Abstract:
A processor can be associated with a memory for storing convolution data. A plurality of M filters from a corresponding plurality of M input channels to a selected one output channel can be provided, wherein each filter can be represented by a corresponding index, m. Each of the M filters can be partitioned into K respective filter partitions, wherein each respective filter partition can be represented by a corresponding index, k. A frequency-domain representation of each filter partition can be provided, wherein each frequency-domain representation of a filter partition comprises N frequency bins and a corresponding frequency-domain filter coefficient, wherein each respective frequency bin can be represented by a corresponding index, n. The memory can store such information in an arrangement suitable for the processor to concurrently receive sufficient information to concurrently convolve a frame of each input signal with the respective filters.
Abstract:
An echo canceller can be arranged to receive an input signal and to receive a reference signal. The echo canceller can subtract a linear component of the reference signal from the input signal. A noise suppressor can suppress non-linear effects of the reference signal in the input signal in correspondence with a large number of selectable parameters. Such suppression can be provided on a frequency-by-frequency basis, with a unique set of tunable parameters selected for each frequency. A degree of suppression provided by the noise suppressor can correspond to an estimate of residual echo remaining after the one or more linear components of the reference signal have been subtracted from the input signal, to an estimated double-talk probability, and to an estimated signal-to-noise ratio of near-end speech in the input signal for each respective frequency. A speech recognizer can receive a processed input signal from the noise suppressor.
Abstract:
An echo canceller can be arranged to receive an input signal and to receive a reference signal. The echo canceller can subtract a linear component of the reference signal from the input signal. A noise suppressor can suppress non-linear effects of the reference signal in the input signal in correspondence with a large number of selectable parameters. Such suppression can be provided on a frequency-by-frequency basis, with a unique set of tunable parameters selected for each frequency. A degree of suppression provided by the noise suppressor can correspond to an estimate of residual echo remaining after the one or more linear components of the reference signal have been subtracted from the input signal, to an estimated double-talk probability, and to an estimated signal-to-noise ratio of near-end speech in the input signal for each respective frequency. A speech recognizer can receive a processed input signal from the noise suppressor.