Can deep learning subvert video codecs?

Yunzhongfa from the Gafei Temple

Quantum position | Public account QBITAI

This fat friend, you have been raised by 4K or even 8K high -definition live broadcast, can you still go back to 1080P?

Especially since this year, what Winter Olympics, the launch of the space station experimental compartment, the clarity is really high. Major platforms also want to write 4K as a standard directly on the face.

Seeing that the 1080P is about to become a yellow flower tomorrow, but I don't know if you thought about such a question:

The clarity of the video is getting higher and higher, but the bandwidth of our video has not increased much?

This involves such a technical keyword--

Audio and video codec technology.

In fact, in order to cope with videos in the 5G era to mention excessive problems and achieve higher -definition, more immersive, lower -delayed interactive experience, the audio and video industry has been updating and iteration in standard encoders such as H.266, AV1 Essence

Not only that, today's audio and video codec technology also takes the Dongfeng of deep learning to explore a new path of new and faster development.

So, is it possible for deep learning to bring disruptive changes to the audio and video industry in the future? With the increasing maturity of artificial intelligence technology, how will the future of audio and video technology develop?

If you are interested in this issue, now, there is a chance to open the little red book and listen to the top experts in the domestic audio and video field.

Technical big coffee sharing

At 19:00 on the evening of July 27, the Xiaohongshu Technical Team will invite the first prize winner of the National Technology Invention and the associate researcher at the School of Computer School of Peking University to participate in the [Redtech Come] technology live broadcast, and "Now and the future" sharing themes.

In 2020, Wang Yishe won the first prize of the National Technology Invention Award for key technologies with ultra -high -definition video polymorphic base. He currently works in the National Engineering Laboratory of Digital Video Code and Patriotic Technology, and has hosted a number of national key research and development plans such as "efficient coding standards for 8K ultra -high -definition video applications" and other countries.

In addition, Master Wang Yishe was from academician of the Chinese Academy of Engineering Gao Wen. Academician Gao Wen devoted himself to studying the key technologies of the video codec for 19 years, led the formulation of the AVS series standards with independent intellectual property rights. The National Engineering Laboratory of Digital and Video Crimination Technology he is responsible has led my country's digital video industry to achieve a huge leap from core technology gaps to internationally leading, and has transported a large number of technologies for the development of the industry.

In addition, Chen Jing, the person in charge of Xiaohongshu's audio and video architecture, will take Xiaohongshu's practice as an example to introduce the landing challenges, innovation practice and development exploration of audio and video technology in the field of industrial application.

As of October 2021, the number of active users of Xiaohongshu exceeded 200 million, with a wide, real, high -time user scene, massive audio and video notes, and rich and complex users to watch behavior data.

Chen Jing has served as a 51Talk chief audio -visual scientist, Google Chrome Media video group engineer, MindSpeed ​​video department senior DSP engineer, Huawei Central R & D department video group engineer, and has published works "In -depth Understanding video compilation technology" and translation "Computer Visual Algorithm: Based on OpenCV's Computer Application Development ".

Wang Yishe and Chen Jing will also start in -depth talks on topics such as the development dilemma of audio and video technology, the possibility of deep learning and audio and video technology:

What are the differences in industrial and academic circles in the exploration of new audio and video technology?

Will deep learning subvert video codecs?

Where is the next air outlet of the audio and video?

Who will define the new generation of codec standards?

Will the pain points such as fluency and massive concurrency ushered in the technical turning point?

What kind of video compression algorithms have been brought in the pursuit of ultra -high -definition?

In addition, at present, the recruitment plan of the top technical talent of Xiaohongshu Redstar has been officially launched. During the Redstar project, Xiaohongshu will host a number of cutting -edge technology live broadcasts.

"Redstar" is the exclusive talent recruitment plan initiated by Xiaohong in writing to the top technical talents in the world. It is committed to providing Redstar with industry competitive salary and benefits, exclusive training programs with sincere sincerity, and practical channels for promoting young people.

In addition to the audio and video topics on the evening of July 27, at 19:00 on August 5th, Xiaohongshu will also invite Zhang Yue, associate professor of West Lake University, and Pang Liang, assistant researcher at the Chinese Academy of Sciences Computing Institute of Calculation Institute. The cutting -edge and practice of language recognition ".

How about it, at 19:00 on July 27th, did the second stage live broadcast on the second phase of "Redtech is here"?

- Finish -

- END -