You can control the voice for announcements (Welcome Message, Ask for Reason, Voicemail and Announcement Only) with different tags by using SSML (Speech Synthesis Markup Language).
We have listed and explained the most important ones. You just have to copy paste the examples and change the text input to meet your needs.
For further examples, you can also see: http://docs.aws.amazon.com/polly/latest/dg/supported-ssml.html.
Break (<break> ): This command inserts a break wherever you want (in between phrases, sentences, etc.).
You can set it to specific seconds and milliseconds:
<speak>Thank you for calling. <break time="3s"/>We appreciate your business.</speak> <speak>Thank you for calling. <break time="100ms"/>We appreciate your business.</speak>
Or to a specific break strength:
<speak>Thank you for calling. <break strength="none"/>We appreciate your business.</speak> <speak>Thank you for calling. <break strength="x-weak"/>We appreciate your business.</speak> <speak>Thank you for calling. <break strength="weak"/>We appreciate your business.</speak> <speak>Thank you for calling. <break strength="medium"/>We appreciate your business.</speak> <speak>Thank you for calling. <break strength="strong"/>We appreciate your business.</speak> <speak>Thank you for calling. <break strength="x-strong"/>We appreciate your business.</speak>
Lang (<lang> ): With this command, a specific word or phrase is spoken in a different language’s tone. In this example, an Italian voice is used. However the name “Bruce Springsteen” should be pronounced in American English:
<speak>Mi piace Bruce Springsteen.</speak>
sounds differently than
<speak>Mi piace <lang xml:lang="en-US">Bruce Springsteen.</lang></speak>
Phoneme (<ph> ): The Yodel voice can spell a specific word or expression (e.g. your brand’s name) in phonetic pronunciation. We support “Ipa” and “X-Sampa” spelling. Please make sure to also add the default text of that word or expression within the tags:
<speak> You say, <phoneme alphabet="ipa" ph="pɪˈkɑːn">pecan</phoneme>. I say, <phoneme alphabet="ipa" ph="ˈpi.kæn">pecan</phoneme>. </speak>
Say-as (<say-as> ): This tag indicates how the input text should be interpreted (e.g. spell out each letter, pronounce each digit of a number, etc.)
<speak>I want to order <say-as interpret-as="cardinal">1,234</say-as> units</speak> <speak>The <say-as interpret-as="original">1,234</say-as> unit is damaged</speak> <speak>Richard's number is <say-as interpret-as="digits">2122241555</say-as></speak> <speak>Richard's number is <say-as interpret-as="fraction">2 1/2</say-as></speak> <speak>Richard's number is <say-as interpret-as="unit">1meter</say-as></speak> <speak>Today is <say-as interpret-as="date" format="dm">4/10</say-as></speak> <speak>Please hold on for <say-as interpret-as="time">1'21"</say-as></speak> <speak>Our office is located at <say-as interpret-as="address">44</say-as> Tehama Street</speak> <speak>Richard's number is <say-as interpret-as="expletive">21</say-as>22241555</speak> <speak>Richard's number is <say-as interpret-as="telephone">2122241555x345</say-as></speak>