advancedchip
New member
Kadeng Electronics announced today that Wuqi Microelectronics (Chongqing Wuqi Microelectronics Co., Ltd.) has successfully applied Cadence Tensilica HiFi 5 DSP to its True Wireless Stereo (TWS) products. TensilicaHiFi 5 is the first IP core optimized for high-performance far-field processing and artificial intelligence (AI) speech recognition processing. Compared with the previous generation HiFi 4 DSP, the audio processing performance of the fifth generation HiFiDSP will be significantly increased 2 fold, and the neural network (NN) processing performance will be improved 4 times. It is an ideal choice for voice control user interfaces such as digital home assistants and car entertainment systems.
Among Wuqi’s TWS products, all audio codec software including CVSD/MSBC/SBC/AAC/LC3/OPUS run on Cadence TensilicaHiFi 5. In addition, Wuqi’s audio pre-processing algorithms have also been ported to TensilicaHiFi 5, including voice activity detection (VAD) based on 1, 2 or 3 microphones, keyword recognition (KWS), and adaptive active noise reduction (Adaptive ANC) , environmental noise reduction (ENC), AI noise reduction, automatic echo cancellation (AEC), automatic gain control (AGC), and segmented equalizer (EQ) for music/voice enhancement. The HiFi5 Neural Network (NN) engine can well support Wuqi's intelligent voice applications. At the same time, it can also support Wuqi's complex noise suppression (NR) algorithm and achieve very low power consumption.
With TensilicaHiFi5IP, when Wuqi’s TWS products are using SBC encoding for music playback, the current will be reduced to less than 3mA under 3.8v voltage conditions. In case of voice call and ENC and AI noise reduction running at the same time, the current will be reduced to less than 6.5mA under the condition of 3.8v voltage.
"The ultra-high-performance, ultra-low-power audio SoC launched by Wuqi provides an advanced and feature-rich solution for the fast-growing TWS market." Liu Yipeng, Director of Product Marketing Group, Tensilica Audio/Voice IP Division, Cadence Said, “HiFi 5 DSP, which is highly optimized for audio, voice and AI processing, can provide high-performance computing, ideal choice to further enhance the rich hearing experience."
“The demand for music enhancement/noise reduction/voice AI applications is endless. At the same time, TWS headset applications have very stringent requirements for low power consumption,” said Hao Lin, CTO of Wuqi Microelectronics. “TensilicaHiFi-5’s powerful audio processing capabilities provide rich possibilities for multi-microphone noise reduction/voice wake-up/adaptive ANC algorithm/sound enhancement algorithm, etc., while maintaining a very low system frequency to save system power consumption. Applying Tensilica HiFi-5, Wuqi TWS’s music playback current can be as low as 3mA. At the same time, Tensilica’s powerful and robust ecosystem is also a strong guarantee for us to quickly implement algorithms on TensilicaHiFi-5.”
TensilicaHiFi 5 DSP supports Cadence Intelligent System Design (Intelligent System Design) strategy, helping to achieve SoC design of excellence.
About the Author
Mark Chen is a China based freelance IC designer, IT and web developer, and writer in technology, personal and family stories, world affairs, and other topics. He served GE Alstom, Siemens and other multinationals in his earlier career. His current career focuses on IC technology and business, incl silicon IP core design (mainly modulator and demodulator), IC technology development reporting (the “World IC Technology Development Monthly Report”), IC and industry portal development and IC technology authoring (the book titled “Introduction to IP Core Design”). For details refer to his personal site http://markchen.int.eu.org/ , his site of IC expertise http://angeliamodem.eu.org/ and his IC business site under http://www.angelia.eu.org/.
Among Wuqi’s TWS products, all audio codec software including CVSD/MSBC/SBC/AAC/LC3/OPUS run on Cadence TensilicaHiFi 5. In addition, Wuqi’s audio pre-processing algorithms have also been ported to TensilicaHiFi 5, including voice activity detection (VAD) based on 1, 2 or 3 microphones, keyword recognition (KWS), and adaptive active noise reduction (Adaptive ANC) , environmental noise reduction (ENC), AI noise reduction, automatic echo cancellation (AEC), automatic gain control (AGC), and segmented equalizer (EQ) for music/voice enhancement. The HiFi5 Neural Network (NN) engine can well support Wuqi's intelligent voice applications. At the same time, it can also support Wuqi's complex noise suppression (NR) algorithm and achieve very low power consumption.
With TensilicaHiFi5IP, when Wuqi’s TWS products are using SBC encoding for music playback, the current will be reduced to less than 3mA under 3.8v voltage conditions. In case of voice call and ENC and AI noise reduction running at the same time, the current will be reduced to less than 6.5mA under the condition of 3.8v voltage.
"The ultra-high-performance, ultra-low-power audio SoC launched by Wuqi provides an advanced and feature-rich solution for the fast-growing TWS market." Liu Yipeng, Director of Product Marketing Group, Tensilica Audio/Voice IP Division, Cadence Said, “HiFi 5 DSP, which is highly optimized for audio, voice and AI processing, can provide high-performance computing, ideal choice to further enhance the rich hearing experience."
“The demand for music enhancement/noise reduction/voice AI applications is endless. At the same time, TWS headset applications have very stringent requirements for low power consumption,” said Hao Lin, CTO of Wuqi Microelectronics. “TensilicaHiFi-5’s powerful audio processing capabilities provide rich possibilities for multi-microphone noise reduction/voice wake-up/adaptive ANC algorithm/sound enhancement algorithm, etc., while maintaining a very low system frequency to save system power consumption. Applying Tensilica HiFi-5, Wuqi TWS’s music playback current can be as low as 3mA. At the same time, Tensilica’s powerful and robust ecosystem is also a strong guarantee for us to quickly implement algorithms on TensilicaHiFi-5.”
TensilicaHiFi 5 DSP supports Cadence Intelligent System Design (Intelligent System Design) strategy, helping to achieve SoC design of excellence.
About the Author
Mark Chen is a China based freelance IC designer, IT and web developer, and writer in technology, personal and family stories, world affairs, and other topics. He served GE Alstom, Siemens and other multinationals in his earlier career. His current career focuses on IC technology and business, incl silicon IP core design (mainly modulator and demodulator), IC technology development reporting (the “World IC Technology Development Monthly Report”), IC and industry portal development and IC technology authoring (the book titled “Introduction to IP Core Design”). For details refer to his personal site http://markchen.int.eu.org/ , his site of IC expertise http://angeliamodem.eu.org/ and his IC business site under http://www.angelia.eu.org/.
Last edited: