This article is from WeChat official account:rct studio, original title: “Content Generation and Consumer Technology: Media, Games and Social”, the title picture comes from Visual China
From the perspective of content generation, this article discusses the relationship between information, digital content and people, as well as the development logic, presentation and future forms of content in the fields of media, games, and social interaction.
When we look at the entire content production and consumption in a holistic way, there will be 4 different content generation stages.
In the first stage, the content is produced by a large-scale professional creative team(PGC), most of these products are for single-person experience. When users who participate in content interaction are added to the product, these experiences are actually not change. For example, most professional news media, streaming media, non-sandbox games and other content.
Community, self-media, sandbox games and other UGC content belong to the second stage and are more extensive Social media, social networks, and communities in the sense span across phases 1 and 2.
In Phase 3, ABC technology(AI, Big Data, Cloud) will help people generate digital content more efficiently. This part of the content will be naturally and dynamically generated, so it can adapt to multi-person interaction at the same time Dynamic experience.
In the fourth stage, a large amount of rich content is automatically generated by AI to support the Metaverse Content consumption needs of users.
From the development spectrum of the real world and the virtual world, the 123rd stage belongs to the digital real world, while the 4th stage belongs to the original virtual world.
On this basis, we further discussed the technology’s application in several main content areas (media, games, community, social, etc.) Change and influence, and how the content is intelligently generated from the outside and the inside.
Technology will bring people new social relationships in the virtual world, as well as new production relationships. The volume of automatically generated content is exploding at an accelerating rate. The era of UGC and PUGC we currently live in will soon pass. Humans and AI will work together to create more dynamic, rich and personalized content.
1. Information, digital content and people
Born to have a relationship with the world. From the point of view of people as the object of observation, in this world, what can have a relationship with people is people and objects. People communicate with each other. In order to obtain information from each other, language is produced; while people in order to obtain information about objects, The tool was born, and the information of the object was obtained by using the tool, thereby using the object to accomplish various purposes, and even turning the object into the tool itself.
The existence of the information itself does not depend on the way, path and object of the information. The information itself is only related to the information carrier. In other words, when we obtain information from the real world, we only obtain information that already exists in some reasonable way.
In the process of acquiring, transmitting and receiving information, in order to improve the efficiency of information acquisition, we use various media. Language is a virtual carrier, sound is a virtual carrier, tool is a physical carrier, and object is a physical carrier. Of course, if a virtual carrier wants to spread more stably, continuously and effectively, it must be a physical carrier.
After this process occurs, the information form will change, because the carrier has changed.
Strictly speaking, text is also a tool. At the same time, in order to allow the language and the underlying thoughts to be spread stably and effectively, tools such as pens, paper, musical instruments, and music scores have emerged as carriers of different information to help information persist and spread among people.
Information has a carrier, and it becomes content.
The essence of content is information, people are information, objects are information, the language that people communicate with is information, videos that people shoot are also information, and what people buy every day is also information. But strictly speaking, the information after the carrier becomes the content.
At the same time, text, images, music, people, objects, etc. are all content. Before the advent of information technology, people could not directly disseminate people and the information they can produce, such as words, clothes, etc., as content. If you want to communicate with someone, you can only visit or wait for the other person to appear.
For texts and images, which are already information carriers, they can be integrated with other tools again, such as people using paper, carving, etc. for multiple dissemination.
2. Information format
For the information itself, what lies at the bottom is logical information, which enables people to recognize and understand other information in a reasonable way.
If we only have information and content such as language, text, image, shape, temperature, etc., we can indeed obtain their information directly, and everyone will have their own way of understanding and describing them. But if there are no similar or identical logical parts, we cannot communicate with others.
Logical information is often not information at the level of object representation, but the way and way in the process of people’s understanding and cognition.
The language we use every day belongs to natural language, it is a finite language. Although the pronunciation of different language units is different, the basic phonemes of the language are limited, usually only about 40. At the same time, because words are formed by a limited number of basic phonemes, the vocabulary that constitutes any kind of natural language is also limited. Therefore, the corresponding sentences are also limited.
In mathematics, logic, and computer science, formal languages are languages defined by precise mathematics or machine-processable formulas. In the theory of formal languages, it is a collection of certain finite-length strings on an alphabet, and a formal language can contain an unlimited number of strings.
