Abnormal sound specifying device, method of specifying abnormal sound, and nontransitory computer-readable storage medium storing computer-readable instructions for arithmetic device

A method for specifying abnormal sound from sound recorded at a vehicle using an arithmetic device and an output device, wherein

• the arithmetic device is configured to access a learned model of artificial intelligence,
• the method comprises:
• the arithmetic device specifying frequency-time data that indicates a chronological change in frequency spectrum of the sound;
• the arithmetic device inputting the specified frequency-time data into the learned model to cause the learned model to specify a type of abnormal sound included in the sound based on the inputted frequency-time data and to cause the learned model to specify a basis range from the inputted frequency-time data, the basis range indicating a frequency range and a time range that are used to specify the type of the abnormal sound;
• the arithmetic device designating a designated range indicating a frequency range and a time range from the specified frequency-time data;
• and
• the arithmetic device determining whether to cause the output device to output the type of the abnormal sound in a determination process, the determination process including, as a determination element, at least a determination on whether the basis range and the designated range overlap each other.

A non-transitory computer-readable storage medium storing computer-readable instructions for an arithmetic device,

• wherein the computer-readable instructions, when executed by a processor of the arithmetic device configured to access a learned model of artificial intelligence and connected to an output device, cause the arithmetic device to:
• specify frequency-time data that indicates a chronological change in frequency spectrum of sound recorded at a vehicle;
• input the specified frequency-time data into the learned model to cause the learned model to specify a type of an abnormal sound included in the sound based on the inputted frequency-time data and to cause the learned model to specify a basis range from the inputted frequency-time data, the basis range indicating a frequency range and a time range that were used to specify the type of the abnormal sound;
• designate a designated range indicating a frequency range and a time range from the specified frequency-time data;
• and
• determine whether to cause the output device to output the type of the abnormal sound in a determination process, the determination process including, as a determination element, at least a determination on whether the basis range and the designated range overlap each other.