In our recent What is ANC Technology & How Does it Work? blog post, we delved into how ANC technology works and the three basic types of active noise cancellation (ANC). This post will highlight another kind of active noise cancellation (although it doesn’t usually go by that name) that works to prevent undesirable sounds from getting into a voice signal.
As a leader in audio, video and IoT design for voice-enabled products, as a DSP engineering firm, Cardinal Peak employs an experienced team of engineers that excels at adding high-quality voice control to any product, as well as developing products that require ANC, whether at the microphone or in headphone applications.
Noise-Cancelling Headphones and Earbuds: Now A Remote Work Staple
Imagine you’re on a conference call (who among us isn’t well versed in Zoom or another video conferencing solution by now?), and there’s a loud noise in the background of your work environment — say a dog barking or a washing machine buzzing. Wouldn’t it be wonderful if:
- the audio device you’re using to listen through would prevent that loud noise from reaching your eardrums?
- the system that’s picking up your microphone would block that noise from being transmitted to listeners on the other end?
Active Noise-Cancelling Microphones
Enter ANC at the mic, or multi-microphone noise cancellation.
Let’s face it, we’ve all experienced making a phone call from a noisy street, busy restaurant or crowded public space in which the symphony of background noises can make it impossible to hear an incoming call. Worse yet, none of us wants to be the person in this situation who is yelling into their phone in an attempt to be heard. While active noise-cancelling microphones can minimize background noise, they also need to be able to reject noise from microphone arrays so users sound good to whoever they are talking to (whether on the phone or talking to a voice assistant).
Unfortunately, capturing background sounds accurately enough to provide the maximum amount of noise reduction is challenging for dual-microphone setups. From frequency responses that aren’t perfect to cancellation waveforms not lining up with the phase of the noise once it reaches your ear in an ideal manner, as well as electronic conversion introducing its own noise, it’s unlikely we’ll see 100% microphone noise reduction with current technologies — but we can significantly cut the background volume.
Feedforward vs. Feedback Microphones
Whether using a feedforward or feedback microphone for noise reduction, ANC at the mic requires a digital signal processor (DSP) or other ANC processing hardware to deal with noise filtering and map noise signals to what users will actually hear. While external feedforward microphones have the best noise sensitivity, they are susceptible to short burst high-frequency background noises, which could be amplified. Feedback microphones, on the other hand, capture sound that more accurately reflects noise the wearer actually hears because the mic is located inside the ear device. The best — and most expensive — option, hybrid ANC, combines the best of both worlds (external feedforward mics and internal feedback mics) to deliver optimal noise cancellation.
How Does Active Noise-Cancellation Technology Work with Voice Signals?
While ANC has been used in headphones for people listening to music for years, advances in ANC technology only recently came to microphones for voice applications. Traditionally, voice signals taken in from a microphone in the phone can be filtered and improved, but making or receiving a call in a noisy location can make the voice signal inaudible.
Part of the challenge is that most noise-canceling headphones utilize feedback ANC technology to generate a cancellation signal that is sent to the headphone. Unfortunately, this technique is limited in the frequencies that can be cancelled. Plus, earphones need significant sound-proofing around the ear to block out the external noise through passive noise cancellation.
To ameliorate this issue, most modern mobile phones leverage a multimicrophone design to cancel out ambient noise from voice signals. Sound is captured from the microphone furthest from a speaker’s mouth — the noise signal — and from one closest to the mouth — the desired voice signal. From there, each signal is processed to cancel the unwanted noise from the desired signal, producing improved voice sound quality.
The same is true of noise-cancelling microphones. In order to achieve directionality, such microphones have at least two ports through which sound enters:
- a front port, which is usually oriented toward the desired sound
- another port that’s more distant
Noise cancellation at the microphone doesn’t help you but allows others to hear you better, picking up your voice while ignoring any background noise.
How Do Noise-Cancelling Microphones Know Which Sounds to Reduce?
But how do noise-cancelling microphones (and microphone noise-cancelling software) know what noises are unwanted versus those that need to come through loud and clear?
While microphone shape and positioning can help with noise reduction, the simplest solution from a technical perspective is to have some kind of frequency domain power analysis so that the microphone noise-cancelling technology can determine the difference between the desired sound and the signal that ought to be cancelled.
Basically, the digital signal processor (DSP) looks at the amount of energy in frequency bands, and when a frequency band or a specific frequency — depending on the DSP algorithm’s accuracy — reaches an excessive level, the DSP puts a cancellation waveform in to cancel out external ambient noises. ANC at the microphone is all about detecting the signal that doesn’t belong and cancelling that undesired signal.
Essentially, one microphone is located close to your mouth to pick up your voice while another further away microphone picks up noises from your surroundings. DSP algorithms then work to remove any external noise, leaving your voice — and your voice only. In a noisy environment, both microphones receive noises at similar levels, but the mic closest to someone’s mouth receives the desired voice signal more strongly. Thus, if the external ambient noise signal is subtracted from the voice signal, much of the undesirable noise is cancelled out while the sound the listener wants to hear is retained.
As the vast majority of us continue to work from home, it’s important to remember that despite these incredible technological advances, noise-cancelling headsets will not completely eliminate undesired sounds. Whether your new workspace has a barking dog, crying baby or noisy appliances, even the best noise-cancelling technologies only reduce external sounds by up to 75% — whoever you’re chatting with will still know you’re WFH. By reducing background noise, communication between callers improves due to less repeated and more accurate information, ultimately improving productivity.
Today, implementing an ANC circuit into a phone or in a headset is the most effective method for providing users with the high-quality voice signal they expect under almost any ambient conditions.
Cardinal Peak’s deep expertise with ANC at the microphone allowed us to recently complete a project that required audio filtering in a noisy environment. If you’re curious about voice assistants and the neural network software and new hardware that allows the use of voice technologies in low-power applications, check out this blog post. For more information on how we can support your next active noise-cancellation technology project from initial architecture through product release, reach out to us today!
If you enjoyed this blog, subscribe to our quarterly newsletter below.