After the Dimensity 9000 helped MediaTek achieve a phased victory in the flagship mobile phone processor market, it can clearly feel that MediaTek’s confidence in the flagship processor market has greatly increased. At an event this Wednesday, MediaTek gave a comprehensive look at its latest Dimensity flagship technology for 2022, previewing the new features of its upcoming flagship mobile platform.
The most noteworthy is the further upgrade of mobile light tracking technology, and the further upgrade of AI is also worth looking forward to. In terms of connection functions, the upgrades including 5G, Bluetooth, Wi-Fi, etc. are not so surprising, but the high-precision positioning of mobile phones is not so surprising. The experience upgrades that can be brought next are worthy of attention.
Mobile optical tracking is further upgraded, energy efficiency and ecology are the keys to commercial use
In 2018, NVIDIA launched the Turing architecture GPU, which brought ray tracing technology into the public eye in one fell swoop, and desktop games have gradually entered the era of ray tracing. The reason why ray tracing has been limited to a few industries such as film and television and advertising for a long time is the high requirements for hardware. The desktop GPU supports ray tracing, which can undoubtedly bring a better gaming experience to ordinary consumers.
When desktop-level GPUs develop ray tracing, more and more desktop-level game features appear on mobile phones, mobile games have become the choice of more and more gamers, and more and more desktop-level games are run on mobile phones.
If you can play games that support ray tracing on mobile phones, it will undoubtedly increase the attractiveness of mobile phones to gamers. The layout of ray tracing technology is undoubtedly an important development direction of mobile games, and MediaTek, as the underlying technology provider, is no exception.
A year ago, MediaTek released the Vulkan Ray Query technical solution, and jointly realized the first demonstration of mobile real-time ray tracing technology with Arm and Tencent Games.
However, for last year’s plan, industry insiders told Leifeng.com: “It is not practical, and power consumption is a big problem.”
It is impossible for MediaTek to be unaware of the challenges faced by mobile light tracking, so MediaTek further introduced its latest light tracking technology this year, and proposed a mobile GPU efficiency enhancement solution. For GPU performance improvement, with GPU peripheral technologies, to achieve the basis for maintaining system resource balance , to meet the computing needs of high-complexity content.
Zou Wenshan, product manager of MediaTek Wireless Communications Division, said, “We will focus on the Ray Query solution, which can better connect existing Vulkan games in depth. At the same time, this solution is also a better means for mobile devices. , because it is more controllable for the performance and power consumption of the mobile phone platform. We will also match a series of algorithms such as super-resolution or noise reduction to assist in making the content of light tracking run better.”
In fact, through hardware acceleration, light-chasing soft shadows, light-chasing reflection effects, and global illumination technology can be realized, which can significantly improve the picture quality of the whole scene, and the experience of mobile games will be further improved.
Today, the hardware of mobile raytracing is gradually entering the usable stage. In January of this year, the Vulkan 1.3 standard was officially released, supporting the Vulkan Raytracing API, indicating that the raytracing technology of mobile GPUs will accelerate the popularization and cover a wider range of content types. In June of this year, Arm’s latest flagship GPU Immortalis-G715 supported hardware-level mobile light tracking for the first time.
However, in addition to the hardware-level challenges in the application of optical tracking technology, the bigger challenge lies in the software and content ecology.
It can be seen that MediaTek’s Dimensity Star Ecological Laboratory, together with the Gold Label Alliance, and OEM manufacturers, are working together to promote the upgrade of the 64-bit ecosystem, and are also deeply cultivating the Vulkan ecosystem.
Computational photography makes photography and photography more “possible”
Gaming performance and image performance have been the focus of smartphone upgrades in recent years, and are naturally the focus of flagship mobile phone processor upgrades. Computational photography is a keyword that has been frequently mentioned in recent years.
Li Junnan, Director of the Technical Marketing Department of MediaTek Wireless Communications Division, introduced that MediaTek has been emphasizing the importance of computational photography since 2018. The first is the computational photography of CV (Computer Vision), using some algorithms to do multi-frame noise reduction. But in 2020, when it encounters a bottleneck, MediaTek will switch to AI processing. It turns out that the potential of AI is even higher than that of CV, so computational photography is from CV to AI. The latest shared AI image semantic segmentation is already post-level, and it can only be done with detailed post-level flexibility. to debug with preference.
MediaTek focused on introducing the AI Image Semantic Segmentation technology imported from the flagship TV technology to the mobile platform. This technology can perform differentiated image processing for different object characteristics in complex scenes, and can also generate objects for the main and background objects. Special effects can significantly reduce computing power requirements while taking into account both effect and energy efficiency.
It is reported that with the help of AI image semantic segmentation technology, the flagship platform of Dimensity can have more accurate focus tracking. Compared with the traditional focus tracking method, the AI image semantic segmentation technology can automatically distinguish the subject and the background, and the shutter parameters can be accurately set for the subject. In order to ensure accurate focus and make the snapshot clearer.
Of course, AI image semantic segmentation technology also has advantages in video beautification processing. Through the AI scene image quality enhancement technology, the scene content in the image can be recognized in real time, and the contrast, sharpness and color of each frame in scenes such as buildings, landscapes, and portraits can be optimized. . On the other hand, the AI area image quality enhancement technology can also perform object recognition and cutting in real time, and adjust the image parameters of different objects in each frame.
