One aspect of the work-at-home/stay-at-home environment of the past year has been an increased emphasis on podcasts, online meetings, and online video events. That’s why there’s more interest in microphone technology among those interested in producing a higher-quality experience. Here’s a short run-down on microphones and how they are typically measured and quantified.
A microphone converts sound waves into electrical impulses. There are various types of microphones and some methods of sound pick up are now obsolete. Land-line telephones often contained a carbon microphone in the handset. It consisted of a diaphragm sensitive to changes in air pressure, which causes a powdered carbon resistive element to expand and contract, producing an electrical signal. A carbon microphone does not produce electrical energy on its own, but instead requires an external dc power source. In a phone, the dc came from the central phone office, but a 9-V battery would also work. The microphone output can easily be displayed in an oscilloscope by simply connecting the two microphone leads to a standard 10:1 passive probe, but you must remember to place a dc power source in series with a carbon microphone.
Another vintage technology is the ribbon microphone. Ribbon microphones generate a signal via an ultra-thin ribbon of electro-conductive material suspended between the poles of a magnet. Sound vibrates the ribbon, inducing a voltage at right angles to both the ribbon velocity and magnetic field direction. Ribbon microphones are also called velocity microphones because the induced voltage is proportional to the velocity of the ribbon and thus of the air particles in the sound wave.
Ribbon microphones are still around but are generally confined to high-end audio work. The lightweight ribbon has a resonant frequency below 20 Hz. This gives them an excellent frequency response in the 20 Hz to 20 kHz human hearing range. They are known for capturing high-frequency detail. Where condenser microphones can often sound subjectively aggressive or brittle in the high end, ribbon mics are often described as having a warm, vintage tone.
Early ribbon mics had a reputation for being fragile. They could also be damaged by applying a dc voltage–a concern because other types of mics required a dc power supply, so accidental application of dc was a possibility. Modern ribbon mics generally are built to fail-safe such scenarios.
Podcasters and YouTube celebrities today are far more likely to employ condenser or dynamic mics. Dynamic–also called moving coil–microphones operate by suspending a coil of wire attached to a diaphragm inside a magnetic field. When sound vibrates the diaphragm, it also vibrates the coil and produces an electrical signal.
One difficulty with moving-coil mics is that a single dynamic membrane does not respond linearly to all audio frequencies. The mass of the diaphragm and coil that is good for recording speech, for example, may not respond quick enough to accurately handle higher frequencies. Consequently, some dynamic mics employ multiple membranes for the different parts of the audio spectrum. Combining the resulting signals correctly is difficult; designs that do so tend to be expensive.
Dynamic mics are generally viewed as good all-round mics. They don’t require a power supply, and they exhibit a cardioid polar response pattern. With a cardioid pattern, the microphone filters out noise from its rear.
Another widely used mic employing a vibrating diaphragm is the condenser mic. Here the diaphragm acts as one plate of a capacitor, and sound vibrations change the distance between the plates. In dc-biased condenser mics, the voltage maintained across the capacitor plates changes with the sound vibrations vis the capacitance equation (C = Q⁄V), where Q = charge, coulombs; C = capacitance, farads; and V = potential difference, volts. The capacitance of the plates is inversely proportional to the distance between them for a parallel-plate capacitor.
The capacitance of the mic element (around 5 to 100 pF) and the value of a bias resistor (100 MΩ to tens of GΩ) form a highs-pass audio filter and low-pass for the bias voltage. Within the time-frame of the capacitance change (as much as 50 msec at 20 Hz audio signal), the charge is practically constant and the voltage across the capacitor changes to reflect the change in capacitance. The capacitor voltage varies above and below the bias voltage. The voltage difference between the bias and the capacitor is seen across the series resistor. The voltage across the resistor is what gets amplified. In most cases, the voltage differential is quite significant, up to several volts for high sound levels. The impedance of the circuit is basically that of the relatively high-resistance bias resistor. So only current gain is usually needed, with the voltage remaining constant.
There is another type of condenser mic called the RF condenser. These work from a low RF voltage, generated by a low-noise oscillator. The signal from the oscillator may either be amplitude modulated by the capacitance changes produced by the sound waves moving the diaphragm, or the diaphragm assembly may be part of a resonant circuit that modulates the frequency of the oscillator signal. Demodulation yields a low-noise audio frequency signal with a low source impedance.
The point of the RF condenser is to get rid of the high bias voltage. This permits the use of a diaphragm with looser tension, allowing a wider frequency response. The resulting diaphragm also has a smaller mass, making the mic more responsive to lower sound pressure levels.
Another type of condenser mic is the electret microphone. Electrets also don’t need an external power supply, but they often contain an integrated preamp that does require power. Instead, they make use of a permanent charge in an electret material–a ferroelectric material that has been permanently electrically charged or polarized. A static charge is embedded in an electret by the alignment of the static charges in the material, much the way a permanent magnet is made by aligning the magnetic domains in a piece of iron.