We already know that pure mathematics and logical language cannot describe the system itself in the system, and the information represented by logical language cannot completely represent the content of information in different contexts.
In addition to using the logic in different systems, people also use analogies, associations and other methods to quickly understand and recognize other information.
In this process, the text, image, and sound information on the surface is the most direct way to obtain. In addition, there are information in the form of smell and touch to help us understand and understand the information and content.
The development of technology allows us to disseminate the information on the surface in a more stable and effective way. From paper to telegram, to the text and images on the network screen, we continue to digitize the information of objects in the real world to become digital content.
At the same time, for people in the real physical world, their information and content are uploaded to the Internet through data and become digital content.
Due to technical reasons, we can only digitize part of the information for the time being, such as basic information (name, age, place of birth, etc.) , Portrait information (photos, videos, etc.), preference information (viewing videos Type, label, etc.), behavior information (positioning in the game), etc., so as to approximate the image of the person in the virtual world.
3. The relationship between information and digital content
As we mentioned earlier, content is information with a carrier. The tricky part is that in the real world, we understand people, objects, and things through fragmented information.
It’s just that in the context of knowledge learning, work, etc., in order to improve the efficiency of information reception, we will integrate these fragmented information to form systematic information.
The information on the Internet is all carried information, so in a broad sense, they are all content. Therefore, in the Internet, the relationship between people and information and content becomes:
People — Digital Content — Digital Content — People
This is a simple model. Different people convert the content in the real world into content in the virtual world. This also represents the information carrier of different people on the Internet, which can be voice or text. , It can also be a picture or video.
At the same time, due to the feature that digital information can be edited and modified many times, people will reprocess the information in the virtual world into new content.
Of course, not all information can be modified. For example, some information related to authenticity, confidentiality, and uniqueness cannot be modified, or permission is required.
People—Digital contentAs the organization continues to expand, the quality of content will decline. When the user’s tolerance for bad content is greater than the value he gets from the content, he will leave and look for other content gathering places.
For social media, the way to deal with this situation is actually a trade-off strategy. Since it focuses on the vitality of the entire social media, rather than pure content appeal, its strategy will tend to focus on the media attribute content produced by the head user and the media attribute content produced by the tail user. These are also two comparisons. Extreme content quality.
Some users in the middle will move to more vertical content gathering places and become communities.
Therefore, the similarity between community and social media is that they are both content-oriented places. Of course, a community can also be composed of a lot of content in different fields, but the difference between a community and social media is that the symbol of a community is equal, and the distribution of content will not be affected by the size of the user’s influence.
Users choose different topics and different people’s content by themselves, and generate many-to-many net-like interaction based on the content under an equal communication relationship.
The difference is that after social media joins Wemedia, centralized and decentralized communication will coexist. But in the early stages of development, the influence of centralized communication will occupy a more important position in social media, while the communication of the community will be flatter.
As more and more people have the attributes of media, self-media was born from social networks.
In fact, from the definition of self-media, it means that popular communicators use the Internet to deliver information and content to an unspecified majority or a specific individual. Therefore, as long as users are provided with content dissemination tools, users can become self-media in any place.
So the question is, why is self-media first born on social networks? In fact, this question is also very easy to answer. Between different users and content, an equal relationship is more conducive to the emergence of self-media. If there is an influential and authoritative media in advance, the personal release of the content will naturally be suppressed by the existing media, so that large-scale dissemination cannot be formed quickly.
AlthoughHowever, in the early days, users will choose to use social media to publish content, and try to use the existing user volume to help them build a heart, but it turns out that there are three ways to help spread from the media:
In social media, based on the existing relationship chain or other factors, The influence of self-media to reach the head or authoritative media grows rapidly.
In the community, through continuous contributions to a certain field or topic, it becomes the self-media.
In social networks, more extensive methods are used to continuously spread and extend the relationship chain.
On the whole, any method can bring growth to the media. Relatively speaking, the self-media on social media is more dependent on the overall user volume. The emergence of self-media content enables users to establish a certain degree of social relationship with self-media through the content of self-media. This relationship is stronger than the relationship established based on the content of any user, but it is better than that based on head media or KOL. The relationship established by the content is weaker.
Communities with more equal content interaction have in fact restricted users and content, so self-media is more of an overlap of existing relationships for the community, rather than bringing new interaction methods.
