FIG. 3 to FIG. 5 are explanatory drawings for describing an example of the voice recognition processing (Step T2 of FIG. 2) executed by the voice recognition module 12 of the server apparatus 10. A specific example of the voice recognition processing executed by the voice recognition module 12 of the server apparatus 10 is described with reference to FIG. 3 to FIG. 5.
FIG. 3 is a figure for describing the reliability (weighting) of the devices 20. The reliability of the devices 20 can be calculated by Reliability=Recording level score+Noise level score+Score of influence of echoes. The reliability becomes higher when the score is higher. The recording level has a range of the scores 0 to 5, and the score is higher when the recording volume is larger. The noise level has a range of the scores 0 to 5, and the score is higher when the surrounding noise is less. With respect to the influence of echoes, the score is higher when a sound output from a speaker is lower.
The request module 23 of each device 20 may transmit the information on the recording level, the noise level, and the influence of echoes to the server apparatus 10 as the recording state information, for example. The voice recognition module 12 of the server apparatus 10 may calculate the reliability of each device 20 based on recording state information. The request module 23 of each device 20 may calculate Reliability=Recording level score+Noise level score+Score of influence of echoes, and then transmit the reliability to the server apparatus 10 as the recording state information.
