AI Solution for Multi - person Collaboration and Asset Version Management in High - quality TV Dramas

In the field of high-quality TV drama production, multi-person collaboration and asset version management have always been difficult problems for production teams. The Wemio content engine of WeLinkirt DaoAI brings new ideas and methods to solve these problems with advanced AI technology.

Industry background and user scenario: In today's film and television industry, high-quality TV dramas are loved by a large number of audiences for their high-quality plots and exquisite production. A high-quality TV drama production team plans to produce a 20-episode series, with each episode about 45 minutes long, and the total number of shots is expected to exceed 5000. The team's production requirement is to shorten the production cycle and reduce the production cost as much as possible while ensuring the quality of the series, and at the same time, ensure the consistency of characters, scenes, and lighting across shots. However, the current industry situation is that there are many problems in multi-person collaboration and asset version management in traditional production methods, resulting in low production efficiency and high costs.

Pain points: Why is it difficult to ensure consistency across shots?

Multidimensional quantification of pain points: First of all, the proportion of character 'face-changing' across shots is relatively high. According to statistics, in the traditional production method, about 20% of the shots have problems with inconsistent character appearances. Secondly, the phenomenon of scene incoherence occurs frequently, and about 15% of the scenes bring an obvious sense of fragmentation to the audience when switching. Moreover, the production cycle is long. A project originally planned to be completed in 6 months often takes 8 months or even longer. The unit cost is also high, with the cost per minute of the finished film being about $1000. In terms of multi-person collaboration, the communication cost is high, and the information transmission is not timely, resulting in low work efficiency. The situation of repeated manuscript revisions is also relatively serious, with an average of 3-4 revisions for each shot.

Root cause analysis: The root cause why the generation layer has difficulty in locking the consistency across shots is that the traditional video generation method mainly relies on pure prompts and lacks consideration of 3D space and physical constraints. In this way, the generation of different shots is relatively independent, and there is no unified base to ensure the understanding of semantics and 3D space. Therefore, inconsistencies are likely to occur in terms of characters, scenes, and lighting. In addition, during multi-person collaboration, due to the lack of effective collaboration tools and version management mechanisms, the information is chaotic, making it difficult to ensure the consistency and accuracy of assets.

Technical principle

In - depth mechanism: The Wemio physical AI content engine of WeLinkirt DaoAI introduces 3D and physical constraints into video generation. In terms of 3D space, it can accurately construct the spatial structure of the scene, making the scenes in different shots coherent in space. For example, when constructing an indoor scene, the engine will consider factors such as the layout of the room and the placement of furniture to ensure that the spatial relationship of the scene is consistent in different-angle shots. In terms of physical constraints, it follows the physical laws of the real world, such as lighting and gravity, making the actions of characters and the performance of scenes more realistic. For example, when dealing with lighting effects, the engine will simulate real lighting changes according to factors such as the time and weather of the scene, so that the lighting effects in different shots are consistent. The DaoAI World model, as a unified base, has the ability to understand semantics and 3D space. It can parse the semantic information in the script and transform it into entities and relationships in 3D space. For example, when the script mentions 'a man is watching TV on the sofa in the living room', the world model can accurately construct the images of the living room, sofa, and man in 3D space and determine their positional relationships. In this way, the world model provides a basis for the consistency across shots/episodes.

Comparison with pure prompt-based generation: Compared with pure prompt-based generation, the Wemio physical AI content engine and the DaoAI World model have obvious advantages. Pure prompt-based generation mainly relies on text descriptions to generate images or videos and lacks consideration of 3D space and physical constraints, so it is prone to problems of inconsistency across shots. The Wemio physical AI content engine introduces 3D and physical constraints into video generation, which can fundamentally solve these problems. For example, when dealing with character actions, pure prompt-based generation may not be able to ensure the coherence and authenticity of actions, while the Wemio physical AI content engine can generate natural and smooth actions according to physical laws. The unified base function of the DaoAI World model provides a unified framework for the entire production process, which can better ensure the consistency across shots/episodes, while pure prompt-based generation lacks such a unified management mechanism.