At the same time, the relationship chain in social networks has also increased trust in the information born from the media, so that the content of the media can be spread more effectively and quickly. For this reason, the combination of social network + self-media also carries some attributes of social media.
As the content continues to increase, it contains more and more dimensional fields. The user’s demand for a certain type of content and the time-sensitive demand for content acquisition are also increasing, which has further promoted the segmentation of content.
At the same time, the original user organization structure was further subdivided and fragmented, and smaller circles and communities were born. Communities tend to be people at the core, and the coreIf the characters don’t give, the vitality of the community will gradually die out.
When the intensity of content and the frequency of interaction with people become higher (real-time or near real-time), the form of the media Will change.
If people interact deeply with information from the real world, it will be in the form of live broadcast, and if people interact deeply with information from the virtual world, it will be in the form of games.
Compared with the content in the media, communities, and social networks, the content of the game will bring users higher information density, deeper information interaction, and more frequent interactions. Of course, content like games also requires more investment and takes longer to produce, which is more similar to movies and animations.
This kind of creative content not only requires people to digitize the real world in a certain way, but also requires people to create native content in the virtual world, and then combine the two to form a new digital content.
Digital content, whether in terms of information volume, editability, etc., far exceeds content derived from the real world. They are not just graphics or video streams, but dynamic and interactive content in real time. This will inevitably put forward higher requirements for people’s production capacity.
From the perspective of content, the forms of different media such as text, pictures, long and short videos determine the way the content is presented, and also affect the relationship between users and content in product forms such as media and communities And interactive methods.
At this stage, the content on various media, communities, and social networks generates a huge amount of content every day, so that the information spread based on the social relationship chain can no longer be guaranteed to be the content they want to see, so people start to use it The machine analyzes the user’s preferences and makes the information to find people more efficiently.
Creative content and game content have structured data. In the long run, this also provides a basis for people to use machines to generate content. In the nearer future, machines will collaborate with users to produce more dynamic and rich content.
Second, the development logic of content generation
1. Different stages of content generation
The development of cloud and AI technology has allowed society toThere has been a subtle but important change in the way of communication and entertainment: from the traditional interaction centered on events to the spontaneous interaction centered on people.
With the proliferation of user-generated content (UGC) and AI technology has gradually become the tool and process of creative realization, we are using new methods When creating content, it also unlocks new media, social, community, and gaming experiences.
If we consider the volume of the content, the production method of the content, and the main form of the content at the same time, we can get such a structure:
In fact, we are currently in stage 1 and stage 2 above. Most of the TV, movies, music, and games in the current entertainment industry are created by a large-scale professional creative team. Most of these products are developed for the single-person experience, which means that when we add users who participate in content interactions in the product, these experiences have not actually changed.
As mentioned above, most of the professional news media, streaming media, non-sandbox games and other content in the current market belong to the first stage, and content such as self-media and sandbox games belong to the second stage. Social media, social networks, and communities in a broader sense span both phases 1 and 2.
Of course, ABC technology (AI, Big Data, Cloud) has been gradually applied to the industry, subject to certain rules and constraints , To help people generate digital content more efficiently, which is also the content generation described in stage 3.
With the support of AI, cloud computing and big data, this part of the content will be naturally generated dynamically, and therefore can adapt to the dynamic experience of multiple people at the same time.
Part 4 is to generate a large amount of rich content automatically by AI to support the content consumption needs of users in Metaverse.
“GTA 5″The development cycle is as long as 5 years, the total capital investment is expected to be 266 million US dollars, and more than 1,000 developers are involved in the production. The capacity of the meta universe may be millions of times that of “GTA 5”, and multiple people interact dynamically in real time. Therefore, the birth and continuous operation of the meta universe must use AI to generate content.
From the development spectrum of the real world and the virtual world, the 123rd stage belongs to the digital real world, while the 4th stage belongs to the original virtual world.
2. Co-create with AI
We know that both the media and the community depend on the digital content produced by users.
In order to generate this part of the content in a faster and differentiated way, computer vision technology allows us to quickly transmit real-world information through photos and videos to the digital world, and use computer graphics technology With the help of, we can see this content on any screen.
In stage 1 and 2 production, only a few people are content producers (relatively speaking). But AI can make content generation more democratized, further lower the threshold for content creation, and make it easier.
