Understanding Microphone Sensitivity

By Jerad Lewis
Sensitivity, the ratio of the analog output voltage or digital output value to the input pressure, is a key specification of any microphone. Mapping units in the acoustic domain to units in the electrical domain determines the magnitude of the microphone output signal, given a known input. This article will discuss the distinction in sensitivity specifications between analog and digital microphones, how to choose a microphone with the best sensitivity for the application, and why adding a bit (or more) of digital gain can enhance the microphone signal.

contrast in the defi nition of sensitivity of digital microphones compared to that of analog microphones. For an analog microphone with a voltage output, the only limit to the size of the output signal is the practical limit of the systems voltage supplies. Although it may not be practical for most designs, there is no physical reason why an analog microphone couldnt have 20 dBV sensitivity, with a 10 V output signal for a reference-level input signal. This sensitivity could be accomplished as long as the amplifiers, converters, and other circuits could support the required signal levels. Sensitivity of a digital microphone is less flexible; it depends on a single design parameter, maximum acoustic input. As long as the full-scale digital word is mapped to the microphones maximum acoustic input (the only sensible mapping, really), the sensitivity must be simply the difference between this maximum acoustic signal and the 94 dB SPL reference. So, if a digital microphones maximum SPL is 120 dB, then its sensitivity will be 26 dBFS (94 dB 120 dB). There is no way to tweak a design to make the digital output signal higher for a given acoustic input, unless the maximum acoustic input is lowered by the same amount. For digital microphones, sensitivity is measured as a percentage of the full-scale output that is generated by a 94 dB SPL input. For a digital microphone, the conversion equation is

Analog vs. Digital

Microphone sensitivity is typically measured with a 1 kHz sine wave at a 94 dB sound pressure level (SPL), or 1 pascal (Pa) pressure. The magnitude of the analog or digital output signal from the microphone with that input stimulus is a measure of its sensitivity. This reference point is but one characteristic of the microphone, by no means the whole story of its performance. The sensitivity of an analog microphone is straightforward and easy to understand. Typically specified in logarithmic units of dBV (decibels with respect to 1 V), it tells how many volts the output signal will be for a given SPL. For an analog microphone, sensitivity, in linear units of mV/Pa, can be expressed logarithmically in decibels:

Sensitivity %FS Sensitivity dBFS = 20 log10 Output DREF

where OutputDREF is the full-scale digital output level. One last very confusing piece of this comparison is the inconsistent usage of peak and rms levels between digital and analog microphones. The microphones acoustic input levels in dB SPL are always rms measurements, regardless of the type of microphone. The output of analog microphones is referenced to 1 V rms, as rms measurements are more commonly used for comparing analog audio signal levels. However, the sensitivity and output level of digital microphones are given as peak levels because they are referred to the full-scale digital word, which is a peak value. In general, this convention of using peak levels to specify the output of digital microphones must be kept in mind when configuring downstream signal processing that may rely on precise signal levels. For example, dynamic range processors (compressors, limiters, and noise gates) typically set thresholds based on rms signal levels, so a digital microphones output must be scaled from peak to rms by lowering the dBFS value. For a sinusoidal input, the rms level is 3 dB (the logarithmic measure of (FS/2) below the peak level; this difference between rms and peak may be different for more complex signals. For example, the ADMP421, a MEMS microphone with pulse-densitymodulated (PDM) digital output, has a sensitivity of 26 dBFS. A 94 dB SPL sinusoidal input signal will give a 26 dBFS peak output level, or a 29 dBFS rms level. As the outputs of digital and analog microphones have different units, comparing one type to another can be confusing; however, they share a common unit of measure in the acoustic domain, SPL. One may have an analog voltage output, another a modulated PDM output, and a third an I2S output, but their maximum acoustic input and signalto-noise ratio (SNR, the difference between the 94 dB SPL reference and the noise level) can be directly compared.

Sensitivity mV / Pa Sensitivity dBV = 20 log10 Output AREF

where Output AREF is the 1000 mV/Pa (1 V/Pa) reference output ratio. Given this information, with the appropriate preamplifier gain, the microphone signal level can be easily matched to the desired input level of the rest of the circuit or system. Figure 1 shows how the microphones peak output voltage (VMAX) can be set to match an ADCs full-scale input voltage (V IN ) with a gain of V IN/VMAX. For example, an ADMP504 with 0.25 V maximum output voltage could be matched to an ADC with 1.0 V full-scale peak input voltage by using a gain of 4 (12 dB).

