Method for synthetic output with prompting sound and text sound in speech synthetic system

Abstract

In this method, a customer supplies the synthetic information to the synthesis system. The synthesis system, by analyzing the text and the rules of the rhyming and applying the pronunciation element in the library, combines the text with sound data. The invention constructs the prompting sound library to set the prompting sound parameter at the customer-developing interface of the sound synthesis system. The customer can add the expected prompting sound into the library to achieve the universal management of the prompting sound.

Description

technical field [0001] The invention relates to the field of speech synthesis, in particular to a method for managing and applying prompt speech during the conversion process from text to natural speech by using a computer. Background technique [0002] In telephone voice systems such as IVRs and call centers, business processes often require the use of pre-recorded prompts and synthetic voices at the same time. The pre-recorded prompts are recorded by real people, the effect is natural, and it can reflect more emotional styles, giving users a cordial feeling. Although the synthetic voice is clear and accurate, there is still a certain gap between the tone and emotion of the real person recording. In a specific application, the pre-recorded voice is used to broadcast relatively fixed content in the voice service system, usually greetings entering the system and prompts for system operation methods. Synthetic speech is used to broadcast texts whose content changes frequentl...

AI Technical Summary

Problems solved by technology

[0004] First, in the process of system integration, two sets of interfaces are usually used to call and play prompt tone and synthesized voice respectively. In the case of relatively complex services provided by the telephone voice system, especially when prompt voice and synthesized voice appear alternately, In the case of frequent switching, the writing of business processes is particularly complicated, which also increases the workload of integrated development

[0005] Second, prompt voices are generally stored and managed in the form of voice files. A voice prompt is saved as a voice file, and generally needs to be converted from a certain voice data format to a specified voice data format. The number of files is large, and management is inconvenient. Matching It is also difficult, and it is easy to make mistakes during the conversion process

[0006] Third, since the prompt tone is a pre-recorded voice, it has different characteristics from the synthesized voice in many aspects such as energy. Therefore, if the prompt tone and the synthesized voice are simply spliced ​​together, the difference between the two voices will be obvious. There will be sound skipping and other phenomena, which will affect the overall effect of the broadcast

[0007] In addition, since the prompt voice file and the synthesized voice file are stored separately, the telephone voice system only has the voice file of the prompt voice, which cannot fully display the entire service process and content from the text perspective, and cannot splicing, adjusting and optimizing the prompt voice and synthesized voice deal with

Images