Speech encoding: US Pat. No. 7,660,712

U.S. Patent No. 7,660,712, issued on February 9, 2010 to Mindspeed Technologies, Inc. of Newport Beach, CA, discloses a speech encoder for voice communication systems.

According to the ’712 patent, speech sounds are typically encoded prior to being transmitted to increase the efficiency of bandwidth usage, for example by removing short-term correlations and redundancies.  The ’712 patent discloses using an analysis-by-synthesis coding approach on a speech signal to apply different gain quantization processes to periodic and non-periodic input speech signals.

According to their website, Mindspeed Technologies “designs, develops and sells semiconductor networking solutions for communications applications in enterprise. access, metropolitan and wide area networks.”  According to the USPTO database, Mindspeed was granted 17 U.S. patents in 2009 and 17 in 2008. 

Mindspeed’s financial statements over the past few years show that the company continues to evaluate its intellectual property portfolio to identify patents which are no longer part of its core business to be used to generate revenue by sale or licensing.  For example, the company used patent sales to raise $12.0 million in 2008 and $3.0 million in the first quarter of fiscal 2009.  Such activity is a good way to monetize intellectual property which would otherwise be lying dormant.