Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Supported SSML Tags

Amazon Polly supports the following SSML tags:

Identifying SSML-enhanced text

<speak>

This tag is supported by long-form, neural, and standard TTS formats.

The <speak> tag is the root element of all Amazon Polly SSML text. All SSML-enhanced text must be enclosed within a pair of <speak> tags.

<speak>Mary had a little lamb.</speak>

Adding a pause

<break>

This tag is supported by long-form, neural, and standard TTS formats.

To add a pause to your text, use the <break> tag. You can set a pause based on strength (equivalent to the pause after a comma, a sentence, or a paragraph), or you can set it to a specific length of time in seconds or milliseconds. If you don't specify an attribute to determine the pause length, Amazon Polly uses the default, which is <break strength="medium"/>, which adds a pause the length of a pause after a comma.

strength attribute values:

time attribute values:

For example:

<speak>
     Mary had a little lamb <break time="3s"/>Whose fleece was white as snow.
</speak>

If you don't use an attribute with the break tag, the result varies depending on text:

Emphasizing words

<emphasis>

This tag is supported only by the standard TTS format.

To emphasize words, use the <emphasis> tag. Emphasizing words changes the speaking rate and volume. More emphasis makes Amazon Polly speak the text louder and slower. Less emphasis makes it speak quieter and faster. To specify the degree of emphasis, use the level attribute.

level attribute values:

For example:

<speak>
     I already told you I <emphasis level="strong">really like</emphasis> that person.
</speak>

Specifying another language for specific words

<lang>

This tag is supported by long-form, neural, and standard TTS formats.

Specify another language for a specific word, phrase, or sentence with the <lang> tag. Foreign language words and phrases are generally spoken better when they are enclosed within a pair of <lang> tags. To specify the language, use the xml:lang attribute. For a complete list of available languages, see Languages Supported by Amazon Polly.

Unless you apply the <lang> tag, all of the words in the input text are spoken in the language of the voice specified in the voice-id. If you apply the <lang> tag, the words are spoken in that language.

For example, if the voice-id is Joanna (who speaks US English), Amazon Polly speaks the following in the Joanna voice without a French accent:

<speak>
     Je ne parle pas français.
</speak>

If you use the Joanna voice with the <lang> tag, Amazon Polly speaks the sentence in the Joanna voice in American-accented French:

<speak>
     <lang xml:lang="fr-FR">Je ne parle pas français.</lang>.
</speak>

Because Joanna is not a native French voice, pronunciation is based on her native language, US English. For example, although perfect French pronunciation features an uvual trill /R/ in the word français, Joanna's US English voice pronounces this phoneme as the corresponding sound /r/.

If you use the voice-id of Giorgio, who speaks Italian, with the following text, Amazon Polly speaks the sentence in Giorgio's voice with an Italian pronunciation:

<speak>
     Mi piace Bruce Springsteen.
</speak>

If you use the same voice with the following <lang> tag, Amazon Polly pronounces Bruce Springsteen in Italian-accented English:

<speak>
     Mi piace <lang xml:lang="en-US">Bruce Springsteen.</lang>
</speak>

This tag can also be used as a substitute for the optional DefaultLangCode option when synthesizing speech. However, doing so requires that you format your text using SSML.

Placing a custom tag in your text

<mark>

This tag is supported by long-form, neural, and standard TTS formats.

To put a custom tag within the text, use the <mark> tag. Amazon Polly takes no action on the tag, but returns the location of the tag in the SSML metadata. This tag can be anything you want to call out, as long as it maintains the following format:

<mark name="tag_name"/>

For example, suppose that the tag name is "animal" and the input text is:

<speak>
     Mary had a little <mark name="animal"/>lamb.
</speak>

Amazon Polly might return the following SSML metadata:

{"time":767,"type":"ssml","start":25,"end":46,"value":"animal"}

Adding a pause between paragraphs

<p>

This tag is supported by long-form, neural, and standard TTS formats.