Typical application scenarios

• Scenario 1: Scriptwriting stage. At this stage, screenwriters can use the screenwriting agent in the intelligent agent pipeline to quickly generate scripts based on the story outline. The difficulty lies in how to make the intelligent agent understand the creative ideas and intentions of the screenwriters and generate script content that meets the requirements. The Wemio content engine can better interact with screenwriters and generate more suitable scripts through the semantic understanding ability of the DaoAI World model.
• Scenario 2: Storyboard design stage. Storyboard artists can use the Wemio physical AI content engine to generate storyboard images with a sense of 3D space and physical realism based on the script. The difficulty lies in how to ensure the consistency of storyboard images across shots, especially when switching between different scenes. By introducing 3D and physical constraints, the engine can ensure the consistency of storyboard images in terms of space and lighting.
• Scenario 3: Video production stage. At this stage, the production team can use the video-output agent in the intelligent agent pipeline to quickly generate video content. The difficulty lies in how to improve production efficiency while ensuring video quality. The physical AI ability of the Wemio content engine and the collaborative work of the intelligent agent pipeline can greatly shorten the production time and ensure the consistency of the video across shots.
• Scenario 4: Editing stage. Editors can use the editing agent to quickly edit according to the storyboard and video content. The difficulty lies in how to maintain the consistency of characters, scenes, and lighting during the editing process. The cross-shot locking ability of the Wemio content engine can help editors easily solve this problem and ensure the quality of the final finished film.

Implementation case

Comparison before and after using for an anonymous client: Before using the Wemio content engine of WeLinkirt DaoAI, a high-quality TV drama production team faced problems such as a long production cycle, high costs, and poor consistency across shots. When producing a 20-episode series, the production cycle was as long as 8 months, the cost per minute of the finished film was about $1000, the proportion of character 'face-changing' across shots was about 20%, and the proportion of scene incoherence was about 15%. After using the Wemio content engine, the production cycle was shortened from 8 months to 4 months, a reduction of -50%. The cost per minute of the finished film was reduced to about $700, a cost reduction of -30%. The proportion of character 'face-changing' across shots was reduced to about 5%, and the proportion of scene incoherence was reduced to about 3%. At the same time, the efficiency of multi-person collaboration was greatly improved, and the number of manuscript revisions was also significantly reduced.

The application of AI technology has brought unprecedented changes to high-quality TV drama production, and the Wemio content engine of WeLinkirt DaoAI has become the key to solving production problems.

Wemio solution and product

Product capabilities and pipeline: The Wemio content engine of WeLinkirt DaoAI mainly involves the Wemio physical AI content engine, the DaoAI World model, and the intelligent agent pipeline from screenwriting to storyboarding, video output, and editing. In the screenwriting link, the screenwriting agent parses the script semantics with the help of the DaoAI World model, laying the foundation for subsequent creation. In the storyboarding link, the Wemio physical AI content engine generates storyboard images with a sense of 3D space and physical realism based on the script and the parsed semantic information, and at the same time ensures cross-shot consistency through the world model. In the video output link, the video-output agent quickly generates video content using the physical AI engine, and due to the introduction of 3D and physical constraints, the cross-shot consistency of the video is guaranteed. In the editing link, the editing agent quickly edits according to the storyboard and video content, and uses the cross-shot locking ability to ensure the consistency of characters, scenes, and costumes in different shots. In terms of multi-person collaboration, the engine supports real-time multi-person collaboration. Team members can operate on the project simultaneously, share the credit pool to reduce costs, recover the permissions of departing employees with one click to ensure data security, and the project transfer and data retention function facilitates project handover and management.

Quantitative results and business value: In terms of quantitative indicators, after using the Wemio content engine, the production cycle was reduced by -50%, the cost was reduced by -30%, and the monthly credit consumption was saved by about -40%. These results have brought significant business value. On the one hand, the shortened production cycle enables the series to be launched into the market faster, seizing market share; on the other hand, the reduced cost improves the profitability of the project. At the same time, the improvement of cross-shot consistency ensures the quality of the series, improves the viewing experience of the audience, and helps to enhance the brand image and market competitiveness.