In the past two years, it may be the most "fame and fortune-winning" chip manufacturer in the smartphone market.
On the one hand, according to data released by Counterpoint Research not long ago, MediaTek achieved the eighth consecutive quarter of global smartphone AP/SoC chipset shipments, and its market share was as high as 39%.
On the other hand, looking at the entire market today, it is not difficult to find that from the entry-level Dimensity 700 and Dimensity 810, to the "middle-level God U" Dimensity 8000 and Dimensity 8100, to the flagship Dimensity 9000 and Dimensity 9000+, MediaTek's 5G SoC has now achieved full coverage of the entire market. Especially in the flagship field, top products such as vivo X80 Pro and Asus ROG Phone 6D show that leading manufacturers recognize the product strength of "Fa Ge" and " Dimensity ".
If you ask an end user, "Why do you think MediaTek 5G SoC is strong today?" Then the answer I got is probably "Medtech chose to have more advanced processes, the processor energy efficiency is higher than that of , and it is more power-saving to play games."
Obviously, this also very subjectively reflects the perception of MediaTek SoC today. But from another aspect, is the improvement of "Fa Ge"'s current product strength and market performance just because of the "correct choice" of process and architecture?
It is obvious that the truth is not the case. With the convening of MediaTek's media communication meeting with purely "technology" theme in mid-October 2022, we can also get a glimpse of this current largest mobile SoC manufacturer, a new flagship product experience that will be launched in many aspects in the future.
self-developed ray tracing technology "turned into positive", hardware ray tracing GPU is already on the road
As we all know, when the Dimensity 9000 was released, MediaTek actually announced their self-developed mobile game ray tracing technology, and even jointly with some game manufacturers to launch some DEMOs that can actually operate.
However, everyone knows that "mobile ray tracing" has not become the main theme of the mobile game market in 2022. On the one hand, this is related to its high development cost and the pressure on the device; on the other hand, MediaTek's mobile ray tracing technology was based on the Vulkan Ray Query software solution, rather than the traditional GPU ray tracing based on hardware acceleration. It is somewhat too advanced in terms of concepts, and the effect is indeed not as extreme as the ray tracing on PC.
But the advanced R&D investment was obviously not in vain, because just today, MediaTek once again emphasized their firm position on "mobile ray tracing". However, this time's technical layout is much more mature in all aspects than when it was announced last time.
First of all, in terms of basic industry specifications, MediaTek chose this year's latest Vulkan 1.3 standard. It formally includes the Vulkan Raytracing API dedicated to ray tracing. The Raytracing API and Ray Query, which was developed by MediaTek last year, can achieve seamless technological connection. Therefore, in terms of the most basic software technology, MediaTek has conducted in-depth research "in advance" and has had certain advantages from the beginning.
Secondly, in terms of hardware design, MediaTek specifically mentioned the latest and strongest flagship GPU solution of ARM, today. Obviously, this is equivalent to directly indicating that the next generation of flagship Dimensity platform will be equipped with Immortalis-G715 GPU to achieve hardware-level optical tracing acceleration support. From this point of view, it is not difficult to find that the new Dimensity flagship platform is likely to focus more on gaming performance than the current model, especially to strengthen support for high-quality mobile games, thereby effectively challenging competitors' position in the flagship market.
Finally, under the premise of synchronous evolution of software and hardware, MediaTek also specifically mentioned the "mobile GPU efficiency enhancement solution". It not only includes software and hardware scheduling technology derived from Vulkan and ARM, but also incorporates MediaTek's self-developed AI super-scoring, game frame insertion and other functions.The purpose is to ensure excellent energy efficiency ratio while turning on "ray tracing" and reduce the lag and heat of the mobile phone. From this point of view, the next generation of MediaTek Dimensity flagship platform will not only have significantly improved 3D, especially ray tracing performance, but also be expected to continue its current reputation in actual game testing.
flagship TV technology was introduced into mobile phones just for clearer images
Some friends may know that MediaTek is not only mainstream in the mobile phone chip field, but they are currently the absolute leader in the field of high-end TV main control. In fact, behind many leading manufacturers including Samsung and Sony , MediaTek flagship TV SoC and related technologies can be seen.
For example, among the latest flagship TV products in 2022, many brands have launched the new image optimization function of "AI depth of field". It can automatically recognize the subject of screen by and only optimize the details of the subject, so that the viewer's visual experience is more natural.
. On MediaTek's subsequent flagship platform, the same underlying technology will be introduced. But this time, this technology called "AI Image Semantic Segmentation" will not be used for mobile phone video playback screen processing, but is mainly reflected in the real-time optimization algorithm of video shooting.