ANALOG MICROPHONE: VMAX PEAK OUTPUT VOLTAGE

PREAMP: GAIN = VIN/VMAX

ADC: FULL-SCALE INPUT LEVEL VIN

Figure 1. Analog microphone input signal chain with preamp to match microphone output level to ADC input level. The sensitivity of digital microphones, with units dBFS (decibels with respect to digital full scale), is not so straightforward. The difference in units points to a subtle

Analog Dialogue 46-05 Back Burner, May (2012)

www.analog.com/analogdialogue

By referring to the acoustic domain, not the output format, these two specifications provide a convenient way to compare different microphones. Figure 2 shows the relationship between an acoustic input signal and the output levels of analog and digital microphones for a given sensitivity. Figure 2(a) shows the ADMP504 analog microphone, which specifies 38 dBV sensitivity and 65 dB SNR. Changing its sensitivity, relative to the 94 dB SPL reference point on the left, would result in sliding the dBV output bar up to decrease sensitivity or down to increase sensitivity.
dBV OUTPUT 0 dB SPL INPUT MAXIMUM ACOUSTIC INPUT 120 110 100 REFERENCE SPL (94dB) DYNAMIC RANGE 90 80 70 60 50 40 NOISE FLOOR OF MICROPHONE 30 WITH 65dB SNR 20 10 0 dB SPL INPUT MAXIMUM ACOUSTIC INPUT 120 110 100 REFERENCE SPL (94dB) DYNAMIC RANGE 90 80 70 60 50 40 NOISE FLOOR OF MICROPHONE 30 WITH 65dB SNR 20 10 0 30 dBFS OUTPUT 0 10 10 20 30 40 50 60 70 80 90 100 110 120 SENSITIVITY (38dBV)

Figure 2(b) shows the ADMP521 digital microphone, which specifies 26 dBFS sensitivity and 65 dB SNR. This illustration of the input-to-output level mapping for a digital microphone shows that the sensitivity of this microphone cannot be adjusted without breaking the mapping between the maximum acoustic input and the full-scale digital word. Specifications such as SNR, dynamic range, power supply rejection, and THD are better indicators of microphone quality than sensitivity.

Choosing Sensitivity and Setting Gain

A high sensitivity microphone isnt always better than a low sensitivity microphone. Sensitivity tells something about the characteristics of the microphone but not necessarily about its quality. A balance between the microphones noise level, clipping point, distortion, and sensitivity determines whether a microphone is a good fit for a particular application. A microphone with high sensitivity may need less preamp gain before the analog-to-digital conversion, but it may have less headroom before clipping than a microphone with lower sensitivity. In near-field applications, such as cell phones, where the microphone is close to the sound source, a microphone with higher sensitivity is more likely to reach the maximum acoustic input, clip, and cause distortion. On the other hand, a higher sensitivity may be desirable in far-field applications, such as conference phones and security cameras, where the sound is attenuated as the distance from the source to the microphone increases. Figure 3 shows how the distance of the microphone from the sound source can affect the SPL. The level of an acoustic signal decreases by 6 dB (one-half) each time the distance from the source is doubled.

SIGNAL-TO-NOISE RATIO

(a)
87dB SPL

69dB SPL
20 SENSITIVITY (26dBFS)

63dB SPL 57dB SPL

SIGNAL-TO-NOISE RATIO

40 50

1
60

16

32

DISTANCE FROM SOURCE (Inches)

70 80 90 100 110 120

Figure 3. Sound pressure level at the microphone is reduced as the distance from the source increases. For reference, Figure 4 shows the typical SPL of various sound sources, from quiet recording studios (below 10 dB SPL) up to the threshold of pain (above 130 dB SPL), the point at which the sound causes pain for the average person. Microphones can rarely cover allor even mostof this range, so choosing the right microphone for the required SPL range is an important design decision. The sensitivity specification should be used to match the microphones output signal level across the dynamic range of interest to the common signal level of the audio signal chain.

(b)

Figure 2. Mapping acoustic input level to (a) voltage output level for an analog microphone; (b) digital output level for a digital microphone.