To add a pause between paragraphs in your text, use the <p> tag. Using this tag provides a longer pause than native speakers usually place at commas or the end of a sentence. Use the <p> tag to enclose the paragraph:

<speak>
     <p>This is the first paragraph. There should be a pause after this text is spoken.</p> 
     <p>This is the second paragraph.</p> 
</speak>

This is equivalent to specifying a pause using <break strength="x-strong"/>.

Using phonetic pronunciation

<phoneme>

This tag is supported by long-form, neural, and standard TTS formats.

To make Amazon Polly use phonetic pronunciation for specific text, use the <phoneme> tag.

Two attributes are required with the <phoneme> tag. They indicate the phonetic alphabet Amazon Polly uses and the phonetic symbols of the corrected pronunciation:

With the <phoneme> tag, Amazon Polly uses the pronunciation specified by the ph attribute instead of the standard pronunciation associated by default with the language used by the selected voice.

For instance, the word "pecan" can be pronounced two ways. In the following example, “pecan” is assigned a different pronunciation in each line. Amazon Polly pronounces pecan as specified in the ph attributes, instead of using the default pronunciation.

International Phonetic Alphabet (IPA)

<speak>
     You say, <phoneme alphabet="ipa" ph="pɪˈkɑːn">pecan</phoneme>. 
     I say, <phoneme alphabet="ipa" ph="ˈpi.kæn">pecan</phoneme>.
</speak>

Extended Speech Assessment Methods Phonetic Alphabet (X-SAMPA)

<speak>
     You say, <phoneme alphabet='x-sampa' ph='pI"kA:n'>pecan</phoneme>.
     I say, <phoneme alphabet='x-sampa' ph='"pi.k{n'>pecan</phoneme>. 
</speak>

Mandarin Chinese uses Pinyin for phonetic pronunciation..

Pinyin

<speak>
     你说 <phoneme alphabet="x-amazon-pinyin" ph="bo2">薄</phoneme>。 
     我说 <phoneme alphabet="x-amazon-pinyin" ph="bao2">薄</phoneme>。
</speak>

Japanese uses Yomigana and Pronunciation Kana.

Yomigana

<speak>
     名前は<phoneme alphabet="x-amazon-yomigana" ph="ひろかず">浩一</phoneme>です。
     名前は<phoneme alphabet="x-amazon-yomigana" ph="ヒロカズ">浩一</phoneme>です。
     名前は<phoneme alphabet="x-amazon-yomigana" ph="Hirokazu">浩一</phoneme>です。
</speak>

Pronunciation Kana

<speak>
     名前は<phoneme alphabet="x-amazon-pron-kana" ph="ヒロ'カズ">浩一</phoneme>です。
</speak>

Controlling volume, speaking rate, and pitch

<prosody>

Prosody tag attributes are fully supported by the standard TTS voices. Neural and long-form voices support the volume and rate attributes, but don't support the pitch attribute.

To control the volume, rate, or pitch of your selected voice, use the prosody tag.

Volume, speech rate, and pitch are dependent on the specific voice selected. In addition to differences between voices for different languages, there are differences between individual voices speaking the same language. Because of this, while attributes are similar across all languages, there are clear variations from language to language and no absolute value is available.

The prosody tag has three attributes, each of which has several available values to set the attribute. Each attribute uses the same syntax:

<prosody attribute="value"></prosody>

The <prosody> tag must contain at least one attribute, but can include more within the same tag.

<speak>
     Each morning when I wake up, <prosody volume="loud" rate="x-slow">I speak  
     quite slowly and deliberately until I have my coffee.</prosody>
</speak>

It can also be combined with nested tags, as follows:

<speak>
     <prosody rate="85%">Sometimes combining attributes <prosody pitch="-10%">can 
     change the impression your audience has of a voice</prosody> as well.</prosody>                  
</speak>

Setting a maximum duration for synthesized speech

<prosody amazon:max-duration>

This tag is currently supported only by the standard TTS format.

To control how long you want a speech to take when it is synthesized, use the <prosody> tag with the amazon:max-duration attribute.

