Computing is the foundation of intelligence
Calculations are the foundation of intelligence and the basis of AI. The Junzhen T30A is aimed at landing the front-end lightweight AI algorithm. The T30A provides a 1.2GHz clocked CPU with a 128-bit vector acceleration engine at the same frequency, and 8-bit tensor. Calculations have been fixed-point optimization. Under such core parameters, the T30A can provide relatively high computational power for visual and audio algorithms. At the same time, the software layer package opens the OpenCV/WebRTC optimization operator, and the underlying ABI of the open vector engine.
Thanks to the appropriate computing power of the T30A, some front-end lightweight AI algorithms can be quickly and cost-effectively landed. For example, the license plate recognition camera, in addition to the main functions of the video, the T30A can also support single-chip support for low-speed license plate recognition. Manufacturers are gradually landing, applied to the general access gates and inside the parking lot.
At the same time, audio and video fusion calculation is also a current mainstream trend. The Junzhen T30A is a chip platform that professionally solves audio and video fusion calculations. It has several core features. First, professionally, the Junzhen T30 has professional imaging. Video encoding, audio computing and other capabilities, which have ultra-low stream H.265/H.264 encoding engine, professional high-quality ISP, up to 5 million resolution, support CMOS sensor of any size pixel; secondly in terms of fusion, Jun The T30A is capable of landing smart audio plus smart video applications on a single chip. It supports digital microphone arrays, supports dual-microphone/three-microphone/four-microphone combo, and has built-in analog microphone audio codecs. It can expand external audio codecs through I2S. In front-end lightweight AI computing capabilities, the T30A still performs strongly for front-end embedded AI. Finally, in applications, the Junzhen T30A is different from other AI concept chips. The T30A has many advantages such as cost, power consumption, and engineering. The product is very easy. Floor-standing. T30A's ultra-low operating power, as low as 300mW, also supports on-screen display and fast boot, cold start can be completed within one second.
Professional video capabilities
The T30A is based on a professional video processing chip. Its built-in high quality ISP can support 3A/3D-NR, support WDR, low light night vision and full color. T30A supports H.265/H.264 encoding, H. The 265 is currently one of the most efficient coding methods. With a variety of bit stream control methods, it can provide the most appropriate video output for a variety of application scenarios, from extremely high quality, to sports recording, to network transmission, and then to many The stream can output at the same time.
The T30A can support up to 5 million resolution sensor inputs, and up to 5 million resolution video processing. This extremely high-definition resolution provides a wide range of video access options for a variety of applications. It is distinguished from toy-grade light video 1M. Rate, to 2K * 2K HD fish eye; Professional security from 3u pixels, to 1.4u mobile video recording, can support.
Innovative audio capabilities
Currently the industry's video processor chip, audio function is generally weak. Junzhen T30A has made a directional enhancement of audio capabilities innovatively. Video-centric, is the foundation of the Junzhen T30A audio and video fusion calculation. Professional video processing capabilities, For pure audio smart plugs powerful wings.
The T30A first innovatively integrates a digital microphone array interface and can support multiple array combinations such as dual-microphone, triple-microwave, and quad-microwave. Based on the digital microphone array, the T30A can be used to get close to the recently-popular 'Little Lovemates'. 'Tmall Elf' and other speaker products voice capabilities. Jun is a joint of the industry's well-known supplier of voice algorithms, the DOA, NR, AEC, Beamforming, De-Reverbration and other integrated in place, directly provide 'shaking machine sound source positioning ''Command word control'' and other functions.
T30A contains audio codec, and can pass the I2S interface to expand the level of audio codec, to achieve fever-level audio quality.
From these unique characteristics, Junzhen T30A is undoubtedly a professional video processor with strong audio potential. Based on these capabilities, a series of products that require video and audio fusion are ready to appear. For example, a speech recognition camera based on the original smart camera. Add voice commands, sound source tracking and other functions. In addition, educational robots and children's story machines that are currently being developed by several brands are adding remote video capabilities to the original audio-only function, which has made the product leap to a level.
Differentiation of innovation to stimulate smart video market
In addition to audio, the T30A adds another series of interesting and differentiated innovations to the classic camera. It provides a versatile video processing platform for the broad pan-video market.
The T30A is compatible with the standard UVC/UAC protocol set. It can be used together with H.265 to create a new high-end USB camera. Due to the versatility of USB, the USB camera can not only be connected to a PC for plug-and-play, but also to a set-top box, Android/iOS device. As an enhanced video input device.
In addition, the screen displays, quick start within one second, power consumption of video operation as low as 300mW, these characteristics provide innovative products, cross-border products offer possibilities beyond imagination, such as interactive cameras, fast-start recording products, mobile Video recording, network videoconferencing terminals, can all bear fruit on the T30A platform.
The Junzhen T30A is the first audio/video convergence computing chip that provides a comprehensive, general-purpose video processing platform for the broad pan-video market. Focusing on audio and video convergence and targeting the front-end lightweight AI calculations, the Junzheng T30A Can cover a variety of applications. In general, in the license plate recognition camera, voice recognition camera, interactive screen camera, educational robots, children's story machines, USB cameras, video conferencing terminals and other products and areas, Jun Zheng T30A will have a great deal.