Speech Technology & Applied Research Corporation

Other

Kill the Cann, LLC

Keith Gilbert

Joel MacAuslan

Richard Scott Goldhor

Karen Payton

Geoffrey F Gross

A system processes data signals consisting of sums of independent signal terms, zero or more of which signal terms may already have been identified, in order to generate one or more additional terms. Deflated versions of the data signals are created by subtracting from the data signals any previously identified signal terms. Additional independent signal terms are computed using a set of reference signals organized into mutually independent partioning support sets. The images of each support set are computed on the data signals. Computed images on a data signal that are non-zero are identified as additional independent signal terms of that data signal.

Using images and residues of reference signals to deflate data signals

In environments (such as acoustic and bioelectrical environments) characterized by multiple simultaneous sources, effective blind source separation from sensor response mixtures becomes difficult as the number of sources increases-especially when the true number of sources is both unknown and changing over time. However, in some environments, non-sensor information can provide useful hypotheses for some sources. Embodiments of the present invention provide an adaptive filtering architecture for validating such source hypotheses, extracting an estimated representation of source signals corresponding to valid hypotheses, and improving the separation of the remaining 'hidden' source signals from the sensor response mixtures.

Hypothesis-based estimation of source signals from mixtures

Deflation and decomposition of data signals using reference signals

In environments (such as acoustic and bioelectrical environments) characterized by multiple simultaneous sources, effective blind source separation from sensor response mixtures becomes difficult as the number of sources increases&#x2014;especially when the true number of sources is both unknown and changing over time. However, in some environments, non-sensor information can provide useful hypotheses for some sources. Embodiments of the present invention provide an adaptive filtering architecture for validating such source hypotheses, extracting an estimated representation of source signals corresponding to valid hypotheses, and improving the separation of the remaining 'hidden' source signals from the sensor response mixtures.

Signal source separation partially based on non-sensor information

A system processes data signals consisting of sums of independent signal terms in order to generate one or more of those terms. The generated terms are computed using a set of reference signals to construct alternative support sets, whose images are computed on the data signals. Computed images that are independent of the residue of the data signal minus the image are identified as independent signal terms, as are the residues. These initial signal terms are used to compute additional terms. Identified terms from different data signals are organized into independent slices, and slices whose terms are supported by sets of reference signals are associated with those supporting sets.

Decomposing data signals into independent additive terms using reference signals

The available audio tracks of a multi-track television program are processed to remove or reduce certain components of the program, especially audience sounds, such as laughter. Such processing may be particularly applicable to 'canned' laughter. This applies to any program which is supplied with a plurality of audio channels, whether provided by cable, DVD, Internet streaming, or otherwise.

Growing community of inventors

Framingham, MA, United States of America

Keith Gilbert