Frequency-Domain Loudness Processing for Variable Audio Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining and modifying the perceived loudness of audio signals, particularly in perceptual coding where block sizes vary, are complex due to the need for constant frequency and temporal resolution, leading to difficulties in sound quality for transient signals.
Innovation Solution
The method involves combining frequency domain data from multiple block sizes to form a longest block size, interpolating gain information, and applying it to shorter block sizes to maintain consistent loudness processing, using techniques like interleaving or replication to create a constant block size for efficient loudness determination and modification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If perceptual coding methods use varying block sizes to improve sound quality for transient signals, then sound quality is improved, but the complexity of loudness determination and modification increases
Solution Approach 1:
The patent segments the frequency domain data into multiple blocks of different sizes (short blocks and long blocks) and processes them separately through different processing paths. Short blocks are combined to form long blocks, which then all undergo uniform loudness processing. This segmentation allows the system to handle varying block sizes while maintaining consistent processing quality.
Solution Approach 2:
Instead of processing each varying block size directly through loudness determination, the patent inverts the approach by combining multiple short blocks into long blocks first, then processing the unified long blocks through the loudness processing pipeline. This inversion simplifies the processing architecture by establishing a constant output format regardless of input block size variations.
2Measurement precision
If constant frequency and temporal resolution is maintained in loudness processing, then measurement precision is improved, but the difficulty of processing varying block sizes increases
Solution Approach 1:
The patent changes the block size parameter dynamically based on the input signal characteristics. Short blocks (256 samples) are used for transient signals requiring high temporal resolution, while long blocks (2048 samples) are used for steady-state signals requiring high frequency resolution. The system transforms varying block sizes into a unified long block format for consistent processing, effectively managing the trade-off between temporal and frequency resolution.
3Adaptability or versatility
If multiple block sizes are used in perceptual coding, then adaptability is improved, but the consistency of loudness processing deteriorates
Solution Approach 1:
The patent performs preliminary combining of short blocks into long blocks before the actual loudness processing stage. This preliminary action ensures that all subsequent processing operates on uniformly sized long blocks, maintaining consistency throughout the processing pipeline while still accepting varying block sizes as input. The combining operation is performed in advance to establish uniformity before processing begins.
Data Source
AI summary
Methods of, apparatuses for, and non-transitory computer readable media having instructions thereon that when executed cause carrying out methods of determining and modifying the perceived loudness of a frequency domain audio signal where the frequency resolution, and corresponding temporal coverage of the frequency domain information is not constant. The frequency (and thus temporal) resolution of the perceived loudness processing is maintained constant at the longest block size. One method includes a block combiner and a loudness modification interpolator.