System and method of analyzing audio data samples associated with speech recognition

Abstract

A particular apparatus includes a first buffer that is configured to store multiple audio data samples and a second buffer that is configured to store the multiple audio data samples. The first buffer is coupled to a first processor that is configured to analyze audio data samples to detect a keyword. The second buffer is coupled to a second processor that is configured to initialize a speech recognition engine (SRE) based on the multiple audio data samples. The first buffer has less storage capacity that the second buffer.

Description

I. CROSS-REFERENCE TO RELATED APPLICATIONS[0001]The present application claims priority from and is a continuation application of U.S. Non-Provisional patent application Ser. No. 14 / 573,402, filed Dec. 17, 2014 and entitled “SYSTEM AND METHOD OF ANALYZING AUDIO DATA SAMPLES ASSOCIATED WITH SPEECH RECOGNITION,” the contents of which are incorporated herein by reference in their entirety.II. FIELD[0002]The present disclosure is generally related to speech recognition.III. DESCRIPTION OF RELATED ART[0003]Advances in technology have resulted in smaller and more powerful computing devices. For example, there currently exist a variety of portable personal computing devices, including wireless telephones such as mobile and smart phones, tablets and laptop computers that are small, lightweight, and easily carried by users. These devices can communicate voice and data packets over wireless networks. Further, many such devices incorporate additional functionality such as a digital still camer...

AI Technical Summary

Benefits of technology

[0008]The CODEC may continue to buffer the audio data until the bus between the application processor and the CODEC is active. When the bus is ready, the audio data at the first buffer may be transferred via the bus to a second buffer at the application processor. The first buffer may have less storage capacity than the second buffer (e.g., to reduce cost associated with the first buffer). For example, while the second buffer may be capable of storing audio data associated with an entire command phrase, the first buffer may not have sufficient capacity to store the entire command phrase. After the keyword is detected and while the speech recognition engine is being prepared, audio data (e.g., portions of the command phrase) may continue to be received at the mobile device. The audio data may be buffered at the first buffer and transferred, in a first in first out manner, from the first buffer to the second buffer via the bus. Thus, the first buffer need not be large enough to store the entire command phrase. Rather, it is sufficient for the first buffer to be large enough to store audio data that is received during a time for the bus to be initialized.

[0014]One particular advantage provided by at least one of the disclosed embodiments is that buffering audio data at the CODEC before providing the audio data to a second buffer at the application processor, as described herein, allows the user to conveniently provide a keyword / command phrase sentence (without waiting for the mobile device to wake up and provide a prompt). Additionally, cost of the mobile device is not significantly increased by this arrangement because a relatively low cost buffer can be used at the CODEC since the buffer at the CODEC does not need to be large enough to store the entire command phrase.

Problems solved by technology

Because of the time it takes to prepare the speech recognition capabilities to receive input, there can be a delay between when the user speaks to wake up the mobile device and when the user is able to provide the command.

For example, while the second buffer may be capable of storing audio data associated with an entire command phrase, the first buffer may not have sufficient capacity to store the entire command phrase.

Images