USD

Build a True Wireless Fitness Hearable—Part 2: Audio Processing

By Stephen Evanczuk

Contributed By Digi-Key's North American Editors

Editor’s Note: Although they have great potential, fitness hearables present significant design challenges in three key areas: biomeasurement, audio processing, and wireless charging. This series of three articles explores each of these challenges one by one and shows developers how they can take advantage of ultra-low-power devices able to more effectively create fitness hearables. Part 1 addressed biomeasurement of heart rate and SpO2. Here, Part 2 looks at audio processing. Part 3 will discuss solutions for wireless charging and power management for fitness hearable designs.

As discussed in Part 1, in-ear smart wireless audio earbuds, also called true wireless hearables, have emerged as popular audio playback devices, particularly during fitness activities when wires can interfere with movement or equipment. By adding health metric measurements to these designs, developers can create "fitness hearables" that deliver both audio playback and health data.

While the addition of biomeasurement features is an exciting development, designers cannot lose sight of the core function of hearables: high quality audio playback. The problem now is how to maintain high quality audio playback, while at the same time adding new functions in such a small form factor with satisfactory battery life.

This article will discuss the role of audio codecs and processors and outline the core elements of an audio system architecture for hearables. It will then introduce a sophisticated audio codec from Maxim Integrated and show how designers can use it to meet user expectations for high quality sound in a compact form factor—with extended battery life.

Audio codecs and processors

Audio codecs and specialized audio processors have served high-performance audio designs for years. Each combines a complete signal chain for sampling, converting, and conditioning an audio signal. While codecs (derived from coder-decoders) have conventionally limited their capabilities to audio signal encoding and decoding using hardwired firmware, audio processors typically build this functionality around a programmable digital signal processor (DSP). Increasingly, the lines between these classes of products have blurred with the emergence of reprogrammable codecs and audio processors with hardwired features. In either case, developers can find powerful audio signal processing devices able to meet the demands of the most discerning audiophile.

The broad popularity of small in-ear audio earphones, or earbuds, has further driven the evolution of these audio signal processing devices toward providing a complete audio subsystem on a chip. Combined with wireless communications and wireless charging technologies, these devices can provide the foundation for true wireless earbuds able to provide users with remarkably rich sound without encumbering cables.

The hearable evolution

Unlike more conventional wired earbuds, however, true wireless earbuds significantly expand the challenges facing their developers. These products must satisfy the listener's need for audio performance while also meeting the mobile user's expectations for convenience and usability. As a result, the design must provide exceptional sound quality and overall features while also providing minimum form factor and maximum battery life. Fortunately, developers can find a wide range of audio codecs and audio processors able to meet this broad set of requirements.

So-called smart earbuds, or hearables, represent a natural evolution of true wireless earbuds. Along with other advanced functional capabilities, hearables add sensors for biomeasurement, motion detection, and other capabilities designed to enhance the user's well-being and ambient awareness.

Although functionally complex, fitness hearable designs can build on a hardware platform of readily available system-on-chip (SoC) devices designed specifically for these low-power applications. As discussed in Part 1 of this series, the Maxim Integrated MAXM86161 biosensor provides all the biomeasurement capabilities required in these products. Similarly, the Maxim Integrated MAX98090 audio codec provides a complete audio subsystem capable of supporting the diverse range of audio capabilities being designed into emerging fitness hearables. Using these devices, along with a Bluetooth (BT) radio frequency (RF) controller and power management integrated circuits (PMICs), developers can implement the hardware foundation underlying sophisticated fitness hearable designs (Figure 1).

Diagram of Maxim capabilities of a true wireless audio earbud with biosensing capabilities (click to enlarge)Figure 1: A fitness hearable extends the capabilities of a true wireless audio earbud with biosensing capabilities but faces the same requirements for high quality sound playback and extended battery life. (Image source: Digi-Key Electronics, based on source material from Maxim Integrated)

Comprehensive audio subsystem

Designed specifically for mobile applications, the MAX98090 audio codec combines ultra-low-power performance with highly configurable audio signal processing capability. Different combinations of analog and digital inputs can be used to feed Maxim Integrated's FlexSound digital signal processor (DSP) at the core of the device. In turn, the device can deliver the FlexSound-transformed audio to separate output signal paths optimized for different types of audio speakers (Figure 2).