The duration of synthesized speech varies slightly, depending on the voice you select. This can make it difficult to match synthesized speech with visuals or other activities that require precise timing. This issue is magnified for translation applications because the time it takes to say particular phrases can vary widely with different languages.

The <prosody amazon:max-duration> tag matches synthesized speech to the amount of time you want it to take (the duration).

This tag uses the following syntax:

<prosody amazon:max-duration="time duration">

With the <prosody amazon:max-duration> tag, you can specify duration in either seconds or milliseconds:

For example, the following spoken text has a maximum duration of 2 seconds:

<speak>
     <prosody amazon:max-duration="2s">
          Human speech is a powerful way to communicate. 
     </prosody>
</speak>

Text placed within the tag, it doesn't exceed the specified duration. If the chosen voice or language would normally take longer than that duration, Amazon Polly speeds up the speech so that it fits into the specified duration.

If the specified duration is longer than it takes to read the text at a normal rate, Amazon Polly reads the speech normally. It doesn't slow down the speech or add silence, so the resulting audio is shorter than requested.

You can include a single sentence or multiple sentences within a <prosody amazon:max-duration> tag, and you can use multiple <prosody amazon:max-duration> tags within your text.

For example:

<speak>
     <prosody amazon:max-duration="2400ms">
        Human speech is a powerful way to communicate.
     </prosody>
     <break strength="strong"/>
     <prosody amazon:max-duration="5100ms">
        Even a simple ‘Hello’ can convey a lot of information depending on the pitch, intonation, and tempo.
     </prosody>
     <break strength="strong"/>
     <prosody amazon:max-duration="8900ms">
        We naturally understand this information, which is why speech is ideal for creating applications where 
        a screen isn’t practical or possible, or simply isn’t convenient.
     </prosody>
</speak>

Using the <prosody amazon:max-duration> tag can increase latency when Amazon Polly is returns synthesized speech. The degree of latency depends on the passage and its length. We recommend using text comprised of relatively short text passages.

Limitations

There are limitations both in how you use <prosody amazon:max-duration> tag and in how it works with other SSML tags:

Pauses and max-duration

When using max-duration tag, you can still insert pauses within your text. However, Amazon Polly includes the length of the pause when calculating the maximum duration for speech. Additionally, Amazon Polly preserves the short pauses that occur where commas and periods are placed within a passage and includes in the maximum duration.

For example, in the following block, the 600 millisecond break and the breaks caused by the commas and periods occur within the 8-second speech:

<speak>
     <prosody amazon:max-duration="8s">
          Human speech is a powerful way to communicate.
          <break time="600ms"/>
          Even a simple ‘Hello’ can convey a lot of information depending on the pitch, intonation, and tempo.
     </prosody>
</speak>  

Adding a pause between sentences

<s>

This tag is supported by long-form, neural, and standard TTS formats.

To add a pause between lines or sentences in your text, use the <s> tag. Using this tag has the same effect as:

Unlike the <break> tag, the <s> tag encloses the sentence. This is useful for synthesizing speech that is organized in lines, rather than sentence, such as poetry.

In the following example, the <s> tag creates a short pause after both the first and second sentences. The final sentence has no <s> tag, but it is also followed by a short pause because it ends with a period.

<speak>
     <s>Mary had a little lamb</s> 
     <s>Whose fleece was white as snow</s> 
     And everywhere that Mary went, the lamb was sure to go.
</speak>  

Controlling how special types of words are spoken

<say-as>

Except for the characters option, the <say-as> tag is supported by long-form, neural, and standard TTS formats. Note that if Amazon Polly is using a neural voice and encounters the <say-as> tag with the characters option at runtime, the affected sentence will be synthesized using the related standard voice. However, the affected sentence will still be billed as if it uses a neural voice.

Use the <say-as> tag with the interpret-as attribute to tell Amazon Polly how to say certain characters, words, and numbers. This enables you to provide additional context to eliminate any ambiguity on how Amazon Polly should render the text.

