The first cases of Active Noise Cancelling (ANC) date back to the 1950’s. They were initially used in helicopter and aircraft pilot headsets. By definition noise is a mixed set of tones at different frequencies and different volumes. The unpredictable nature of noise demands real-time processing which is ideally suited to analog-based solutions.
It is important to pick up the noise at the ear since the phase of the noise is the most critical parameter. In theory and practice sound waves from different sources add their magnitude, i.e. getting the phase wrong can actually create more noise. Getting the phase perfectly right (with the right amount of volume/magnitude) will totally cancel the noise.
For this reason centralized systems have very limited performance. All high performance ANC solutions measure the unwanted noise at each ear separately.
Figure 1: Theory of noise cancellation
In feed-forward architectures, the noise pick-up microphones are located outside the ear cup and point typically away from the ear. Noise travelling toward the ear is picked-up before reaching the ear and as it carries on toward the ear, the electronics generates an equivalent inverted signal that it plays back through the speaker. The objective being that the noise and the inverted signal arrive in the ear canal at exactly the same time, with the same magnitude but phase shifted by 180 degrees. The wanted audio signal (speech or music) can simply be added to the anti-noise signal.
In feedback architectures, the noise pick-up microphones are located inside the ear cup and point typically toward the ear. Noise travelling toward the ear is picked-up almost at the same time as it arrives at the ear. Real-time response with minimum delay is essential to achieve good ANC performance.
In feedback operation the noise pick-up microphone collects not only the unwanted noise, but also collects the desired audio signal. The first step is to remove this audio signal to be sure to treat only the unwanted noise. A filter is constructed to generate an inverted noise signal and the result added back to the audio signal. Feedback systems invariably impact the sound frequency spectrum of the desired audio signal and, if designed right, can enhance the audio performing an equalization function.
Feedforward and Feedback have pros and cons as highlighted in this list. To highlight the most characteristic differences:
- Feedforward has a sweetspot with typically one frequency being able to cancel to quite a high level; figures above 30 dB are not uncommon.
- Feedforward also has a wide bandwidth of 2 kHz with some of the higher performance solutions actively cancelling up to 4 kHz.
- Feedback is typically flatter over its active bandwidth, typically limited to 1 kHz max, but achieving over 20dB for the higher performance solutions.
- Feedforward typically doesn’t affect the audio playback signal and allows an assisted hearing function (inverting the ANC behavior to amplify the surrounding noise).
- Feedback typically enhances the sound of the audio signal enabling bass or treble boost. Feedback also reduces noises generated in the body (heart pumping, occlusion effect, teeth, etc.).
In terms of system design, feedforward is typically easier to carry out since it is less susceptible to headset resonances; however care must be taken with the repeatability of fitting to different people’s ears (risk of uncontrolled leakage). Feedback by definition can better handle leakages.
A new generation of headsets using Hybrid technology (a combination of feedforward and feedback) allows for the best of both worlds at the cost of a more complex electro-acoustic circuit design and system cost.
The ams portfolio continues to grow to ensure the best fit for each customer application. The portfolio can now address mono-stereo, wired-wireless in/on/over-ear with feedforward, feedback and hybrid topologies.
The AS3421 is designed to interface directly to common wireless headsets. Similar to the AS3410 and AS3430, it is designed for both feedforward and feedback architectures, an IC capable of handling hybrid architectures.
The device has One Time Programmable ROM (OTP-ROM) to be used in standalone mode, and/or can be connected via I2C to the host controller and the device settings downloaded during power on phase.
The IC has a new low power mode allowing the user to completely turn off the ANC features of the device, while retaining the high quality audio chain (line-in through to headphone driver). This ensures minimum power overhead with ANC off, without the need for a bypass.
The microphone gain is programmable up to four times and is used to correct not only for the microphone sensitivity variations, but also for the electro-acoustic tolerances of the headset.
The headphone amplifier is a low noise, true ground Class AB amplifier that can be used in either cap free single ended mode or bridge tied configurations.
Line-in is designed for differential inputs from a wireless SoC. It is associated with a gain control to free up resources within the SoC.
It is also useful to note that the filter construction between wired and wireless chips is identical, making life easy for customers wanting to evolve their current headset designs from AS3410/30 to AS3421.
In systems with outward facing microphones, the device has an assisted hearing or "monitor" mode allowing the ANC behavior to be inverted and actually amplify the surrounding noise. This is useful when listening to your neighbor/stewardess etc. without having to take off the headset.
Here are some application examples that use stereo Bluetooth ANC speaker drivers (Figure 2).
Figure 2: Examples of products that use Bluetooth ANC speaker drivers
Interestingly enough, even in a mono environment where one user's ear is open to the surrounding noise, ANC can bring a significant benefit. This curve (Figure 3) describes the "Intelligibility", i.e. the likelihood of hearing a word correctly depending on the signal to noise ratio. A 10 dB improvement corresponds to 40% better chance of hearing the information piped through the headset – quite significant when considering phone numbers, addresses, people's names and other important information.
Figure 3: Word "Intelligibility" based on signal-to-noise ratio
A comprehensive development platform has been developed by ams. The customer is expected to develop their own headset prototype integrating the noise pick-up microphones, then, using the ams development platform, they can develop optimized filters and adjust the system performance through a standard USB interface.
In summary, as much as 30% of adult hearing loss may be due to noise. Hearing loss not only has health risks but also creates behavioral side effects that will interfere with learning and cognitive performance. Early prevention is critical and the good news is that it's easy with ANC!