Diagram of Maxim Integrated MAX98090 audio codec designed for in-ear wearables (click to enlarge)Figure 2: Designed specifically for in-ear wearables, the Maxim Integrated MAX98090 audio codec integrates a comprehensive set of input, output, and processing capabilities to provide a complete audio subsystem able to meet hearables’ power and size limitations. (Image source: Maxim Integrated)

The MAX98090 digital audio interface (DAI) subsystem supports sample rates from 8 kilohertz (kHz) voice audio to 96 kHz high resolution audio in a variety of standard pulse-code modulation (PCM) formats. In a typical design, digital audio input would pass directly from the source to the DAI subsystem. For analog sources, the MAX98090 provides a multichannel analog front-end comprising input multiplexers, mixers, preamplifiers, and programmable-gain amplifiers (PGAs). Analog and digital inputs all connect to separate left and right channel mixers, each feeding into dedicated analog-to-digital converters (ADCs). The ADC’s output for left and right channels in turn passes into the DAI subsystem, which ultimately supplies the digital audio to the FlexSound DSP core.

The DSP core provides the essential signal processing functionality needed in audio playback products but typically not supported in traditional audio codecs. Users expect their in-ear wearables to provide sufficient volume to overcome a noisy environment, such as a gym, while still providing a clean audio signal at all volume levels. The MAX98090 FlexSound DSP core meets these requirements with a playback subsystem comprising multiple stages including a separate seven-band parametric equalizer, automatic level control (ALC), and multiple filters for left and right channels (Figure 3).

Diagram of Maxim Integrated MAX98090 audio codec FlexSound DSP core (click to enlarge)Figure 3: At the heart of the Maxim Integrated MAX98090 audio codec, the company's FlexSound DSP core provides dedicated multi-stage paths for processing separate left and right audio channels. (Image source: Maxim Integrated)

These capabilities provide a highly flexible audio processing capability able to meet the varied requirements of each application. For example, besides its seven-band mode, the equalizer can also be enabled for three- or five-band operation that might be needed in products with simpler user interfaces. Similarly, the ALC integrates a programmable dynamic range control (DRC) capability able to prevent both clipping of the audio signal on the high end and amplification of background noise on the low end. For cleaning up digital audio data, the device's digital filter set includes a finite impulse response (FIR) filter for music and high rate audio, as well as an infinite impulse response (IIR) filter for 8 kHz or 16 kHz voice applications. In addition, a DC blocking high-pass filter stage can be included in the FIR music and IIR voice filters to attenuate low frequency sound.

At the output of the DSP core a dedicated digital-to-analog converter (DAC) for left and right channels passes the resulting analog signal to the MAX98090's output mixers. As with its input subsystem, the MAX98090 supports a broad array of audio output configurations and speaker types with its integrated Class D speaker output drivers, Class H headphone output drivers, and configurable Class AB drivers. For each output type, developers simply set associated registers to configure the MAX98090 to drive stereo or mono output from left or right channels to the output driver appropriate for their individual design.

Enhanced low power hearable

For a fitness hearable product, developers would typically configure the MAX98090 to use its Class H headphone output to drive micro speakers, or emerging microelectromechanical systems (MEMS) speakers such as Usound’s UT-P 2017 developed specifically for in-ear applications. In a fitness hearable, digital audio would stream through the Bluetooth connection directly to the MAX98090's digital audio input subsystem. As a result, developers can save power by configuring the MAX98090 to bypass the headphone subsystem's built-in mixer since the analog and line input options would not be required in a base configuration (Figure 4).

Diagram of Maxim Integrated MAX98090 audio codec lower power configuration (click to enlarge)Figure 4: For playback devices such as audio earbuds, the Maxim Integrated MAX98090 audio codec can operate in a lower power configuration that streams digital audio directly to the device's integrated headphone output subsystem. (Image source: Maxim Integrated)

In this configuration, the MAX98090 consumes only about 6 milliwatts (mW). To reduce power consumption even further, the MAX98090 headphone output subsystem can be configured to operate in a special low-power mode that drops power consumption to about 3.85 mW.