The <say-as> tag uses one attribute, interpret-as, which uses a number of possible available values. Each uses the same syntax:

<say-as interpret-as="value">[text to be interpreted]</say-as>

The following values are available with interpret-as:

Fractions

Amazon Polly interprets values within the say-as tag that have the interpret-as="fraction" attribute as common fractions. The following is the syntax for fractions:

Dates

When interpret-as is set to date, you also need to indicate the format of the date.

This uses the following syntax:

<say-as interpret-as="date" format="format">[date]</say-as>

For example:

<speak>
     I was born on <say-as interpret-as="date" format="mdy">12-31-1900</say-as>.
</speak>

The following formats can be used with the date attribute.

Telephone

Amazon Polly attempts to interpret the text you provide correctly based on the text’s formatting even without the <say-as> tag. For example, if your text includes "202-555-1212," Amazon Polly interprets it as a 10-digit telephone number and says each digit individually, with a brief pause for each dash. In this case, you don't need to use <say-as interpret-as="telephone">. However, if you provide the text “2025551212” and want Amazon Polly to say it as a phone number, you would specify <say-as interpret-as="telephone">.

The logic for interpreting each element is language-specific. For example, US and UK English differ in how phone numbers are pronounced (in UK English, sequences of the same digit are grouped together, as in "double five" or "triple four"). To see the difference, test the following example with a US voice and with a UK voice:

<speak>
     Richard's number is <say-as interpret-as="telephone">2122241555</say-as>
</speak>

Pronouncing acronyms and abbreviations

<sub>

This tag is supported by long-form, neural, and standard TTS formats.

Use the <sub> tag with the alias attribute to substitute a different word (or pronunciation) for selected text such as an acronym or abbreviation.

This uses the syntax:

<sub alias="new word">abbreviation</sub>

In the following example, the name "Mercury" is substituted for the element's chemical symbol to make the audio content clearer.

<speak>
     My favorite chemical element is <sub alias="Mercury">Hg</sub>, because it looks so shiny. 
</speak>

Improving pronunciation by specifying parts of speech

<w>

This tag is supported by long-form, neural, and standard TTS formats.

You can use the <w> tag to customize the pronunciation of words by specifying the word’s part of speech or alternate meaning. This is done using the role attribute.

This tag uses the following syntax:

<w role="attribute">text</w>

The following values can be used for the role attribute:

To specify the part of speech:

For example, depending on its part of speech, the US English pronunciation of the word "read" varies based on the tag:

<speak>
     The word <say-as interpret-as="characters">read</say-as> may be interpreted 
     as either the present simple form <w role="amazon:VB">read</w>, or the past 
     participle form <w role="amazon:VBD">read</w>.
</speak>

To specify a specific meaning:

This difference in pronunciation and meaning can be heard if you synthesize the following:

<speak>
    Depending on your meaning, the word <say-as interpret-as="characters">bass</say-as> 
    may be interpreted as either a musical element: bass, or as its alternative meaning, 
    a freshwater fish <w role="amazon:SENSE_1">bass</w>.
</speak>

Adding the sound of breathing

<amazon:breath> and <amazon:auto-breaths>

This tag is supported only by the standard TTS format.

Natural-sounding speech includes both correctly spoken words and breathing sounds. By adding breathing sounds to synthesized speech, you can make it sound more natural. The <amazon:breath> and <amazon:auto-breaths> tags provide breaths. You have the following options:

Manual Mode

In manual mode, you place the <amazon:breath/> tag in the input text where you want to locate a breath. You can customize the length and volume of breaths with the duration and volume attributes, respectively:

To set a breath sound using the defaults, use <amazon:breath/> without attributes.

For example, to use attributes to set the duration and volume for a breath to medium, you would set the attributes as follows:

<speak>
     Sometimes you want to insert only <amazon:breath duration="medium" volume="x-loud"/>a single breath.
</speak>

To use the defaults, you would just use the tag:

<speak>
     Sometimes you need <amazon:breath/>to insert one or more average breaths <amazon:breath/> so that the 
     text sounds correct.
</speak>

You can add individual breathing sounds within a passage, as follows:

<speak>
     <amazon:breath duration="long" volume="x-loud"/> <prosody rate="120%"> <prosody volume="loud"> 
     Wow! <amazon:breath duration="long" volume="loud"/> </prosody> That was quite fast. <amazon:breath 
     duration="medium" volume="x-loud"/> I almost beat my personal best time on this track. </prosody>
</speak>

Automated Mode

In automated mode, you use the <amazon:auto-breaths> tag to tell Amazon Polly to automatically create breathing noises at appropriate intervals. You can set the frequency of the intervals, their volume, and their duration. Place the </amazon:auto-breaths> tag at the beginning of the text that you want to apply automated breathing to and then close the tag at the end.

