Why do we always see lightning before hearing thunder?
As a child the answer might be, the eyes are in front of the ears. After starting to get in touch with physics, the answer becomes that light travels much faster than sound. Even, based on this basic common sense, many physics solution questions were done.
▲ Picture from: unsplash
There is a picture first, and then there is sound. This seemingly irregular phenomenon also commonly appears in many industries.
For example, in the film industry, silent films appeared at the very beginning, only with pictures, and then developed to engrave the sound track on the film, and then there were sound films.
Nowadays, there are some gaps in the development of video (picture) and audio (sound).
4K, 8K, HDR, 10bit and other video specifications are rapidly becoming popular, and almost become the standard equipment for the audio-visual experience of TVs and mobile phones.
The sound is more tortuous and rough. It encountered a choice between a fish and a bear's paw. The fish seems to be "portable", while the bear's paw is "sound quality".
After the TWS true wireless headset trend swept the world, most people compromised the sound quality due to portability. Up to now, the Bluetooth transmission rate that has been developed for many years has not yet fully unlocked the lossless sound quality.
On the just-concluded INNO DAY, OPPO adopted a self-developed Bluetooth audio chip MariSilicon Y (Mariana Y), which completely solved the thorny problem of "portability" and "quality" that lingered on the road of audio development.
High-quality audio, is it necessary?
Or to put it another way, does high-quality audio make sense to the average person?
Previously, the threshold for obtaining high-quality audio such as HiFi and lossless was not low. It required a lot of work on the front end, decoding, transmission, and power amplifier. Only after such a complete set can lossless sound quality be obtained.
Except for some enthusiast players, ordinary people rarely enjoy it. But with the popularization of HiFi, lossless music, and the emergence of miniaturized devices, more and more people have come into contact with lossless high-quality audio.
As the saying goes, "it is difficult to go from luxury to frugality", after comparison, people flock to high-quality music, which is actually somewhat similar to the popularization of 4K and 8K picture quality.
As the consumption level of audio content continues to increase, looking forward to a better audio experience has gradually become an expectation of mainstream consumer groups.
In the "2022 Audio Product Usage Status Survey Report", nearly 60% of the respondents hope to obtain high-resolution and lossless sound quality, and more than 60% of consumers believe that lossless sound quality will determine their purchase behavior.
At the same time, mainstream streaming music platforms are also quietly expanding their lossless music libraries. Spotify has launched Spotify HiFi, and Apple Music has also launched lossless audio in recent years. NetEase Cloud and QQ Music have already launched high-quality audio services such as HiRes (high-resolution audio) and lossless audio, and are constantly expanding their music library.
Whether it is the general trend of the industry or the wishes of consumer groups, the popularization or popularity of high-quality music is a tide of audio development, which is ready to go and only owes the east wind.
And this Dongfeng is how to combine lossless with the TWS true wireless portable that has become the mainstream listening habit.
With the general trend of wireless, the level of Bluetooth transmission rate has become a major factor that directly affects the sound quality. However, compared with the popularity of TWS headsets, the development of Bluetooth standards is a bit slow.
Until now, the public Bluetooth standard has not achieved the minimum rate of 2.25Mbps required for lossless audio.
Even if the mainstream streaming media platforms have a lot of lossless resources, hearing them is a bit against the trend, and still requires a cable and an external DAC (converter).
In view of this, many manufacturers or consumer groups are expecting the emergence of a new public higher-standard Bluetooth protocol to break this speed barrier.
But OPPO's approach is more extreme, directly solving the problem of "lossless sound quality trapped in wired" from the chip and Bluetooth standards.
MariSilicon Y (Mariana Y) brings 12Mbps ultra-high-speed Bluetooth, which can cover the current highest specification of 192kHz/24bit lossless audio, which is 4 times that of standard Bluetooth.
In addition, in order to release the Bluetooth rate of MariSilicon Y, OPPO also customized URLC audio coding technology, with 12Mbps high-speed Bluetooth, it seems that hardware and software cooperate to open lossless audio transmission.