To satisfy the typically limited power budget of in-ear wearables, developers can also selectively disable individual input and output blocks in the MAX98090. During idle periods, the device can be programmatically placed in shutdown mode, which consumes only a few microamps. In this mode, the device's I2C serial interface remains active, allowing developers to load new configurations before restarting the device by setting a bit in its shutdown register. At this point, the device returns to full active mode in only 10 milliseconds (ms), providing a nearly instant-on experience for the user.

For the fitness hearable system design, developers would connect the MAX98090 through its I2C serial interface to an ultra-low-power Bluetooth-enabled microcontroller such as ON Semiconductor’s RSL10 (see, "Rapidly Deploy a Battery-Powered Bluetooth 5 Certified Multi-Sensor IoT Device"). The comprehensive set of input, processing, and output blocks integrated into the MAX98090 means that only a few additional components are required to complete this system integration (Figure 5).

Diagram of Maxim Integrated MAX98090 audio codec hardware interface design (click to enlarge)Figure 5: Developers can implement the Maxim Integrated MAX98090 audio codec hardware interface design with only a few additional components. (Image source: Maxim Integrated)

With minimal effort the basic playback design described above could be enhanced to support additional features such as the use of audio input for Bluetooth connected voice assistant interfaces or mobile phone conversations. To capture a user's voice, such a design might use low-power electret microphones such as Knowles’ FG series 50 microamp (μA) microphones, or MEMS analog microphones such as TDK InvenSense’s 25 μA ICS-40310 or Vesper Technologies’ 5 μA VM1010.

With a few additional register settings, developers could configure the MAX98090 to accept sound input from these analog microphones or from digital microphones as appropriate. Separate analog and digital microphone input stages provide the necessary front-end stages for analog signal conditioning or digital control (Figure 6).

Diagram of Maxim MAX98090 complete analog front end and digital interface (click to enlarge)Figure 6: The Maxim Integrated MAX98090 audio codec provides a complete analog front end (A) and digital interface (B) for connecting analog and digital microphones, respectively. (Image source: Maxim Integrated)

Following the input stage, the digitized data stream enters the FlexSound DSP core's separate record subsystem that precedes the DSP playback subsystem described earlier. Like its playback feature, the record feature provides multiple sequential stages of processing. In this case, the processing comprises a set of digital filters including an IIR voice filter, FIR music filter, and DC blocking filter (Figure 7).

Diagram of Maxim Integrated MAX98090 audio codec multi-stage record path (click to enlarge)Figure 7: Along with its analog and digital input support, the Maxim Integrated MAX98090 audio codec includes a multi-stage record path within the company's FlexSound DSP core. (Image source: Maxim Integrated)

The DSP playback system subsequently combines this recorded sidetone with the primary digital audio music stream for further processing and delivery to the MAX98090 output subsystem.

Conclusion

True wireless fitness hearables need to be capable of delivering the extensive functionality needed to meet users' expectations for the latest features, while at the same time working within tight power and size constraints. With respect to audio playback, the Maxim Integrated MAX98090 audio codec combines analog and digital input and output subsystems with a sophisticated audio digital signal processor to provide the comprehensive audio functionality needed in fitness hearable designs. As shown, using the MAX98090 together with similarly optimized SoC devices, developers can build a flexible hardware foundation for sophisticated fitness hearables.

Disclaimer: The opinions, beliefs, and viewpoints expressed by the various authors and/or forum participants on this website do not necessarily reflect the opinions, beliefs, and viewpoints of Digi-Key Electronics or official policies of Digi-Key Electronics.

About this author

Stephen Evanczuk

Stephen Evanczuk has more than 20 years of experience writing for and about the electronics industry on a wide range of topics including hardware, software, systems, and applications including the IoT. He received his Ph.D. in neuroscience on neuronal networks and worked in the aerospace industry on massively distributed secure systems and algorithm acceleration methods. Currently, when he's not writing articles on technology and engineering, he's working on applications of deep learning to recognition and recommendation systems.

About this publisher

Digi-Key's North American Editors