All Products
Search
Document Center

FAQ

Last Updated: Sep 30, 2020

This topic provides answers to commonly asked questions about Speech Synthesis Markup Language (SSML).

How can I use SSML in long-text-to-speech synthesis?

  • If you need to control the breaks in the long text to be synthesized, we recommend that you first make pauses based on the punctuation marks in the text, such as periods (.), exclamation points (!), and question marks (?). Then, add SSML-based tags to the text as needed. The following example is provided for your reference.

    Original text:

    As the college entrance examination is approaching, many students who are about to take the exam are prone to anxiety symptoms at different levels, such as emotional irritability, memory decline, difficulty in concentration, anorexia, and insomnia.
    How can students overcome anxiety about the forthcoming exam?
    Yu Minhong, the president of New Oriental Education & Technology Group Inc., offered 6 suggestions on overcoming anxiety in an online video recorded in NetEase Cloud Classroom.
    Yu Minhong: Hello, my dear students! In this tough time as we keep struggling with COVID-19, the gaokao has been postponed by one month. Before long you will participate in the gaokao of this year.
    I understand that you feel more or less nervous because the gaokao may have a lasting impact on your life. The exam results directly determine the universities that you are admitted to.
    So, here today, I would like to make a few suggestions to help you better prepare for the gaokao.

    Text added with SSML-based tags:

    <speak>As the college entrance examination is approaching, many students who are about to take the exam are prone to anxiety symptoms at different levels, such as emotional irritability, memory decline, difficulty in concentration, anorexia, and insomnia.<break time="700ms"/></speak>
    <speak>How can students overcome anxiety about the forthcoming exam? Yu Minhong, the president of New Oriental Education & Technology Group Inc., offered 6 suggestions on overcoming anxiety in an online video recorded in NetEase Cloud Classroom.<break time="700ms"/></speak>
    <speak>Yu Minhong:<break time="400ms"/>Hello, my dear students!<break time="700ms"/></speak>
    <speak>In this tough time as we keep struggling with COVID-19, the gaokao has been postponed by one month. Before long you will participate in the gaokao of this year.<break time="700ms"/></speak>
    <speak>I understand that you feel more or less nervous because the gaokao may have a lasting impact on your life. The exam results directly determine the universities that you are admitted to.<break time="700ms"/></speak>
    <speak>So, here today, I would like to make a few suggestions to help you better prepare for the gaokao.<break time="700ms"/></speak>

  • If you need to modify the pronunciation of a heteronym in the text to be synthesized, we recommend that you add SSML-based tags to the text segment that contains the heteronym. The text segment between two commonly used punctuation marks is regarded as the proper segment to which tags can be added to. Original text:

    According to the plan issued by the Guangzhou Land Resource and Planning Committee, the school will cover an area of 59,593 square meters and its main entrance will be on Tiankun 2nd Road. In the eastern part of the school, there will be a five-storey multi-use building for elementary education, a five-storey laboratory and teaching building for secondary education, an eleven-storey dormitory, and a six-storey administrative building.

    Text added with SSML-based tags:

    According to the plan issued by the Guangzhou Land Resource and Planning Committee, the school will cover an area of 59,593 square meters and its main entrance will be on Tiankun 2nd Road. In the eastern part of the school, there will be a five-storey multi-use building for elementary education, a five-storey laboratory and teaching building for secondary education,
    <speak>an eleven-storey dormitory, and a six-storey <phoneme alphabet="py" ph="xing2">administrative</phoneme> building. </speak>