In the recent pandemic, people took to virtual meetings using their computers and laptops. However, most often, the substandard quality of the audio led to an unsatisfactory experience. That’s because people’s expectation of consumer devices has increased significantly. They want to make high-quality calls from wherever they may be. They could be on the street, in an open office, or in a crowd.
People expect their devices to have ANC or active noise cancellation, transparent hearing, and voice control. However, these require more sophisticated and better microphones.
For instance, people engaged in video conferencing, want their experience to be as close as possible to a real, face-to-face meeting. Now this depends, to a great extent, on the audio quality, and people expect high-quality audio without having to put on additional devices, such as headphones.
Achieving good quality audio requires the application to use a combination of high-quality hardware and software. It is necessary to have algorithms that provide good noise reduction, reverberation reduction, enhanced beamforming, and good direction of arrival detection. These are essential for high-fidelity transmission and audio recording in a wide variety of conditions and situations. However, the quality of the entire chain is dependent on the primary sensor, the microphone.
Most good-quality microphones that have been around for a long time tend to be large and expensive, and primarily confined to audio recording studios. However, consumer equipment typically requires microphones that are mass-produced with tight manufacturing tolerances, and physically small. MEMS microphones suit these requirements very well.
For a microphone to be qualitatively described as good, it must possess some performance characteristics. The first among them is the SNR or signal-to-noise ratio. SNR of a microphone is the difference in its output between a standard reference signal input and the microphone’s self-noise. All elements of a microphone contribute to its self-noise. This includes the MEMS sensor, package, ASIC, and the sound ports.
SNR is important when the microphone is detecting sounds or voices that are at a distance from it. This is because the input signal decreases with distance, as sound level halves at twice the distance. Further, signal losses can come from the system design, room conditions, and sound channel. A good microphone with a large SNR can capture sound even at large distances. This helps with capturing input signals for algorithms, voice commands, and recording.
The next important characteristic of a good microphone is its THD or total harmonic distortion. This refers to the presence of harmonics in the microphone’s output that are not present in the input signal. The point where the THD reaches 10% is important as this represents AOP or acoustic overload point. At this point, the output from the microphone contains clipping and other noises, because the signal is too loud for the microphone.
The latest MEMS technology allows building of studio-level microphones for consumer devices like laptops. This has been aptly demonstrated by Infineon using their XENSIV IM69D127 and IM73A135 MEMS microphones that are allowing OEMs to build laptops with the next level of audio quality.