Animated Avatar Rendering Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current artificially intelligent devices, such as smart speakers, primarily interact with humans through audio-only responses, lacking visual cues which limits the realism and effectiveness of human-machine interaction, and require substantial computational resources for rendering animations, leading to delayed user experience.
Innovation Solution
A system and method for rendering an animated avatar that decomposes audio signals into phonemes, synchronizes animations with audio, and uses a multiplicative factor to determine rendering times for clips, allowing parallel rendering and natural break points in speech to reduce latency and enhance interaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If animations are rendered sequentially with high computational resources, then rendering quality is improved, but user wait time and latency increase
Solution Approach 1:
The patent divides the animation rendering process into multiple independent clips, each corresponding to a segment of the audio response. These clips can be rendered in parallel rather than sequentially, significantly reducing total rendering time while maintaining quality. Each clip is processed independently and then stitched together to form the complete animation.
Solution Approach 2:
The system performs preliminary actions by pre-rendering animation clips and storing them before they are needed. When a response is generated, pre-rendered clips can be quickly selected and assembled, eliminating the need for real-time rendering and thus reducing user wait time while preserving rendering quality.
2Speed
If computational resources are allocated for real-time animation rendering, then interaction responsiveness is improved, but system resource consumption increases
Solution Approach 1:
Animation clips are rendered in advance and stored in a cache, so that when an audio response is generated, the corresponding visual animation is already available for immediate playback. This eliminates the need for high computational resources during real-time interaction while maintaining fast responsiveness.
Solution Approach 2:
Instead of rendering unique high-quality animations for every possible response, the system creates a library of pre-rendered animation clips that can be reused across multiple responses. This copying approach significantly reduces computational resource consumption while maintaining visual quality and interaction responsiveness.
3Productivity
If animation clips are rendered in parallel, then rendering time is reduced, but synchronization complexity with audio increases
Solution Approach 1:
The audio response is segmented into discrete units (such as phonemes or syllables) that correspond to pre-rendered animation clips. Each segment has a defined start and end time, making it straightforward to synchronize multiple parallel-rendered clips with their corresponding audio segments without complex coordination.
Solution Approach 2:
The patent introduces an intermediary synchronization layer that matches audio segments to corresponding pre-rendered animation clips based on timing information. This intermediary layer handles the coordination between parallel rendering processes and audio playback, simplifying the overall synchronization complexity while maintaining high rendering throughput.
Data Source
AI summary
There are provided systems and methods for rendering of an animated avatar. An embodiment of the method includes: determining a first rendering time of a first clip as approximately equivalent to a predetermined acceptable rendering latency, a first playing time of the first clip determined as approximately the first rendering time multiplied by a multiplicative factor; rendering the first clip; determining a subsequent rendering time for each of one or more subsequent clips, each subsequent rendering time is determined to be approximately equivalent to the predetermined acceptable rendering latency plus the total playing time of the preceding clips, each subsequent playing time is determined to be approximately the rendering time of the respective subsequent clip multiplied by the multiplicative factor; and rendering the one or more subsequent clips.