Multichannel Interference Cancellation Using Relative Transfer Functions

The patent segments the interference cancellation process into two distinct stages: a first cancellation stage that processes all microphone signals simultaneously using full multichannel processing, and a second cancellation stage that processes residual interference in each channel separately. This segmentation allows the system to achieve effective interference reduction while reducing overall computational complexity by dividing the problem into manageable parts with different processing requirements.

The patent introduces an intermediate processing stage where the output of the first cancellation stage serves as input to the second cancellation stage. This intermediary step allows the system to progressively reduce interference, with the first stage handling the bulk of the cancellation and the second stage refining the result, thereby achieving effective interference reduction without requiring full multichannel processing for every stage.

The patent applies segmentation by dividing the echo cancellation process into two stages: a global cancellation stage that processes all microphone signals together to reduce overall echo, and individual channel cancellation stages that process residual echo in each microphone signal. This allows the system to effectively handle multiple microphones for improved speech enhancement while keeping computational complexity manageable through staged processing.

The patent implements partial action by applying full multichannel processing only in the first cancellation stage, while the second stage uses simpler per-channel processing. This partial application of complex processing is sufficient to achieve effective echo reduction, avoiding the need for excessive computational resources while maintaining speech enhancement performance with multiple microphones.