Analog Dialogue 46-05 Back Burner, May (2012)

dB SPL 130 THRESHOLD OF PAIN

120

PROPELLER AIRCRAFT

110 HEAVY MACHINE SHOPS SUBWAY TRAINS, NIAGARA FALLS STORES AND NOISY OFFICES

100

90

80

AVERAGE FACTORY

floor) signals it can faithfully reproduce (120 dB 33dB = 87 dB for the ADMP441). This dynamic range can be reproduced with a 15-bit data word. A 1-bit shift of the data in a digital word results in a 6 dB shift in the signal level, so even a 16-bit audio processor with a 98 dB dynamic range could use 11 dB of gain or attenuation before the original dynamic range is compromised. Note that in many processors, the digital microphones maximum acoustic input is mapped to the DSPs internal full-scale level. In this case, adding any amount of gain reduces the dynamic range by an equal amount and lowers the systems clipping point. Using the ADMP441 as an example, adding 4 dB of gain in a processor with no headroom above full scale would cause the system to clip with a 116 dB SPL signal. Figure 5 shows a digital microphone, with either I2S or PDM output, connected directly to a DSP. In this signal chain, no intermediate gain stage is necessary because the microphones peak output level already matches the DSPs full-scale input word.

70 NORMAL CONVERSATION AT 1m

60

50 AUDIENCE NOISE, MOVIE THEATER QUIET HOMES 30

DIGITAL MICROPHONE: 0dBFS PEAK OUTPUT DSP: 0dBFS PEAK INPUT

40

20

GOOD BROADCAST STUDIOS

10

TARGET PERFORMANCE FOR RECORDING STUDIOS AND CONCERT HALLS THRESHOLD OF HEARING

Figure 5. Digital microphone input signal chain connected directly to a DSP.

Conclusion
This article explained how to understand a microphones sensitivity specification, how to apply it to a systems gain staging, and why, although sensitivity is related to SNR, it is not an indication of the microphones quality as is SNR. Whether designing with an analog or digital MEMS microphone, this should help a designer choose the best microphone for an application and to get the fullest performance from that device.

Figure 4. Sound pressure level of various sources.1 Analog microphones have a wide range of sensitivities. Some dynamic microphones might have sensitivity as low as 70 dBV. Some condenser microphone modules have integrated preamps so they have extra high sensitivity of 18 dBV. Most analog electret and MEMS microphones have sensitivity between 46 dBV and 35 dBV (5.0 mV/Pa to 17.8 mV/Pa). This level is a good compromise between the noise floorwhich can be as low as 29 dB SPL for the ADMP504 and ADMP521 MEMS microphonesand the maximum acoustic inputwhich is typically about 120 dB SPL. An analog microphones sensitivity can be tuned in the preamp circuit that is often integrated in the package with the transducer element. Despite the perceived inflexibility of a digital microphones sensitivity, the level of the microphone signal can be easily adjusted with gain in the digital processor. With digital gain, there is no danger of degrading the noise level of the signal as long as the processor has a sufficient number of bits to fully represent the dynamic range of the original microphone signal. In an analog design, every gain stage will introduce some noise into the signal; it is up to the system designer to ensure that each gain stage is quiet enough to keep its injected noise from degrading the audio signal. As an example, we can look at the ADMP441, a digital (I2S) output microphone with a maximum SPL of 120 dB (26 dBFS sensitivity) and an equivalent input noise of 33 dB SPL (61 dB SNR). The microphones dynamic range is the difference between the largest (max SPL) and smallest (noise

References
Designing with MEMS Microphones. http://ez.analog.com/ community/ask_the_expert/archived/mems-microphones. Lewis, Jerad. AN-1112 Application Note. Microphone Specifications Explained. Analog Devices, 2011. MEMS Microphones. http://www.analog.com/en/audio video-products/mems-microphones/products/index.html.
1

John Eargle, The Microphone Book, Elsevier/Focal Press, 2004.

Author
Jerad Lewis [[email protected]] is a MEMS microphone applications engineer at Analog Devices. He joined the company in 2001 after getting his BSEE from Penn State University. Since then, Jerad has supported various audio ICs, including SigmaDSP, converters, and MEMS microphones. He is currently pursuing an MEng degree in acoustics at Penn State University.

Analog Dialogue 46-05 Back Burner, May (2012)