Product of the Week: Knowles’ AI-Enabled TWS Development Kit
August 23, 2021
Developing True Wireless Stereo (TWS) solutions and looking for tools that will accelerate development, cut time to market, and give your product that artificially intelligent edge?
For original equipment manufacturers (OEMs), the AI-enabled TWS development kit from Knowles Corporation is designed to do just that. The kit is integrated with features multiple components that enable active noise cancellation (ANC), HD audio, voice command, and AI-enabled conversation enhancement capabilities.
- - Knowles SiSonic MEMS Microphone Arrays
- - An IA8201 AISonic Bluetooth-enabled processing platform with DNN acceleration blocks
- - Voice vibration sensors
- - Independent wireless L and R channels
- - A detachable earbud design, with pre-configured and pre-tuned Knowles earbuds
- - Options for testing different sensor configurations
- - Speaker driver assemblies with Knowles Balanced Armatures
The kit is ideal for applications such as cellphones, smartphones, laptop computers, sensors, digital still cameras, portable music recorders, and other portable electronic devices that thrive off strong wideband audio performance and RF immunity. Per the company, the TWS development kit provides a fast path to achieving HD audio streaming, ANC compatibility, small size, and low power consumption.
The Knowles AI-Enabled TWS Development Kit in Action
The 3.5 mm x 2.65 mm x 0.98 mm Everest SPH0690LM4H-1 digital MEMS microphones, which make up the TWS kit’s SiSonic arrays, come equipped with a high SNR of 69 dB, a high acoustic overload point (AOP) at 130 dB sound pressure level (SPL), and flat frequency response. The SPH0690LM4H-1 also features an acoustic sensor, low noise input buffer, and sigma-delta modulator, and equips a 1.536 MHz clock.
The Voice Vibration Sensor supports hybrid ANC and ambient modes, in addition transparency mode performance. These features are enabled, in part, by Sony Semiconductor Solutions Corporation’s CXD3781 Codec and Knowles’ hybrid driver. Together, these pair with the IA8201 Bluetooth processing subsystem to support both traditional dynamic HD audio woofer and a balanced armature (BA) tweeter configurations.
The voice capture signal chain consists of multi-microphone beamforming, voice vibration sensor processing, and the application of voice algorithms. To execute these "Conversation enhancement" AI algorithms, the IA8201 incorporates a block called AITransparency+ that features hardware-based deep neural network compute blocks capable of more than 100 million AI calculations per second.
This selective acoustic enhancement of conversational speech signals – without perceptible latency – is leveraged by commercial software that comes packaged with the kit such as Alango Technologies’ Voice Communication Package (VCP). This allows voice pickup in TWS devices such as the earbuds included in the kit.
Getting Started with the Knowles AI-Enabled TWS Development Kit
For more information on the “next-generation hearing” partnership between Knowles Corporation and Chatable, watch the “This Week in Hearing” segment featuring Giles Tongue, CEO of Chatable, and Andy Bellavia, Director of Market Development for Knowles Corporation.
- TWS Development Platform Product Page: www.knowles.com/tws-development-platform
- Knowles ANC White Paper: www.knowles.com/docs/default-source/default-document-library/wp05-anc-with-ba_nb0e31d731dff6ddbb37cff0000940c19.pdf
- Multimode Diigital Bottom Port SiSonic SPH0690LM4H-1 Microphone: www.knowles.com/docs/default-source/default-document-library/sph0690lm4h-1_everest_datasheet_.pdf
- Knowles AISonic Audio Edge Processor IA8201: www.knowles.com/docs/default-source/default-document-library/knowles-ia8201-product-brief-final9d761b731dff6ddbb37cff0000940c19.pdf
- Alango Technologies Voice Communications Package (VCP) Product Page: www.alango.com/products-software-vcp.php