Real-Time Mixed Audio Decomposition Using Chunked AI Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for decomposing mixed audio signals into separate tracks in real-time playback situations suffer from significant latency, making it difficult to achieve high-quality decomposition without delaying the continuous playback of mixed audio data, especially when the length of the audio content is unknown or variable.
Innovation Solution
A method that decomposes a continuous stream of mixed audio data using an AI system, allowing playback of decomposed audio data to start before the entire stream is received, by loading chunks of data into a buffer larger than the chunk size, enabling quick decomposition and playback with minimal latency, and progressively improving audio quality as more data is received.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional AI systems preprocess entire audio files before decomposition, then decomposition quality is improved, but latency increases significantly (tens of seconds to minutes)
Solution Approach 1:
The patent divides the audio stream into smaller chunks that can be processed incrementally. Instead of waiting for the entire audio file to be received, the system processes decomposition on partial data segments, allowing output to be generated progressively with much reduced latency while maintaining acceptable decomposition quality through iterative refinement as more data becomes available.
2Manufacturing precision
If the audio buffer size is increased to improve decomposition quality, then more audio data is available for analysis, but playback latency increases
Solution Approach 1:
The patent implements a dynamic buffer management system where the buffer size and processing parameters are adjusted adaptively based on the incoming audio stream characteristics and processing progress. This allows the system to optimize between quality and latency in real-time, using larger buffers when quality is prioritized and smaller buffers when low latency is critical, rather than committing to a fixed buffer size.
3Manufacturing precision
If conventional techniques wait for complete audio content before processing, then all audio information is available for decomposition, but real-time playback capability is lost
Solution Approach 1:
The patent performs preliminary decomposition actions on the portions of audio data that are already available, rather than waiting for complete audio content. This allows the system to generate output progressively as data arrives, enabling real-time playback capability while continuously improving decomposition quality as more audio information becomes available for subsequent processing iterations.
Data Source
AI summary
The present invention relates to a method and a device 10 for processing mixed audio data, including decomposing in real-time with low latency, in which a continuous stream of mixed audio data is received from an audio source 14, a first chunk of the stream of mixed audio data is loaded into a buffer, the audio data contained in the buffer is decomposed to obtain first decomposed audio data representing audio signals of a predetermined timbre, and a first chunk of output data is obtained from the first decomposed audio data, preferably for direct playback via speaker 26.