5G new dual access, dual SIM dual standby and then upgrade
For users who use two calling cards on one mobile phone, to this day, missing calls on one card while the other card is disconnected or disconnected from the network is still a pain point faced by dual-card users. Although there are already dual-SIM dual-standby solutions on the market, if two sets of baseband and RF hardware are used, it will be a huge challenge for the mobile phone space.
MediaTek’s so-called 5G new dual-pass is to use only one set of baseband and radio frequency hardware to realize the “dynamic dual-lane” software architecture to realize the new dual-pass of 5G.
It is reported that MediaTek’s 5G new dual-pass has the advantages of flexible architecture and wider coverage, supports more than 100 frequency band combinations, and can cover more network standards, including SA (SA+SA and SA+LTE), NSA (NSA) +LTE), LTE (LTE+LTE), etc. You can use one card to talk, and one card to surf the Internet at high speed. Under the same frequency band of two cards, the full-time dual-communication ratio is more than 30% faster than the dual-communication network speed.
MediaTek’s full-time dual-pass can realize dynamic dual-lane data transmission, supporting more than 30 frequency band combinations, while time-sharing dual-pass can realize “downlink simultaneous + uplink time-sharing” data transmission, and can support more than 70 frequency band combinations.
But how to choose between full-time dual-pass and time-sharing dual-pass? Qiu Yongkai, Product Manager of MediaTek Wireless Communication Division, said, “We provide a framework for any frequency band combination, which can be imagined as a frequency band combination table, on which we can set each frequency band combination as full-time dual-pass or time-sharing. Dual-pass, that’s part of the resilient architecture.”
Of course, for a better 5G experience, Qiu Yongkai revealed that MediaTek will also apply 5G+AI technology to further enhance the experience. “ We also see the potential of 5G+AI. We believe that in some application scenarios, through AI technology, we can achieve better performance than the current non-AI solutions. For the products we are about to launch, we also use 5G+AI. The technology improves the gain of some scenes.”
Talking about the mobile phone satellite communication, which has attracted a lot of attention this year, Qiu Yongkai said, ” There are currently no unified standards for mobile phone satellite communication solutions on the market. We are very optimistic about satellite communication technology, but we hope that it will return to the standard. “
It is reported that MediaTek made a prototype that can be connected to a satellite two years ago. A few months ago, MediaTek also implemented IoT satellite communication.
Wi-Fi 7 is not absent
In 2022, the wireless communication technology that must be mentioned is Wi-Fi 7. Wi-Fi 7 has new features such as higher throughput, lower network latency, and higher network utilization efficiency. Wi-Fi 7 supports advanced 4K QAM (Quadrature Amplitude Modulation) modulation technology, the new MLO (Multi-Link Operation) technology can realize link aggregation, and supports a higher 320MHz bandwidth in the 6GHz frequency band.
It is reported that MediaTek will soon release products that support Wi-Fi 7, including mobile phones and routers.
Bluetooth develops an open ecosystem
Closely related to the experience of mobile games, high-fidelity Bluetooth headsets have also become an integral part of mobile phones. In order to achieve a better audio-visual experience, MediaTek uses Bluetooth high-bandwidth technology and a high-definition Bluetooth audio encoder that can support higher sampling rates, allowing Bluetooth headsets to link smartphones to present high-quality sound.
However, unlike some companies in the industry that develop relatively closed audio transmission protocols and authentication methods, MediaTek actively builds an open ecosystem that can adapt to major mainstream Bluetooth chip platforms.
In terms of spatial audio, MediaTek will cooperate with third parties to provide corresponding functions.
Possibilities of high-precision navigation on mobile phones
MediaTek has been investing resources in the high-precision navigation of mobile phones and continuously improving the experience of mobile phone navigation.
In 2022, MediaTek will focus on its MPE (MEMS-sensor Positioning Engine) fusion technology, which uses micro-electromechanical sensors (accelerators, gyroscopes, magnetometers) to integrate global satellite navigation systems, which can be widely used in driving, walking and running. Daily navigation application scenarios for smartphone users.
With the help of MPE technology, it can still provide accurate navigation in areas with poor mobile phone signals such as tunnels, viaducts, and urban buildings.
The reason why it can achieve high-precision navigation without signal is that first, when positioning outdoor GNSS, there is a calibration value to locate, and use the calibration value to calibrate the sensors in the mobile phone, including accelerometer, gyroscope, magnetometer, etc. it is good. When there is no satellite signal when entering the room, MPE and sensors can be used to continuously provide users with stable navigation output positioning points.
However, if the magnetometer cannot be recalibrated when there is no satellite signal for a long time, it will switch to the traditional NLP (Network Location Provider) technology, and the accuracy will be reduced.
In fact, the high-precision positioning of mobile phones has already had some practical applications in common navigation, such as lane-level navigation on highways in some cities. The lane-level navigation realized with the help of high-precision navigation can reduce the number of drivers taking the wrong way. Case.
This article is reprinted from: https://www.leiphone.com/category/chips/ZmthJXHQ3VfhPaEJ.html
This site is for inclusion only, and the copyright belongs to the original author.