You can use the following optional attributes with the <amazon:auto-breaths> tag:

By default, the frequency of breathing sounds depends on the input text. However, breathing sounds often occur after commas and periods.

The following examples show how to use the <amazon:auto-breaths> tag. To decide which options to use for your content, copy the applicable examples to the Amazon Polly console and listen to the differences.

Newscaster speaking style

<amazon:domain name="news">

The newscaster style is available only for the Matthew or Joanna voices, which are available only in American English (en-US), Lupe, in US Spanish (es-US) and Amy, in British English (en-GB). It is only supported when using Neural format.

To use the newscaster style, you use SSML tags and the following syntax::

<amazon:domain name="news">text</amazon:domain>

For example, you might use the newscaster style with the Amy voice as follows:

<speak> 
<amazon:domain name="news"> 
From the Tuesday, April 16th, 1912 edition of The Guardian newspaper: 

The maiden voyage of the White Star liner Titanic, the largest ship ever launched, has ended in disaster. 

The Titanic started her trip from Southampton for New York on Wednesday. Late on Sunday night she struck 
an iceberg off the Grand Banks of Newfoundland. By wireless telegraphy she sent out signals of distress, 
and several liners were near enough to catch and respond to the call.
</amazon:domain> 
</speak> 

Adding dynamic range compression

<amazon:effect name="drc">

This tag is supported by long-form, neural, and standard TTS formats.

Depending on the text, language, and voice used in an audio file, the sounds range from soft to loud. Environmental sounds, such as the sound of a moving vehicle, can often mask the softer sounds, which makes the audio track difficult to hear clearly. To enhance the volume of certain sounds in your audio file, use the dynamic range compression (drc) tag.

The drc tag sets a midrange "loudness" threshold for your audio, and increases the volume (the gain) of the sounds around that threshold. It applies the greatest gain increase closest to the threshold, and the gain increase is lessened farther away from the threshold.

This makes the middle-range sounds easier to hear in a noisy environment, which makes the entire audio file clearer.

The drc tag is a Boolean parameter (it's either present or it isn't). It uses the syntax: <amazon:effect name="drc"> and is closed with </amazon:effect>.

You can use the drc tag with any voice or language supported by Amazon Polly. You can apply it to an entire section of the recording, or for only a few words. For example:

<speak>
     Some audio is difficult to hear in a moving vehicle, but <amazon:effect name="drc"> this audio 
     is less difficult to hear in a moving vehicle.</amazon:effect>
</speak>

Using drc with the prosody volume Tag

As the following graphic shows, the prosody volume tag evenly increases the volume of an entire audio file from the original level (dotted line) to an adjusted level (solid line). To further increase the volume of certain parts of the file, use the drc tag with the prosody volume tag. Combining tags doesn't affect the settings of the prosody volume tag.

When you use the drc and prosody volume tags together, Amazon Polly applies the drc tag first, increasing the middle-range sounds (those near the threshold). It then applies the prosody volume tag and further increases the volume of the entire audio track evenly.

To use the tags together, nest one inside the other. For example:

<speak>
     <prosody volume="loud">This text needs to be understandable and loud. <amazon:effect name="drc">
     This text also needs to be more understandable in a moving car.</amazon:effect></prosody> 
</speak>

In this text, the prosody volume tag increases the volume of the entire passage to "loud." The drc tag enhances the volume of the middle-range values in the second sentence.

Speaking softly

<amazon:effect phonation="soft">

This tag is currently supported only by the standard TTS format.

