US 10,896,021 B2 · Dynamically preventing audio underrun using machine learning

Abstract

The disclosure is directed to a process that can predict an audio glitch, and then attempt to preempt the audio glitch. The process can monitor the systems, processes, and execution threads on a larger system or device, such as a mobile device or an in-vehicle device. Using a learning algorithm, such as deep neural network (DNN), the information collected can generate a prediction of whether an audio glitch is likely to occur. An audio glitch can be an audio underrun condition. The process can use a second learning algorithm, which also can be a DNN, to generate recommended system adjustments that can attempt to prevent the audio glitch from occurring. The recommendations can be for various systems and components on the device, such as changing the processing system frequency, the memory frequency, and the audio buffer size. After the audio underrun condition has abated, the system adjustments can be reversed fully or in steps to return the system to its state prior to the system adjustments.

Claims22

Dependent claims

A method to correct for an audio underrun condition on an audio system, comprising:

analyzing environment parameters of said audio system, wherein said environment parameters include a processing system frequency and a memory frequency;
predicting a probability of an audio underrun condition utilizing said analyzing, wherein said predicting includes:
evaluating said environment parameters utilizing a deep neural network (DNN) trained to detect said audio underrun condition;
inferring, by said DNN, said probability of said audio underrun condition utilizing said evaluating;
comparing said probability of said audio underrun condition to an audio underrun threshold;
and
predicting said audio underrun condition when said probability meets or exceeds said audio underrun threshold;
and
when said probability of said audio underrun condition meets or exceeds said audio underrun threshold,
determining one or more audio correction parameters using said predicting;
and
adjusting at least one environment parameter from a respective original value utilizing said one or more audio correction parameters.

9 more elements

A computer program product having a series of operating instructions stored on a non-transitory computer-readable medium that directs a data processing apparatus when executed thereby to perform audio underrun correction operations on an audio system, said operations comprising:

analyzing environment parameters of said audio system, wherein said environment parameters include a processing system frequency and a memory frequency;
predicting a probability of an audio underrun condition utilizing said analyzing, wherein said predicting includes:
evaluating said environment parameters utilizing a deep neural network (DNN) trained to detect said audio underrun condition;
inferring, by said DNN, said probability of said audio underrun condition utilizing said evaluating;
comparing said probability of said audio underrun condition to an audio underrun threshold;
and
predicting said audio underrun condition when said probability meets or exceeds said audio underrun threshold;
and
when said probability of said audio underrun condition meets or exceeds said audio underrun threshold,
determining one or more audio correction parameters using said predicting;
and
adjusting at least one environment parameter from a respective original value utilizing said audio correction parameters.

9 more elements

An audio correction system, comprising:

a receiver operable to receive an audio portion;
and
an audio system operable to play said audio portion, wherein said audio system comprises:
an audio processor operable to store said audio portion in an audio buffer and output said audio portion,
a processing unit operable to execute one or more processing threads and control said audio system, and
an audio correction processor operable to generate an analysis result of environment parameters of said audio system and said processing unit, generate a prediction result of an occurrence of an audio underrun condition utilizing said analysis result, and adjust said audio system and said processing unit utilizing said analysis result and said prediction result, wherein said environment parameters include a processing system frequency and a memory frequency, and said audio correction processor is operable to generate said prediction result by:
receiving said environment parameters from said audio system as input in a deep neural network (DNN) trained to detect said audio underrun condition;
inferring, by said DNN, a probability of said audio underrun condition utilizing said environment parameters;
comparing said probability of said audio underrun condition to an audio underrun threshold;
and
predicting said audio underrun condition when said probability exceeds said audio underrun threshold.

8 more elements

Dependencies are inferred from the claim text.

Dynamically preventing audio underrun using machine learning

Life of this patent

Dynamically preventing audio underrun using machine learning

Life of this patent

Abstract

Claims22