In the speech recognition market, according to the market research firm Strategy AnalyTIcs, by 2022, it is estimated that there will be more than 62 million devices in the global consumer market with personal voice assistants. In order to insert this flag into the market, the semiconductor industry has launched a new generation of solutions. For example, Infineon combines radar, MEMS microphones and audio processors to further enhance the speech recognition performance of MESM microphones, while STMicroelectronics is the developer of voice interface and keyword detection algorithms - Sensory, and communications. DSP Group, a wireless chipset solution provider, is jointly developing efficient speech detection processing microphone technology.
Lv Pei, an analyst at the IEK Electronics and Systems Research Group of the ITRI, said that although the voice assistant is still in its infancy, the market will penetrate into smart appliances, in-vehicle systems and even more Internet of Things devices. In the application. At present, semiconductor manufacturers, telecom operators and other operators who want to enter the speech recognition market, in addition to actively developing related applications, how to improve speech recognition performance is also the key development direction in the future.
To this end, the semiconductor industry has accelerated its layout. Infineon combines radar, silicon microphone sensors, and XMOS audio processors with audio beamforming and radar target presence detection to provide far-field voice capture to ensure a wide range of voice control devices The ideal voice recognition to perform digital voice assistance, and thus improve the multi-person conversation, the MESM microphone can not accurately identify the location of the voice source, and can not be separated from the noise emitted by the object.
It is reported that Infineon's 60GHz 2Tx/4Rx radar IC with the accompanying antenna and 70dB SNR microphone will help overcome these obstacles. The microphone uses the company's dual-backplane MEMS technology for far-field speech capture and beamforming. In addition, the SNR of the microphone is improved and the performance will be further enhanced. The XMOS audio processor analyzes the signal data from the company's digital microphone array and adjusts the angle and distance data provided by each microphone. Form a beam. Through the combination of radar and XMOS beamformer, the microphone can accurately target the target to a specific object even if the object is moving and has a fuzzy noise.
On the other hand, the idea is to work with DSP Group and Sensory to publish a technology for efficient speech detection and processing of microphones. The component integrates ST's low-power MEMS microphone, DSP Group's ultra-low-power speech processing chip, and Sensory's speech recognition firmware in a tiny system-in-package (SiP), which is lightly achieved through the company's advanced packaging technology. Measurement package, long battery life and advanced features.
At the same time, the new microphone uses DSP Group's HDClear low-power audio processing chip, which can greatly reduce energy consumption and extend the battery life of battery-powered devices for many years without charging or replacing the battery; and the system immediately executes commands without prior identification commands. Voice commands will respond more quickly.
Chamfered Radiator,Galvanized Chamfered Radiator,Corrosion Resistant Chamfered Radiator,Power Transformer Chamfered Radiator
Shenyang Tiantong Electricity Co., Ltd. , https://www.ttradiator.com