Avatar Dance Generation via Audio Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for controlling avatars in VR/AR environments require large memory databases to generate dance moves, making them unsuitable for edge devices and limiting creative responsiveness to music.
Innovation Solution
A method that extracts high-level and latent audio features from audio signals to generate joint angle distribution matrices using Gaussian parameters, allowing avatars to improvise dance steps without a predefined database, suitable for edge devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a database storing a large number of preset dance moves is maintained, then the avatar can generate dance moves, but memory usage increases and implementation on edge devices becomes difficult
Solution Approach 1:
The patent extracts only the essential rhythmic features from audio signals (beat detection, tempo extraction) rather than storing complete dance move databases. This extraction approach reduces memory requirements while maintaining the core functionality of generating dance moves that match the music rhythm.
Solution Approach 2:
The patent replaces the mechanical database storage and retrieval system with an audio analysis and synthesis system. Instead of storing preset dance moves and selecting them based on music characteristics, the system analyzes audio signals in real-time and generates corresponding dance moves through computational processing, eliminating the need for large memory databases.
2Productivity
If predetermined hand-crafted features are used to select dance moves from the database, then dance moves can be generated, but the avatar cannot dance creatively
Solution Approach 1:
The patent implements dynamic dance move generation that adapts to the characteristics of the input music. The system extracts rhythmic features from the audio signal and uses these features to dynamically generate dance moves that match the music's tempo, beat patterns, and rhythm, rather than selecting from static preset combinations. This dynamic approach enables creative dance performances tailored to each music input.
3Adaptability or versatility
If a large database of preset dance moves is maintained, then dance moves can be selected and recombined, but the system complexity increases
Solution Approach 1:
The patent extracts only the essential rhythmic characteristics from audio signals (beat timing, tempo, rhythm patterns) rather than maintaining complex databases of dance moves. This extraction focuses on the minimum necessary information to generate dance moves, significantly simplifying the system while preserving the core functionality of creating music-synchronized dance performances.
Data Source
AI summary
The disclosure provides a method for generating action according to an audio signal and an electronic device. The method includes: receiving an audio signal and extracting a high-level audio feature therefrom; extracting a latent audio feature from the high-level audio feature; in response to determining that the audio signal corresponds to a beat, obtaining a joint angle distribution matrix based on the latent audio feature; in response to determining that the audio signal corresponds to a music, obtaining a plurality of designated joint angles corresponding to a plurality of joint points based on the joint angle distribution matrix; and adjusting a joint angle of each of the joint points on the avatar according to the designated joint angles.