Synaptics voice technology added to Baidu mobile accessory platform
Chinese search engine giant Baidu has selected Synaptics’ AudioSmart far-field voice DSPs (digital signal processors) for its new DuerOS Mobile Accessory (DMA) platform to enable a wide variety of market applications, such as Bluetooth speakers, headphones, wearables, and automotive accessories when combined with a Bluetooth connectivity IC.
The DMA platform was announced at the Baidu Create 2018 AI Developer Conference, and adding voice capabilities with Synaptics’ AudioSmart low-power far-field DSP provides the same features available through the DuerOS mobile phone assistant app to Baidu’s DMA-enabled Bluetooth devices, when they are paired with a smartphone. These features include music streaming, flash briefing, navigation, weather updates, calendar management, and smart home control.
“The DuerOS Mobile Accessory platform is a major update to Baidu’s conversational AI system, and it is through partnerships with industry-leading, innovative companies like Synaptics that we are able to strengthen and grow our partner ecosystem as well as accelerate technological development, further enabling consumers to realize cutting-edge voice-interactive experiences with DuerOS-powered hardware products,” said Donghui Jie, Deputy Manager of AI Ecosystem, Baidu DuerOS Business Unit.
Baidu has also announced the launch of DuesOS 3.0, the newest version of the platform, which supports 25 million monthly active devices and more than 200 partners, making it China’s most vibrant conversational ecosystem, according to the announcement.
“Synaptics is proud to be a long-time partner of Baidu, and has collaborated with them on many far-field voice-enabled devices including the Skyworth X1 Pro sound bar and several smart speakers from brands including Suning, Doss, Linekong and Kuk, and we very much look forward to future engagements,” said Saleel Awsare, Corporate Vice President and General Manager, Audio and Imaging Business, Synaptics.
Scientists with Baidu’s Deep Voice project published research comparing two voice cloning methods which could be used for personalizing interactions between people and machines, or to spoof voice biometric systems, earlier this year.