A voice evaluation method, device and readable medium

Abstract

The embodiment of the present invention discloses a speech evaluation method, device and readable medium, which relate to the field of speech recognition technology. In the method provided by the present invention, the speech information to be evaluated of any target object is obtained; A speech evaluation model: evaluate the speech information of the target object to be evaluated according to the determined speech evaluation model. In this way, different speech evaluation models can be selected for different objects, so that the determined speech evaluation model is more suitable for the evaluation of the target object's pronunciation, and then a reasonable speech evaluation result is given for the target object, which improves the degree of improvement. The motivation of the target audience to learn the language.

Description

technical field [0001] The invention relates to the technical field of voice recognition, in particular to a voice evaluation method, device and readable medium. Background technique [0002] With economic globalization, people know more and more about the humanities and cultures of various countries, and it is more and more common to learn other languages. In order to meet the language learning needs of the target audience, some voice evaluation software applications are born, such as English Liulishuo application, etc. In the process of learning, whether the learner's pronunciation is standard or the level of pronunciation is the key factor to measure a learner's learning level. [0003] At present, the commonly used speech evaluation method is to use the speech corpus of other countries to train the speech evaluation model, and then evaluate the learner's speech based on the trained speech evaluation model. The speech evaluation model has strict requirements. When the lea...

AI Technical Summary

Problems solved by technology

[0003] At present, the commonly used speech evaluation method is to use the speech corpus of other countries to train the speech evaluation model, and then evaluate the learner's speech based on the trained speech evaluation model. The speech evaluation model has strict requirements. When the learner's pronunciation improves However, when the upgrade result calculated by the speech evaluation model is not reached, the speech evaluation model will still output a result that cannot be upgraded, which seriously affects the enthusiasm of learners, especially beginners, and even leads to frustration for beginners. poor learner experience

In addition, it is difficult to learn the pronunciation of other languages ​​to achieve the pronunciation of native speakers in other countries. Therefore, it is unreasonable to use the above-mentioned pronunciation evaluation method to evaluate the pronunciation of learners.

Images