Home > Article > Technology peripherals > Prinxinchi launches SPV60 series of end-side AI audio weak vision processors to improve the auditory experience of the visually impaired
On August 28, the "3rd Dishui Lake China RISC-V Industry Forum" hosted by VeriSilicon was held at the Dishui Lake InterContinental Hotel in Shanghai. At this forum, Zhuhai Pulinxinchi Technology Co., Ltd. (hereinafter referred to as "Pulinxinchi") officially released a new generation of end-side AI audio/weak vision processor chip SPV60 series
According to reports, the SPV60 series mainly focuses on end-side intelligent audio and weak vision fields. The chip adopts a CPU+NPU+uDSP multi-core heterogeneous architecture, integrates the Anders D25F RISC-V CPU core, has a main frequency of 400MHz, and has built-in Principe's new self-developed uDSP and the third-generation AI neural network processor core NPU , the computing power of NPU reaches 100GOPS. In addition, the chip also integrates high-performance audio AD and DA, as well as a 0V direct drive headphone amplifier module. Among them, the dynamic range of audio AD exceeds 105dB, THD N
The SPV60 series is a multi-core architecture processor with three cores inside. You can flexibly choose among these three cores based on customer-developed algorithms or using our own audio algorithms. It has abundant storage resources, including 512K SRAM and some dual FLSH applications, and the audio link is also very complete. This processor is mainly used in the audio field, including audio call headsets, live broadcast sound cards, wireless intercoms and karaoke, etc. We have developed many mature audio algorithms for our customers, especially algorithms based on our independently developed AI audio noise reduction and echo cancellation, which run very efficiently in our processors. Hu Yingzhe, CEO of Prinxinchi, further introduced
In addition, the SPV60 series also provides a series of complete supporting tools, including: software development kit (SDK), Spacestudio, voice artificial intelligence platform and open source platform
Hu Yingzhe believes that the SPV60 series has strong competitiveness in four aspects:
High-performance audio metrics. The audio indicators are within the professional level range, the AD/DA dynamic range exceeds 105dB, and the THD N is less than 0.001%
2. Powerful computing power. Including a CPU main frequency of about 400MHz and an NPU computing power of about 0.1T, which can support the acceleration of various algorithm operators
3. Various matching algorithms. Artificial intelligence combines traditional noise reduction algorithm, echo cancellation algorithm, howling suppression algorithm and residual sound recognition algorithm
4. Optimized development tools
"Many of our customers sometimes need to use professional DSP chips in the audio field. Our chips are as open as possible based on RISC-V. In this way, customers can quickly and cheaply use this chip in many audio-related fields. Chip." Hu Yingzhe said
According to reports, the SPV60 series released today has already gone through small-batch trial production with customers and will quickly enter the mass production process
In addition to SPV60 chips in audio-related fields, Prinxinchi also has a series of chips, which are mainly used in three fields: (1) Speech recognition products: SPV1108 has the ultimate cost-effectiveness, SPV2016 has high performance, and SPV3042 has ultra-low power consumption. (2) Audio noise reduction products: SPV6040, which combines the noise reduction function of AI technology; SPV6068 has rich interfaces and AI noise reduction functions. (3) Professional audio products: SPV6140 and SPV6240
Although Pulinxinchi was founded in 2018, it has focused on developing RISC-V chips from the beginning. The SPV40 series released now is Prinxinchi’s fourth RISC-V chip
Principal CEO Hu Yingzhe said that regarding the question of why RISC-V was chosen, he said: "What do our customers need? They need to buy Cortex-M4 or even Cortex-M7 computing power at the cost of Arm Cortex-M0 chips, and this is what RISC-V can help us achieve in the field of Internet of Things. In addition to cost, what is more important is the flexibility of RISC-V. Our four generations of chips are all based on RISC-V, and we We have developed many algorithms in the audio field ourselves, so we need customized instruction sets. For example, batch extensions, NAAS instruction sets, etc., which require high flexibility of the CPU. This is exactly what RISC-V does in edge computing Advantages. In addition, many customers require that the development process be very fast. For example, the tools are easy to use, and the entire system may be built with just a quick meal. This allows customers to use our chips in a very fast and low-cost way. In addition, like Many customers are already accustomed to using development and testing tools like Keil, so we work with RISC-V IP original manufacturers to add some of our customized instruction sets to these integrated development environments so that customers can quickly get started."
The content about Pulinxinchi needs to be rewritten
Pulinxinchi is a company established in Zhuhai in 2018 by several PhDs from the United States. The company's chip products mainly focus on the end-side IoT field, including smart sensing and audio processing. Since its establishment, Prinxinchi has always been guided by market demand and based on MCU intelligent sensing and intelligent algorithm core technologies, it has provided competitive and complete solutions in the fields of smart home, traditional home appliances, consumer electronics and other fields. As of 2021, Prinxinchi's self-developed SPT513 series products have successfully entered the domestic first-tier brand supply chain. Its partners include Xiaomi, Honor, Himalaya, Walmart, Transsion Holdings and other well-known brands, with cumulative shipments exceeding 30 million units.
The rewritten content is: Editor: Xinzhixun-Rurounijian
The above is the detailed content of Prinxinchi launches SPV60 series of end-side AI audio weak vision processors to improve the auditory experience of the visually impaired. For more information, please follow other related articles on the PHP Chinese website!