Audio Signal Phase Modification Using Padded Spectral Blocks

The patent applies preliminary action by detecting transients before phase vocoder processing and applying specific handling (such as time-stretching or zero-padding) to transient-containing blocks beforehand. This prevents temporal aliasing and waveform dispersion from occurring in the first place, rather than correcting them after processing. The method identifies transient blocks and applies different processing parameters to these blocks compared to non-transient blocks, thereby preserving audio quality while maintaining processing efficiency.

The patent implements local quality by applying different processing strategies to different types of audio blocks. Transient-containing blocks are identified and processed with special parameters (such as modified time-stretching factors or zero-padding) to preserve their quality, while non-transient blocks are processed with standard efficient phase vocoder parameters. This localized adaptation of processing parameters maintains overall system efficiency while protecting transient quality where needed.

The patent applies partial action by performing transient detection and selective processing only on blocks that contain transients, rather than applying complex processing to all blocks. The system uses efficient detection mechanisms to identify transient blocks and applies quality-preserving processing only where necessary, leaving the majority of non-transient blocks to be processed with simpler, more computationally efficient methods. This selective approach maintains audio quality for critical transient portions while minimizing overall computational energy consumption.