The vast majority of mics made today are electrets. Their applications range from lapel mics to microphones built-in to small sound recording devices. Electret microphones lend themselves to inexpensive mass-production, while inherently expensive non-electret condenser microphones are made to higher quality.
There are several other exotic mics. For example, the microphone in your cell phone is condenser type where a pressure-sensitive diaphragm is etched directly into a silicon wafer by MEMS processing and is usually accompanied with an integrated preamplifier. Digital MEMS microphones have built-in analog-to-digital converter (ADC) circuits on the same CMOS chip for ready integration with digital circuity.
Sophisticated test equipment is available for checking out microphones used in calibrated measurements. But ordinary bench-top instruments suffice for gauging the performance of mics applied in more every-day uses. The key requirement for mic measurements is the availability of a repeatable amount of sound. It is also necessary to gauge sound levels, necessitating the use of a sound pressure level (SPL) meter.
For example, to measure the output impedance of a condenser mic, first set up the sound source and measure the mic output amplitude without any load–a high-impedance scope probe should work fine without loading the mic. Next load the mic output with a resistor equal to your best guess of the output impedance. Measure the output amplitude again. If it is half the original value your guess was correct. Otherwise solve the equation
B ×R = A × (R + X)
where A = unloaded amplitude, B = loaded amplitude, R = the load resistor used, and X = output impedance.
Note that mic output impedance can vary with sound frequency. Dynamic microphones, for example, may exhibit different frequency response depending on the electrical impedance they drive. Professional microphones of both condenser or dynamic types these days are low impedance; from 50 to 200 Ω for condenser mics and around 600 Ω for some dynamic models. Professional condenser microphones typically have active amplification in the mic, so their source impedance stays constant across the audio band. But there’s a limit to the amount of current the amp can source, so the minimum load impedance is often around 1 KΩ, below which the mic starts distorting. Generally, though, the exact input impedance has little effect on the microphone sound.
That said, a thorough check of mic impedance would use a pure sine wave sound source and be conducted at several frequencies.
In addition, most mic properties can only be accurately ascertained in a sound anechoic chamber. Typical performance specs include directionality, frequency response, signal-to-noise ratio (SNR), equivalent self-noise level, and harmonic distortion. An audio sweeping signal generator is quite helpful in running these tests, particularly frequency response.
Few of the above parameters are specified for consumer-grade mics. They are more important for recording studio mics and those used in scientific measurements. For example, though consumer-grade mics may express sensitivity in terms of dB (-32dB±3dB is a typical spec), studio condenser mics will express the mic self-noise level as an equivalent acoustical noise level stated in units of dB A-wtd. A-weighted decibels, abbreviated dB A-wtd, are an expression of the relative loudness of sounds in air as perceived by the human ear. In the A-weighted system, the decibel values of sounds at low frequencies are reduced. For example, a microphone having a self-noise rating of 13 dB A-wtd has a noise floor equivalent to the signal that would be picked up by an ideal microphone placed in a sound field of 13 dB A-wtd. Studio-grade large-diaphragm condenser microphones generally have self-noise ratings in the range from 7 dB A-wtd to 15 dB A-wtd.
With the self-noise level determined, it’s possible to gauge the useful dynamic range of a microphone. The definition of dynamic range is the interval in decibels between the 0.5% THD level and the A-weighted noise floor of the microphone. However, the actual measurement of dynamic range can be problematic. The difficulty is that mics don’t start exhibiting meaningful harmonic distortion levels until they see extremely loud sounds. And the necessary sound may be so loud that the loudspeaker generating it may start to distort before the microphone does. Studio quality condenser microphones have total dynamic ranges as high as 125 or 130 dB.
Similarly microphone self-noise also defines the resulting SNR of the recorded signals. SNR in dB is defined as 10 times the logarithm of the ratio of a standard signal’s power Psignal to the noise power of the microphone Pnoise created by its self-noise: SNR=10×(Psignal/Pnoise).
The standard signal is commonly generated by a sound source putting out a 94 dB SPL tone at 1 kHz. SNR is a relative measure, valid only for a given signal level, while self-noise is an absolute measure of the microphone quality. SNR at a calibrated SPL will, however, give a measure of self-noise, because it is obtained by subtracting the self-noise from the standard signal’s level.
A typical measurement sequence might start by producing a 1-kHz sine signal via a loudspeaker such that the measured volume close to the mic is 94 dB SPL, determined with a sound level SPL meter. The mic output gets recorded on a laptop (use of a laptop allows use of battery power and eliminates 60-Hz hum problems) via a recording program such as Audacity. The loudspeaker is turned off and the program continues to record, hopefully picking up only microphone noise. Then an A-weighting is applied to both the signal and noise portions of the recording. Then calculate the dB RMS of the A-weighted silence/noise-only part and the dB RMS of the A-weighted sine recording. Subtracting the two numbers gives the A-weighted (re. 94 dB SPL) mic SNR.