Microsoft Technology Licensing, LLC

Jeffrey Eames Taylor

Robert D Donner

Christopher Blomfield-Brown

Joseph Cox Ballantyne

Shanmugam Mohanraj

Systems and methods for processing data from priority-based operating system threads within a realtime component are described. The component submits blank buffers to hardware. The component receives data from priority based operating system threads and processes and writes the data from within an independent second real-time thread to buffers before the buffers into which the data is written are used by the hardware. Hardware buffers are created offset from memory page boundaries such that the least significant bits of any memory address referencing the start of the buffer uniquely identify it and can be used as an index into a circular buffer queue to determine which buffer is currently being processed. Data is always processed and written into a buffer that is a predetermined range of buffers in front of the buffer currently being processed by the hardware.

Systems and methods for running priority-based application threads on a realtime component

Modem-equipped computers can initiate an audio channel using the modem data connection. The connection is initiated with a protocol that does not require any additional modem hardware or telephone line features, and is not tied to any proprietary hardware/software compression or transmission schemes. The protocol negotiates an audio data compression/decompression scheme and then establishes an audio channel using a socket. Headers are compressed and decompressed using a predetermined compression scheme. Compressed audio data is delivered to the remote computer where it is decompressed using the negotiated compression/decompression. The protocol significantly reduces the latency which disrupts normal speech patterns when voice data is sent over a data connection. This protocol also reduces the bandwidth required to send voice over a data connection.

Voice/audio data communication with negotiated compression scheme and data header compressed in predetermined scheme

A method for optimizing the design of a digital filter for use in a microprocessor that utilizes sequential multiply and add instructions as distinct from single multiply and accumulate cycles. The method involves alternating the steps of multiplying and adding samples of the input signal and the coefficients that are assigned to the corresponding samples of the input signal. This add and multiply sequence is more efficient on most general-purpose computers than the more common load, multiply, add sequence used for this type of filter operation.

High efficiency digital filter using sequential multiply and add operations

A sample rate converter for converting a digital signal having a particular sample rate frequency into a signal having a different, specified sample rate frequency. The converter includes an interpolation function for increasing the sample rate frequency of the input signal by an interpolation factor (L), so as to produce an intermediate signal having an intermediate sample rate frequency. The intermediate signal is then filtered with a predefined single-stage, low-pass filter to eliminate high frequency noise introduced as a result of the interpolation. The filtered intermediate signal is then supplied to a decimation function, which extracts samples from the intermediate signal in accordance with a decimation factor (M), thereby producing an output signal having the desired sample rate frequency. The filter is configured so as to be optimized to provide the best conversion results for a plurality of critical input/output sample rate conversion pairs, which are used to define the filter cutoff frequency. The converter can be equipped with a single predefined filter, which is then used for all conversions, or the converter can be equipped with a plurality of predefined filters, and the most appropriate filter selected based on the desired level of quality for the conversion. Since the filter is predefined and fixed, there is no need to recalculate the filter every time the converter is executed or when a different input/output rate is specified. In operation, the input signal is interpolated to an intermediate sample rate by inserting L-1 zero value data points between the existing data points. This intermediate signal is then low pass filtered, but the filter skips all of the intervening zero value data points, thereby greatly enhancing the efficiency of the converter. The filtered signal is then decimated, by removing every M-1 out of every M data samples.

Sample rate converter

Modem-equipped computers which can initiate an audio channel using the modem data connection. The connection is initiated with a new protocol called the voice-over-data protocol. The new protocol does not require any additional modem hardware or telephone line features, and is not tied to any proprietary hardware/software compression or transmission schemes. The voice-over-data protocol negotiates an audio compression/decompression scheme and then sets up an audio channel over an existing data connection using a socket. Compressed audio data is then delivered to the remote computer where it is decompressed and output. The voice-over-data protocol significantly reduces the latency which disrupts normal speech patterns when voice data is sent over a data connection. This protocol also reduces the bandwidth required to send voice over a data connection.

Growing community of inventors

Bellevue, WA, United States of America

Jeffrey Eames Taylor