Why is video shooting? You should know that for modern top mobile platforms, the automatic optimization technology of single-photos based on AI is very mature, but no matter how complex the mobile phone's photography algorithm is, it can only process multiple frame synthesis of dozens of photos at a time, and then correct and strengthen the details of AI scenes, colors, textures, etc. on a single picture. The calculation amount is not particularly large.
However, video shooting is different. If you still use the idea of processing photos to synthesize and reduce the noise before and after frames of each video, and enhance the image generated in each frame with complete AI details, color, and texture, then the calculation amount will not only be very large, but will also accumulate as the video shooting continues. In the end, it will either cause the mobile phone's computing power to be overwhelmed and stutter, or it will cause the processor power consumption and heat generation to become super high during video shooting, and even cause the mobile phone to overheat and force the camera to be turned off.
With AI image semantic segmentation technology, the situation is completely different. With this algorithm originating from flagship TV, the future MediaTek SoC can automatically identify different parts in the sky, background, characters and other pictures when shooting videos, thereby more intelligently calling AI algorithms of different complexities for different parts. In this way, the power consumption of real-time video AI enhanced shooting can be greatly reduced.
Not only that, using this technology, MediaTek SoC can even provide additional "blessing" to the original focus performance of the camera. Because it can now accurately distinguish the "subject" and "background" in viewfinder , so that the algorithm can actively "guid" CMOS to find the correct focus position. Obviously, this will make great sense for scenes like macro photography where the physical distance between the foreground and the background may be small ( ranging sensor does not work).
Several leading communication technologies help to provide all-round flagship experience
At the end of the normal technical communication meeting, MediaTek also previewed many of their latest achievements in basic communication technology in smartphones.
First in the field of G communications, MediaTek, as the earliest manufacturer to implement 5G "dual SIM dual standby", also released their "5G new dual connectivity" solution today.
Unlike traditional dual standby or dual pass technology, MediaTek's "5G new dual pass" does not require two sets of high-cost, large-volume two sets of basebands and two sets of RF hardware. Just a single set of basebands and RF can achieve the experience of dual card 5G on the network, one card calls and another card without missing connections, and continuous networking.
At the same time, due to the self-developed "dynamic dual-lane" software architecture, the "5G new dual-pass" can support more than 100 global frequency band combinations, and can be compatible with various network standards including SA, NSA, LTE, etc., and can also increase the dual-slot dual-pass network speed by more than 30%.
Secondly in the WiFi and Bluetooth fields, MediaTek has prepared mobile, router, and device (headphone) solutions for WiFi7 and the next generation LHDC V. It is reported that compared with WiFi6, WiFi7 will be able to increase network bandwidth several times, reduce latency by 20%, and significantly improve network interference anti-interference capabilities.
, and LHDC V Bluetooth lossless audio encoding, as the latest generation of industry open standard, will bring Hi-Res wireless audio transmission capability of 24bit/192KHz, and at the same time, the audio encoding rate will be increased to more than 1000Kbps, surpassing the existing competitor's 990Kbps level, thereby achieving richer Bluetooth audio details.
Finally, MediaTek also announced the MPE (MEMS-sensor Positioning Engine) fusion technology for the next generation of high-precision lane navigation, precise motion tracking and other scenarios. It combines the accelerator in the mobile phone, gyroscope , magnetometer and satellite positioning system , which can be used in scenarios such as running tracking, driving tunnel navigation, urban building navigation, etc., effectively improving the accuracy of mobile phone positioning when satellite signals are poor, and greatly improving the user experience of the new flagship models in various personalized navigation scenarios.
Summary: "Fa Ge" who understands technology will inject more vitality into the market
Yes, MediaTek did not release any actual chip or reference design today. But judging from the published technologies, they are equivalent to having already made a relatively complete "spoiler" with many features of the new generation of flagship platforms.
It is not difficult to find that compared with the current Dimensity 9000 and Dimensity 9000+, focusing on the "high energy efficiency ratio" in various scenarios will still be the most important design direction of the new Dimensity flagship platform. But in addition, MediaTek has also become more active in optimizing the functional experience of the new platform in scenarios such as extreme gaming, extreme audio and video, and high-end business.
This means that from the product side, it may be expected to see more outstanding segmented product forms such as top-level imaging models, hard-core gaming phones, and even folding screen models on the next generation of Dimensity flagship platforms. As for user experience and consumers' "impression", the next generation of Dimensity flagship platform is also expected to make everyone realize MediaTek's strength in underlying technology, ecological chain integration, and various detailed functions.
Perhaps at that time, when people talk about MediaTek and Dimensity, they no longer think of as simple as "advanced architecture and good process selection", but they will relish the various functional advantages based on MediaTek's self-developed technology. This will obviously bring more choices to the current booming high-end and top smartphone market and inject more market competitiveness.