The positioning of MariSilicon Y is a SoC with connection properties. He can be the protagonist on the earphone side, because it is a completely independent chip that can be used with different models of mobile phones; on the mobile phone side, he can also cooperate with the main SoC.
OPPO has broken through the barriers of wireless lossless coexistence through self-developed chips and independent audio coding, and high-quality audio is almost readily available. At this point, the hotly debated question of whether high-quality audio is necessary will be self-defeating after trying it.
MariSilicon Y Advanced Audio Experience, Beyond Wireless and Lossless
In 2021, OPPO will release its first self-developed chip, MariSilicon X.
In 2022, MariSilicon X, which focuses on imaging, will be officially commercialized. Most of the new OPPO products such as Reno and Find have MariSilicon X, which creates a unique OPPO imaging.
The first self-developed chip is aimed at the imaging field, which is actually OPPO's insight into the industry and user needs. It can be said that it not only conforms to the consistent characteristics of OPPO products, but also meets the needs of users for mobile phone shooting capabilities.
The same is true for MariSilicon Y.
While breaking the tradition of wireless and lossless incompatibility, OPPO has also given MariSilicon Y more and deeper uses and layouts.
Similar to the film industry, a little later than computational photography in the image direction, the audio field has also begun to move closer to the field of AI algorithms in recent years.
At present, audio processing is still done by the DSP (Digital Signal Processor) in the main SoC. The current popular "spatial audio" and "active noise reduction" are actually based on this, and the NPU in the main SoC has not been called yet. The ever-improving AI did not participate in the audio processing.
It can be said that before the release of MariSilicon Y, "computing audio" has not really kicked off.
While MariSilicon Y has 12Mbps high-speed Bluetooth and URLC independent encoding, it also has a built-in NPU unit, which is designed to explore or lay out the wave of computing audio.
From the perspective of objective data, its built-in DSP computing power reaches 25GOPS, and the built-in NPU unit computing power reaches 590GOPS, that is, 590 GOPs = 59 billion calculations per second, while the Apple AirPods Pro (first generation) H1 chip has a computing power of about 9 The computing power of GOPS is 65 times that of MariSilicon Y. Both of them are ahead of the same type of products in the audio field, and have a certain redundant design.
In the future, it is difficult to predict what kind of breakthroughs computing audio can bring to the listening experience and sound presentation methods, but it is certain that audio, like pictures, will give more "personality" elements.
MariSilicon Y will use stronger NPU computing power to analyze audio data, separate vocals from specific instruments, and achieve customization of different timbres.
Currently in OPPO's test, MariSilicon Y has achieved the separation of vocals, bass, drums and other four independent tracks.
Compared with the separation of traditional multi-track audio capabilities, the addition of NPU has greatly simplified and reduced the burden.
After separating different audio tracks, MariSilicon Y brought an attempt to "personalize" the sound, and brought custom panoramic sound and universal panoramic sound.
Custom panoramic sound can customize the spatial position of different audio tracks (such as vocals), creating a new "spatial audio", which is somewhat similar to the EQ value previously adjusted for music.
The omnipotent panoramic sound is actually a universal post-production of sound. Effects such as panoramic sound and stereo sound require a unique format in music production, otherwise there is almost no panoramic sound.
MariSilicon Y can use AI computing power to separate multiple audio tracks, re-render the space, and transform it into a new auditory experience of stereo panoramic sound.
As for the strong AI computing power of MariSilicon Y, OPPO still has more room for exploration and imagination, such as separating human voices through AI to uniquely reduce or enhance noise.
MariSilicon Y is actually similar to MariSilicon X. Its super computing power and capabilities are more like an advance layout. Its capabilities are far beyond our current predictions or imaginations about computing audio capabilities.
The iteration of the chip is not like hardware or software. It has a high frequency and often needs two or three years of deployment before it falls into the routine replacement. And OPPO's MariSilicon plan has just started, and it has just begun to lay out.
Introducing some advanced concepts into MariSilicon Y, such as AI computing power, the use of NPU, etc., is actually a rather rigorous and steady approach.
From the current point of view, what MariSilicon Y brings is not only a wireless, lossless advanced audio experience, but also an attempt at computing audio, but more like an advanced layout for the improvement of AI in the audio-visual experience in the next few years.
