Testing method based on batch speech recognition and TTS text synthesis

Abstract

The invention discloses a testing method based on batch speech recognition and TTS text synthesis. The method comprises the steps: creating a software environment in a to-be-tested intelligent terminal, automatically operating a test APP, completing the ASR function testing and TTS function testing on the to-be-tested intelligent terminal, performing the ASR reverse speech recognition function testing after the TTS function testing, and finally forming a clear and reliable comparison result, and providing help for the later improvement of an intelligent voice system in the to-be-tested intelligent terminal. The testing mode integrates the whole process of speech recognition and text synthesis from the transmission of test data to the result statistics, covers the multi-scene testing of thespeech recognition and text synthesis, improves the speech recognition testing speed and range, guarantees the applicability and reliability of the speech recognition application environment of the intelligent terminal later, is flexible in testing, and can achieve the testing of the intelligent speech control systems of various types of intelligent terminals at any time and any places.

Description

technical field [0001] The invention relates to a voice recognition testing method, in particular to a testing method targeting intelligent terminals based on batch voice recognition and TTS text synthesis, which belongs to the field of intelligent terminal control. Background technique [0002] With the maturity of voice technology, many smart terminals controlled by voice have begun to appear in our lives, such as smart speakers, smart phone voice assistants, smart home appliances, smart toys, smart cars, etc. These products control devices based on voice , making the use process more convenient and more advantageous in experience. [0003] At present, smart terminals mainly use automatic speech recognition technology (ASR) for speech-to-text testing, or use text synthesis technology (TTS) for text-to-speech testing. The testing process is relatively simple, and the testing scope is mainly limited to existing products. In the standard function test, the test target is s...

AI Technical Summary

Problems solved by technology

[0003] At present, smart terminals mainly use automatic speech recognition technology (ASR) for speech-to-text testing, or use text synthesis technology (TTS) for text-to-speech testing. The testing process is relatively simple, and the testing scope is mainly limited to existing products. In the standard function test, the test target is simple and the test form is scarce. The accuracy and reliability of speech recognition in a simple environment can meet the requirements of use. However, with the complexity of the use environment, the speech recognition process is subject to greater challenges. The security of intelligent terminal use in the environment brings great hidden dangers