According to PR Newswire, being fueled by the rising demand for voice enabled devices, the global smart speaker market is likely to grow at a CAGR of 43.68% in terms of unit shipment during 2016–2022.
Early innovators as well as market leaders of these newly marketed products are demanding extremely high standards of sound quality, with minimum noise along with a complex combination of 4 to 8 channel ADC for spatial beamforming applications. Fabless provider and market leader Mnano, has chosen Dolphin Integration for its innovative generation of multi-channel audio A/D Converter, YANZI-A-hexADC-VD.01, which combines an embedded Voice Activity Detector with a complete set of audio features.
In Wireless applications such as Smart speakers, SoC architects must deal with the complicated issue of power consumption. A standard practice is to maintain the audio components in active mode, listening for voice almost 95% of the time, thus reducing precious battery lifetime.
Dolphin Integration’s patented Voice Activity Detector, WhisperTrigger™, when embedded in the audio converter consumes a mere 25 µA in the Always-On. This is a resounding success in achieving negligible power consumption.
The combination of a high dynamic range of 100 dB, together with an embedded linear regulator, ensures the best far field voice recognition and resilience against noise caused by audio, master clocks or by the power supply.
“Meeting the right balance between silicon costs, audio quality and low power consumption is a challenging task” stated Mr. Guoxin ZHANG, Dean at Mnano. “The YANZI-A-hexADC-VD.01, a feature-rich audio ADC from Dolphin Integration, packs all these advantages into our smart speaker offering”.
In a voice-enabled device apart from high power consumption, the number of channels adds challenges on channel synchronization. Indeed, most of the speech recognition solutions use advanced algorithms such as beam-forming and echo cancelation and they may be sensitive to the phase mismatch between the channels, resulting in degradation of the performance of the algorithm. Thanks to the phase alignment feature, fabless companies can be rest assured of an exceptional sound quality for end users.
“We are highly confident that the combination of Dolphin Integration’s Multi-channel A/D Converter and Mnano’s expertise in system design will result in advanced SoCs capable of meeting the stringent demands of the next generation of Wireless Smart Speakers” said Frederic Renoux, Executive VP Marketing and Sales at Dolphin Integration.
“We continuously strive to help companies like Mnano benefit from truly cost-effective silicon IPs which can meet the performance requirements of new generation audio devices. “
Shenzhen Micro-nano, Integrated Circuit and System Application Research Institute, was established in February 2013, in Shenzhen. It has received the strong support of Shenzhen’s semiconductor backbone enterprises, a key technical industrialization platform for the close cooperation between government, industry, universities and research institutes. It has actively developed key cutting-edge technologies in the field of integrated circuits and perceptual computing and profound technical accumulation in areas such as voice perception and machine vision perception. Since 2016, the Micro and Nano Institute has collaborated with the National IC Designing Industrial Base in Shenzhen to undertake the national core fire platform and support the innovation and entrepreneurship ecosystem of integrated circuits.