With the help of AI tools, more people will become creative workers. We can convert specific instructions, images and logical processing methods into production assets in a lightweight and modular manner, thereby greatly reducing the burden of coding, drawing, animation and other tasks.
We know that logic is the basis for decision-making, and language, action and hearing are the objects of this basis; while text, images and music are the media, which we can present on the screen in a digital way. Therefore, the digital content that can be generated by the machine is divided into logical (RL/SL/USL), language(NLG+TTS), action(MG+CG) etc.
The GPT-3 released by OpenAI this year supports multi-modal human-computer interaction, and it learns very dynamically, and can generate text that far exceeds the performance of GPT-2. And TTS, NLG, MG, etc. allow machines to generate corresponding media.
At the same time, AI can quickly learn the content of the script by understanding the information of the game or movie script, and under the given constraints, generate a reasonable plot line and the dialogue and translation of the corresponding NPC.
In fact, the difference from general artificial intelligence is that although the content of a game or movie looks very ambitious, in fact the script may be several hundred pages, and the amount of information contained in it is acceptable.
Morpheus Cloud-Storyline/Dialogue Generation
Under a given knowledge framework, AI becomes able to gradually understand the creator’s intentions, not only to control the different responses of NPC to user interactions, but also to create personalized encounters and dynamic plot content with users .
In addition, if the plot or narrative is regarded as broad content, then people can upgrade the existing graphic and video content into real-time interactive content with the assistance of AI. Whether it is an object or an avatar, everything in the content will gradually realize “everything can interact”.
Traditional digital content does include the digitization of people, so we can use digital images and enter the virtual world.Interaction of line content.
When the intelligent body created by AI actively participates in the changes of the virtual world and the interaction with the real world, it becomes a “virtual human”.
Because virtual humans are digital content, when “virtual humans” with a certain degree of intelligence appear, they can autonomously generate and change a certain degree of digital content. At this time, the relationship between media production and people’s supply and demand for content will change.
With the assistance of AI, the content generated by “virtual people” can be more time-sensitive and scaleable than real media practitioners. Therefore, the form and content of social media will change first. Second, changes in the content of social media will bring about changes in the organizational structure of users, thereby affecting changes in the social relationship chain.
On this basis, when we in the real world establish a social relationship, it is probably because a “virtual person” is interacting with another “virtual person”, bringing up both sides of the real world, and a new one is born Social scenes and new social networks.
Of course, the form of the community will also change simultaneously. The data learned by AI is not only structured, but also strictly equal.
So for “virtual people”, the topics, content and fields of the community can be completely mastered and generated by them. We can boldly imagine how real people interact with the content generated by virtual people in a community composed of virtual people? If the content comes from games, can we directly enter the game and compete with virtual users in the community? Does this bring a new game entry and scene? …
In fact, AI not only needs to make decisions on the logical side, but also needs to pass this type of higher-level instructions to the image, so that the image can move according to human interaction.
Deep learning can make AI-controlled character movements more coherent and smooth. People can adjust various parameters during their movements to make them show different effects.
In addition, people are also trying to use AI to generate digital music and art. The music generation system Jukebox released by OpenAI is a neural network system. Using different genres, artists and lyrics as input, Jukebox can output new music samples created from scratch. For art, people usually refer to the art generated using GAN technology as GANism art (GANism).
In 2018, Adobe held a press conference that truly demonstrated how to use AI in art and design. Whether it is the smart matting realized by Fast Mask, Moving Stills that can generate dynamic videos from static images, or the automatic generation of various artistic fonts, all designers and artists today are dumbfounded.
3. How does consumer technology change content generation?
1. Create more interactive media content
Current media almost have digital products, and traditional paper media also have electronic versions. At the same time, they have also opened up new media communication methods, integrating text, pictures, long and short videos, etc. through digital creation. Content.
For media with independent sites, choosing the platform where users gather is almost a must. Regardless of whether it is self-media or social media, the core proposition of the media is: how to use less investment and expenditure in exchange for greater communication influence?
And any platform with media attributes will consider the next question: How to help creators maintain content differentiation while improving content output efficiency?
In fact, in the graphic era, the platform basically provides a rich text editor; in the (short) video era, the platform tool attributes The strength determines the richness and efficiency of content that users can use.
