics of the
fundamental.

Rev. A | Page 3 of 4
AN-1112 Application Note
The input signal for this test is typically at 105 dB SPL, which is (>70 dBV for analog microphones and >80 dBFS PSR for digital
11 dB above the reference SPL of 94 dB. THD is measured at a microphones). As shown in Figure 6 (taken from the ADMP504
higher SPL than other specifications because, as the level of the data sheet), PSR increases to a level just under −50 dB at high
acoustic input signal increases, the THD measurement typically frequencies.
increases as well. A rule of thumb is that the THD triples with 0
every 10 dB increase in input level. Therefore, THD less than
3% at 105 dB SPL means that the THD will be less than 1% at –10

95 dB SPL. –20

This measurement is not to be confused with total harmonic –30

distortion plus noise (THD + N), which measures not only the

PSR (dB)
level of the harmonics but also all other contributions to noise –40

on the outputs. –50

POWER SUPPLY REJECTION (PSR) –60

The power supply rejection (PSR) of a microphone is a mea-
–70
surement of its ability to reject noise present on the power
supply pins from the signal output. PSR is measured by –80

10140-005
100 1k 10k
superimposing a 100 mV peak-to-peak square wave at 217 Hz
FREQUENCY (Hz)
on the VDD pin of the microphone. The PSR specification is
Figure 6. Typical Power Supply Rejection vs. Frequency
a measurement of this signal’s strength at the output of the
microphone. This specification is shown at 217 Hz because, in
MAXIMUM ACOUSTIC INPUT
GSM phone applications, the 217 Hz switching frequency is
typically a dominant source of noise on the power supply. The maximum acoustic input is the highest acoustic sound
pressure level (SPL) that the microphone can input before
The difference between PSR and power supply rejection ratio
clipping. SPLs higher than this specification cause severe
(PSRR) is that PSR shows the absolute level of the signal on the
nonlinear distortion of the output signal. The clipping point is
output, while PSRR is a ratio of the input level to the output
defined as the point at which THD is 10%. Maximum acoustic
level. Consequently, PSRR is expressed as a negative number
input is specified in peak SPL, not rms level.
with units dB. PSR is a positive value with units dBV for analog
microphones and units dBFS for digital microphones. The maximum acoustic input of the Analog Devices MEMS
microphones is 120 dB, which corresponds to a sound pressure
The data sheets for MEMS microphones also show a graph of the
level in air of 20 Pa.
PSR across the 100 Hz to 10 kHz frequency range. These
microphones have excellent low frequency noise rejection
REVISION HISTORY
3/12—Rev. 0 to Rev. A
Changes to Sensitivity Section ........................................................ 1
Changes to Power Supply Rejection Ratio Section and
Maximum Acoustic Input Section ................................................. 4
3/11—Revision 0: Initial Version

©2011–2012 Analog Devices, Inc. All rights reserved. Trademarks and

registered trademarks are the property of their respective owners.
AN09766-0-3/12(A)

Rev. A | Page 4 of 4