An example of great success in self-developed chips is Apple. After more than ten years of painstaking efforts in the core-making business, from Jobs to Cook, the chip team has been built bit by bit.
The ecological synergy that has been gradually gathered in recent years and the extreme energy efficiency performance have gradually made the A-series and M-series chips famous.
At the same time, with a considerable amount of shipments, the overall cost is lower than the direct purchase of third-party chips, and the old A-series chips have begun to be deployed in monitors and TV boxes.
In addition to these main SoCs, Apple has also designed and developed a series of small chips for user wearable devices to provide a better experience.
The original intention of OPPO’s core building is also to provide a better user experience and increase the technical moat.
OPPO’s core-making started in just three years, and it is still a stage of gradual accumulation. MariSilicon X is an NPU, which can be regarded as a coprocessor, and MariSilicon Y is here to become a SoC, which also means that OPPO has computing and connection for the first time. Bluetooth SoC platform design capabilities, and it is still very difficult to design RF chips in the industry.
This time, MariSilicon Y adopts TSMC's N6RF process, which is the same as Apple's H2 chip, and is the audio Bluetooth connection chip with the highest process in the industry.
It is reported that OPPO and Apple are the only two technology companies in the world that use RF chip manufacturing technology. After such functions and processes are implemented, the cost of MariSilicon Y is actually very high, and third-party chips are more cost-effective in comparison.
In response to this contradiction, Jiang Bo, senior director of OPPO chips, said, "If you want to provide users with more valuable products, you have to pay a higher cost."
In other words, for OPPO, the value of self-developed chips to bring users a unique experience is much higher than the benefits of lowering costs.
At present, OPPO is not just a mobile phone manufacturer, but has involved quite a few product forms, and is constantly deploying self-developed chips. In fact, in the long run, it is also gradually deploying to the integration of the Internet of Everything.
The chip may not be a cost-benefit, but a "coup" in the product. The potential energy it brings to the product or brand requires a slow process of accumulation and release.
The foundation of the Internet of Everything is actually chips at the bottom layer, and the ultimate experience depends on what capabilities the chips can bring. The birth of MariSilicon Y also represents a further step in OPPO's vertical integration capabilities.
MariSilicon X's mobile phone image is better, so MariSilicon Y is to make the sound of the mobile phone better, and the experience of the mobile phone is improved, which is composed of many better-looking and better-sounding components.
This is the second step of OPPO's self-developed chip. From a small point of view, MariSilicon Y will have important significance for improving OPPO's audio and Bluetooth experience in the next few years, and it is an important bonus item in the segmented field.
Looking at the big picture, OPPO's attitude towards self-developed chips is like playing Go. The winner may be determined after a hundred hands, so this second hand does not mean that OPPO will immediately become the only player in the market.
But the outcome of a game of chess is related to every step, and for chip research and development, there are no masters, only masters.
In an internal speech, OPPO CEO Chen Mingyong said:
When we make chips, we never hope for miracles. If there is a miracle that can be accomplished overnight, then anyone can do it. It is precisely because it is difficult to make good chips that we have done so that we can form an advantage in user experience for a long time. This process is not so easy, it takes time, and there will be many doubts in the process, but we must maintain a normal mind, respect objective laws, and make solid progress step by step. At the same time, maintain long-term optimism and short-term caution. No matter what happens in the future, it makes sense for OPPO to make chips.
To put it bluntly, in the current competition in the consumer electronics industry, it is very simple to pass the competition. It takes hard work to satisfy consumers, but if you want to surpass competitors and exceed user expectations, you must plan three to five years in advance, act two or three years in advance, and persist A long-term effort of ten years or more, and in the process, we must overcome the temptation to take shortcuts to save costs.
In the workplace, those who can achieve success can often overcome short-term temptations and choose to delay gratification. Making chips is similar, there are no mushrooms, only ten-year trees.
#Welcome to pay attention to Aifaner's official WeChat public account: Aifaner (WeChat ID: ifanr), more exciting content will be presented to you as soon as possible.