The core of video lies in visual effects, so computer vision technology can give full play to its potential in this field, allowing a variety of visual effects to be added to recorded videos or live broadcasts in a simple and effective way .
The tool accumulates initial users for the platform, and then settles its content, distributes and pushes it through algorithms or other mechanisms, so that a content product with media attributes appears.
Whether such content products are going to social media or the community can only be judged based on whether the content and interaction are biased towards users or the content itself. If it is a user, it is a social media attribute; if it is a content, it is a community orientation.
For the media, the video is followed by an immersive experience of real-time interaction, or the interactive object of the content becomes an agent in the virtual world. Since the interactive and immersive nature of games is the strongest among digital content, the former method is more likely to be a real-time game community, or bring gamified interactive methods into media products.
The latter is because virtual humans themselves are digital content. When “virtual humans” with a certain degree of intelligence appear, they can autonomously generate and change a certain degree of digital content, media production relations and people’s supply and demand for content. The relationship will change, which brings about new interactions.
At the same time, we can also see that from traditional TV content and news media toNowadays, short videos and content as media, their creation threshold and dissemination mechanism are becoming more and more automated and lightweight. This is the credit of the technology, which makes content generation and distribution more intelligent, thereby releasing the user’s content creativity, allowing everyone to create real-time interactive content in a very convenient and efficient way.
In terms of interactivity, YouTube’s interaction is more like moving the remote control of a traditional TV to the screen.
Although UGC has brought a lot of content to YouTube, the interaction between users and content is still based on the traditional “search + recommendation” logic.
As a benchmark for long videos, YouTube is a video site that hardly needs to pay network fees, because its biggest advantage is actually the free bandwidth provided by Google Peering for YouTube. From professional media reporters to anyone on the street, as long as you want to create and share content in the form of video through the Internet, no matter how long the video is, YouTube will provide you with high-definition, high-bitrate, and Qualcomm technical services.
Therefore, YouTube is purely post-paid for content creators, and it will almost never appear as a domestic video website, bypassing the creator’s share plan and directly paying high-value content producers with high signing fees.
Can the logic of YouTube be used to reproduce it in China? The answer is actually very obvious. In China, the cost of bandwidth, cloud computing, and other data cannot be reduced by enterprises in the most direct way. Therefore, it is necessary to limit the duration of content or use other methods for commercialization.
We can also see many video-oriented media and communities. At the beginning, they will tell a story about burning money to accumulate users and then commercializing users. But not all companies can realize this story, because it is not only related to technology, but also related to the interaction between users and the content itself.
For TikTok, short video does reduce the bandwidth cost by a large part, and it also integrates real-time interaction into the user’s viewing experience, rather than the interaction between users and content like graphics Limited to comments. Users can scroll and browse at willView videos, and interact with their favorite videos quickly and easily.
In fact, TikTok has fundamentally changed the way we think about content, and content will no longer be something we passively accept. Simple and immersive interaction methods combined with strong recommendation algorithms can allow us to experience the content world created by different people without interruption. Once you don’t like it, you can quickly withdraw, and the system will recommend content again.
If you do not regard TikTok as a media or community, but as a gamification portal, it will not only lead to various digital real worlds, but also lead to the original virtual world. Long videos and short videos are different forms of interaction of different types of content, but short videos can actively interact with users at a faster refresh frequency.
We can also see that when users are producing this type of video content, they will use various images and visual techniques to quickly achieve content differentiation. These technologies or visual effects, when assisting the generation of a single content, also have a social attribute, and everyone can use this effect to share different content.
In other words, when generating content, technology can create more ways of interaction between content and users, further increasing the diversity of content. At the same time, technology can make the interaction between users and content more dynamic, and constantly create content that players like.
The tricky point is that whether it’s a long video or a short video, users may not experience dynamic content generation in a certain video content, but they can constantly feel the difference in the same theme, music and scrolling. Dynamic appearance of content.
Allowing users to experience dynamic interactions outside the content is actually achieved by technologies such as recommendation algorithms.
If users want to have deeper dynamic interactions within the content, then we need technology to understand the logic of the content, not just make the content different at the visual and image levels.
This requires more AI technology to achieve: decision generation on the logic side and dynamic interaction between logic and image actions.
When the logic of content generation is implemented, we can think that this