Dialogue-Based Effect Generation With Real-Time Preview
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face a complex and time-consuming process when creating video and picture effects due to the need to learn how to use effect creation tools, which hinders their efficiency and effectiveness in generating media content.
Innovation Solution
A method and apparatus that utilize a session interface with a virtual object to generate effects through a dialogue-based interaction, allowing users to input messages describing the desired effect, receive a corresponding response, and preview the generated effect in real-time, reducing the learning curve and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional effect creation tools are used, then users can create video and picture effects, but users need to learn how to use the tools which makes the process complex and time-consuming
Solution Approach 1:
The patent introduces a large language model as an intermediary between the user and the effect creation system. The user interacts with the LLM through natural language chat, and the LLM translates user intent into effect generation parameters and control data, eliminating the need for users to learn complex tool interfaces
Solution Approach 2:
The system enables self-service effect creation by allowing users to describe their desired effects in natural language. The LLM automatically interprets the description, generates appropriate effects, and applies them without requiring users to manually configure technical parameters or learn specialized tools
2Productivity
If traditional effect creation tools are used, then effects can be created, but the process is complex and learning curve is steep
Solution Approach 1:
The patent replaces the traditional mechanical interaction model (clicking buttons, adjusting sliders, navigating menus) with a natural language-based interaction model. Users communicate their effect creation needs through text chat, and the LLM converts this into system commands, dramatically simplifying the interface complexity
3Manufacturing precision
If users want to preview effects before finalizing, then effect quality can be improved, but additional time is required for iteration
Solution Approach 1:
The system performs preliminary effect generation and preview before finalizing the effect. The LLM generates effect parameters and provides a preview to the user, allowing them to evaluate and request modifications before the effect is fully applied, ensuring quality while streamlining the iteration process
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
Embodiments of the disclosure relate to a method, an apparatus and a device for generating an effect and a computer-readable storage medium. The method proposed herein includes: presenting a session interface with a virtual object; obtaining a first message via the session interface, the first message configured to describe a to-be-generated effect; presenting a second message from the virtual object in the session interface, the second message including an interaction component corresponding to a target effect, and the target effect generated based on the first message; and presenting a preview window of the target effect in response to a first operation on the interaction component. In this way, the embodiments of the disclosure can generate and edit the effects by means of a dialogue, thereby improving efficiency of generating effects.