Voice customization method and device thereof, voice recognition method and device thereof

Abstract

The invention discloses a voice customizing method and device, and a voice identification method and device. The voice customizing method and device can be used for correlating different application instructions in different application scenarios with same voice so as to realize one voice customization, achieve application of the voice to the multiple application scenarios and improve the voice customizing efficiency and the utilization rate. The voice customizing method comprises the steps of recording the voice, identifying the recorded voice to obtain a voice message, building the association of an application instruction selected from a preset application instruction set for the voice by a user and the voice message and storing the association, wherein one application instruction in the application instruction set corresponds to one application scenery.

Description

technical field [0001] The invention relates to the communication field, in particular to a voice customization method and its device, a voice recognition method and its device. Background technique [0002] At present, in the voice customization process, the voice recognition system can only simply analyze the user's voice characteristics, and match the analysis results with the preset corpus to recognize the recorded voice, and assign a unique voice to the voice after the recognition is successful. corresponding application instructions. [0003] The voice customization method provided by the prior art has the following technical defects: [0004] Technical defect 1. For a voice, assign a unique corresponding application command to it. Therefore, the application scenario is relatively single; for different application commands, the unique corresponding voice needs to be recorded separately. When the number of application commands is large, the voice recording work is more...

AI Technical Summary

Problems solved by technology

[0004] Technical defect 1. For a voice, assign a unique corresponding application command to it. Therefore, the application scenario is relatively single; for different application commands, the unique corresponding voice needs to be recorded separately. When the number of application commands is large, the voice recording work is more complicated. low efficiency and utilization

[0005] Technical defect 2. The voice recognition system analyzes the user's voice characteristics, and matches the analysis results with the preset corpus to recognize the voice. This recognition method requires the voice recognition system to spend a long time for training. Reach a certain speech recognition rate, and if the recorded speech is mixed with external environmental noise, it is difficult to accurately recognize the speech, so the speech recognition rate is low

Images