To specify that input text should be spoken in a softer-than-normal voice, use the <amazon:effect phonation="soft"> tag.

This uses the syntax:

<amazon:effect phonation="soft">text</amazon:effect>

For example, you might use this tag with the Matthew voice as follows:

<speak>
     This is Matthew speaking in my normal voice. <amazon:effect phonation="soft">This 
     is Matthew speaking in my softer voice.</amazon:effect>
</speak>

Controlling timbre

<amazon:effect vocal-tract-length>

This tag is currently supported only by the standard TTS format.

Timbre is the tonal quality of a voice that helps you tell the difference between voices, even when they have the same pitch and loudness. One of the most important physiological features that contributes to speech timbre is the length of the vocal tract. The vocal tract is a cavity of air that spans from the top of the vocal folds up to the edge of the lips.

To control the timbre of output speech in Amazon Polly, use the vocal-tract-length tag. This tag has the effect of changing the length of the speaker’s vocal tract, which sounds like a change in the speaker’s size. When you increase the vocal-tract-length, the speaker sounds physically bigger. When you decrease it, the speaker sounds smaller. You can use this tag with any of the voices in the Amazon Polly Text-to-Speech portfolio.

To change timbre, use the following values:

The following example shows how to change the vocal tract length to change timbre:

<speak>
     This is my original voice, without any modifications. <amazon:effect vocal-tract-length="+15%"> 
     Now, imagine that I am much bigger. </amazon:effect> <amazon:effect vocal-tract-length="-15%"> 
     Or, perhaps you prefer my voice when I'm very small. </amazon:effect> You can also control the 
     timbre of my voice by making minor adjustments. <amazon:effect vocal-tract-length="+10%"> 
     For example, by making me sound just a little bigger. </amazon:effect><amazon:effect 
     vocal-tract-length="-10%"> Or, making me sound only somewhat smaller. </amazon:effect> 
</speak>

Combining Multiple Tags

You can combine the vocal-tract-length tag with any other SSML tag that is supported by Amazon Polly. Because timbre (vocal tract length) and pitch are closely connected, you might get the best results by using both the vocal-tract-length and the <prosody pitch> tags. To produce the most realistic voice, we recommend that you use different percentages of change for the two tags. Experiment with various combinations to get the results you want.

The following example shows how to combine tags.

<speak> 
     The pitch and timbre of a person's voice are connected in human speech.
     <amazon:effect vocal-tract-length="-15%"> If you are going to reduce the vocal tract length, 
     </amazon:effect><amazon:effect vocal-tract-length="-15%"> <prosody pitch="+20%"> you 
     might consider increasing the pitch, too. </prosody></amazon:effect>  
     <amazon:effect vocal-tract-length="+15%"> If you choose to lengthen the vocal tract, 
     </amazon:effect> <amazon:effect vocal-tract-length="+15%"> <prosody pitch="-10%"> 
     you might also want to lower the pitch. </prosody></amazon:effect>
</speak>

Whispering

<amazon:effect name="whispered">

This tag is currently supported only by the standard TTS format.

This tag indicates that the input text should be spoken in a whispered voice rather than as normal speech. This can be used with any of the voices in the Amazon Polly Text-to-Speech portfolio.

This uses the following syntax:

<amazon:effect name="whispered">text</amazon:effect>

For example:

<speak>
     <amazon:effect name="whispered">If you make any noise, </amazon:effect> 
     she said, <amazon:effect name="whispered">they will hear us.</amazon:effect>
</speak>

In this case, the synthesized speech spoken by the character is whispered, but the phrase "she said" is spoken in the normal synthesized speech of the selected Amazon Polly voice.

You can enhance the "whispered" effect by slowing down the prosody rate by up to 10%, depending on the effect you want.

For example:

<speak>
     When any voice is made to whisper, <amazon:effect name="whispered">
     <prosody rate="-10%">the sound is slower and quieter than normal speech
     </prosody></amazon:effect>
</speak>

When generating speech marks for a whispered voice, the audio stream must also include the whispered voice to ensure that the speech marks match the audio stream.