Layered Audio Generation With Multi-Channel Spectrograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio content generation in video games is time-consuming, expensive, and computationally intensive, particularly for creating varied audio assets, and existing computational solutions often require single-output generation, increasing time and costs.
Innovation Solution
A method using a generative model trained on multi-channel graphical representations of audio assets, specifically spectrograms, to efficiently generate layered audio assets with reduced computing power, allowing for quick and autonomous creation of varied audio assets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual crafting of audio assets is used to create variations, then audio quality and variation are improved, but time consumption and cost increase
Solution Approach 1:
The patent uses a trained generative model to automatically generate copies/variations of audio assets based on learned patterns from training data, replacing manual crafting while maintaining quality consistency
Solution Approach 2:
The system generates audio variations by manipulating parameters such as pitch, tempo, and timbre through the generative model, creating diverse audio assets from a single source without manual intervention
2Loss of time
If computational audio generation is used to generate assets on the fly, then time consumption is reduced, but processing complexity and cost increase
Solution Approach 1:
The system performs preliminary training of the generative model during asset creation, storing the trained model for rapid generation of variations later, reducing real-time processing complexity
Solution Approach 2:
The trained model is saved and reused to generate multiple audio variations efficiently, avoiding repeated training processes and reducing computational overhead
3Device complexity
If single-output generative models are used, then model simplicity is maintained, but productivity decreases due to sequential generation
Solution Approach 1:
The patent segments the audio asset into multiple independent layers (e.g., different instrument tracks, voice channels), allowing parallel generation of each layer through the same generative model
Solution Approach 2:
The system combines multiple independently generated audio layers into a complete audio asset, achieving multi-output capability while reusing the same simple generative model for each layer
Data Source
AI summary
A method of generating audio assets, comprising the steps of: receiving an input multi-layered audio asset comprising a plurality of audio layers, generating an input multi-channel image, wherein each channel of the input multi-channel image comprises an input image representative of one of the audio layers, training a generative model on the input multi-channel image and implementing the trained generative model to generate an output multi-channel image, wherein each channel of the output multi-channel image comprises an output image representative of an output audio layer, and generating an output multi-layered audio asset based on a combination of output audio layers derived from the output images.