Unlock Audio Black Technology: Tencent Conference And "Pain Point" Of Track

In the cloud conference track, Tencent conference is undoubtedly a dark horse. Since it was officially released on December 25, 2019, Tencent's conference only took 245 days, and the number of users exceeded 100 million.

The rapid development of Tencent conference benefited from the huge demand for online office created by the epidemic. At that time, as a video conference product, Tencent conference solved the problem of remote office for many enterprises. But in fact, Tencent Conference on their positioning, has been far more than a product so simple.

Chen Weidan, conference product manager of Tencent, told the 21st century economic report that in the post epidemic era, people have developed the habit of video conference, so even when they return to the office, they will choose video conference to communicate in many scenes. Therefore, the hybrid meeting of online and offline collaborative meetings will be a standard configuration for enterprises in the future.

However, in the traditional conference room scene, there are many challenges and pain points in the video conference, such as more noise in the conference room, the sound signal attenuation of long-distance microphone pickup and so on.

In view of these pain points, the Teana Laboratory of Tencent conference has recently officially launched the Teana voice module program. Based on this scheme, the video conference can achieve 12 meters ultra long distance pickup, and can eliminate more than 200 kinds of conference noise. In the aspects of double talk and echo cancellation, the scheme also achieves a breakthrough upgrade.

Audio "black technology"

"Tencent Teana" is a real-time audio technology brand released by Tencent in September last year, and Teana laboratory is the research and development team of real-time audio communication and processing under Tencent conference.

Yu Tao, an expert in technical research and development of Teana laboratory, told the 21st century economic report that in the technical path, Teana laboratory has been making continuous noise reduction through the voice signal itself, starting from the near field, making breakthroughs in the middle field and the far field. In this field, the laboratory has accumulated long-term experience.

"On the other hand, starting from the overall situation, in fact, not only hardware, but also circuits, acoustics and algorithms are also included. In short, we adopt a multi-dimensional and multimodal approach to research audio and video," Yu Tao said. "In terms of technical route, we have taken a more comprehensive approach, and will continue to present some black technology in the near future.".

In fact, Teana voice module is one of the black technologies launched by Teana laboratory. It is reported that, for the first time, Teana voice module combines room sound field perception, acquisition and spatial sound field reconstruction technology, and creatively solves the real-time audio and video technical problems such as remote multi person audio communication, multi speaker real-time tracking, multi person automatic viewfinder and other real-time audio and video technology problems in the complex scene of conference room.

For example, in the same conference room, as the distance between people and audio and video equipment becomes longer, the sound will gradually weaken in the process of transmission. Therefore, when participating in enterprise training, work report and other enterprise meetings, the most common phenomenon is that they can't hear clearly or even hear.

In view of this, the Teana voice module scheme embeds multiple MEMS microphone boards into the conference screen, and combines with the audio processor, it can achieve 180 degree wide-angle, 12 meter ultra long-distance pickup, which is equivalent to that the participants can communicate face-to-face even if they are four floors apart.

In addition, based on the deep learning model, Teana voice module can eliminate more than 200 kinds of sudden noises in conference room scenes, such as the sound of air conditioner and fan, the sound of handwriting on the screen, keyboard sound, finger tapping on the table, water cup, turning paper and other conference room scenes.

Yu Tao said that it is not easy to eliminate so much noise. In the process of technology development, Teana laboratory established a noise library by repeatedly testing the scheme in actual user scenarios. With the noise library, it is necessary to optimize the technology to eliminate the noise.

According to the reporter's understanding, Teana laboratory also jointly organized the international conference speech far field competition with the academic and industrial circles this year. It is through this competition that the technology of far-field voice noise reduction is continuously iterated, and the commercial implementation of Teana voice module is achieved.

Give the key to the partner

From the perspective of Tencent conference, in order to better user experience, the threshold of real-time audio and video technology is actually rising, but from the perspective of industry, the threshold of hardware equipment manufacturers to improve conference experience is decreasing.

This is because in the conference room scenario of software and hardware integration, Tencent conference rooms provides software capabilities. At the same time, Tencent conference also clearly stated that it would not do hardware by itself. Therefore, Tencent conference will also cooperate with hardware equipment manufacturers to create intelligent conference rooms.

Shang Shidong, general manager of Tencent Tianlai laboratory, told reporters that Tencent's business model decided not to make hardware for the conference. "Our business model is to make connections and help all kinds of peripherals and users to have a clear and smooth meeting experience in various scenarios.".

In fact, Tencent conference has many mature cases in other industries. For example, chip manufacturers, mobile phone manufacturers, or Microsoft computers, provide a "turnkey" solution that allows partners to directly use their technical capabilities.

So for hardware manufacturers, when they solve the problem of poor user experience in complex acoustic scenes, they can enjoy the most cutting-edge and high threshold technical solutions directly through the technical authorization of Tencent conference. It is understood that the voice of nature solution released this time has been authorized to open to Tencent conference rooms dedicated hardware partners.

In fact, "doing something and not doing something" is an important principle of Tencent in developing b-end business. Tang Daosheng, senior executive vice president of Tencent and CEO of cloud and smart industry group, also said on several public occasions that ecological co creation will be an important strategy of Tencent to B.

Therefore, facing the cloud conference scene, Tencent conference positioning is very clear, that is, focusing on products, technology and other software levels, and through opening up its own technical capabilities, it cooperates with hardware manufacturers to serve enterprise customers.

Shang Shidong said that for the Teana voice module, the first choice is the large screen manufacturer of all in one, because the large screen has more comprehensive functions, including audio, video, interaction, whiteboard, etc., it will be more smooth when integrating the Teana voice module.

According to its disclosure, the first batch of partners of Teana voice module has entered the stage of market promotion. The large screen under mass production includes 86 inches and 65 inches, corresponding to medium and large conference rooms and small and medium-sized conference rooms respectively.

Shang Shidong said that at present, there are still great differences between video conferencing and face-to-face communication. Constantly narrowing the gap is also the goal of Teana laboratory. In its view, the future level of real-time audio and video industry can make the experience of video conference close to face-to-face communication, and even surpass in some fields.

?