Certain characters may not be supported and there is a character limit on text to speech usage. You may find an improvement by breaking up larger text into several smaller chunks.
The text to speech elements also allow use of Speech Synthesis Markup Language (SSML). For comprehensive documentation about SSML, W3C has a recommendation at version 1.0. We currently support the following elements:
p and s
The use of p and s elements is optional. Where text occurs without an enclosing p or s element CallFlow will attempt to determine the structure using language-specific knowledge of the format of plain text.
<text>
<p>
<s>This is the first sentence of the paragraph. </s>
<s>Here's another sentence. </s>
</p>
</text>
voice
We currently offer several voices and languages for use with CallFlow, below is a list of valid voices. See the below example for details on how to use a voice.
-
Nigella - English female
-
Pauline - American female
-
Priya - Indian female
-
Cory - Irish female
-
Heidi - German female
-
Adriana - Spanish female
-
Gabrelle - French female
Alternatively you can select a language from the list below. If the name attribute is present, this one will be ignored
- en-GB - British English
- en-US - American English
- en-IN - Indian English
- en-IE - Irish
- de-DE - German
- es-ES - Spanish
- fr-FR - French
<text>
<voice name="pauline">Speak in an American voice.</voice>
<voice xml:lang="fr-FR">Parler en Francais.</voice>
</text>
N.B. certain characters may not be supported and there is a character limit on text to speech usage.
break
Break is an empty element you may use to insert a pause into the speech. It supports two optional attributes, time or strength. Strength may be one of these following values: "none", "x-weak", "weak", "medium" (default value), "strong", or "x-strong".
<text>
A medium pause
<break strength="medium">
followed by a three second pause
<break time="3s">
bye.
</text>
emphasis
Text enclosed in an emphasis element will have additional prominence or stress. It allows a level attribute. Defined values are "strong", "moderate" (default value), "none" and "reduced".
<text>
Call Flow is
< emphasis level="strong">really</emphasis>
cool!
</text>
say-as
The say-as element helps define the enclosed text content. For example, this can be used to read out currency, time and numbers. It supports two attributes, interpret-as and format. Say-as support varies with language.
<text>
<say-as interpret-as="net">
sdkportal@bt.com
</say-as>
<say-as interpret-as="telephone">
+44123456789
</say-as>
<say-as interpret-as="spell">
spell this out
</say-as>
<say-as interpret-as="time" format="hms">
12:57:43
</say-as>
</text>
The full supported list is:
- acronym
- address
- cardinal - e.g. 12th or (Roman numerals) VIII
- currency - e.g £4.99 or $4.95
- date - formats are: d,dm,dmy,m,md,mdy,my,y,ym,ymd
- decimal
- digits
- duration - h,hm,hms,m,ms,s
- fraction - e.g 1/3 will be read as 'one third'
- letters
- measure - e.g. hm,hr,dB,lb,MHz
- name
- net - e.g. emails or web address
- spell
- telephone
- time - e.g. h,hm,hms
- words - meant for acronyms read as words
prosody
The prosody element permits control of the pitch, speaking rate and volume of the output. All attributes are optional. The permitted attributes are as follows:
- rate - the speed of speech - "x-slow", "slow", "medium", "fast", "x-fast", or "default"
- range - variability of the spoken pitch - "x-low", "low", "medium", "high", "x-high", or "default"
- volume - the volume of the spoken text - "silent", "x-soft", "soft", "medium", "loud", "x-loud", or "default"
<text>
<prosody rate="-50%">
say this slowy
</prosody>
<prosody rate="-50%">
+44123456789
</say-as>
<prosody range="x-low">
say this with low variability
</prosody>
<prosody range="x-high">
say this with high variability
</prosody>
<prosody volume="x-loud">
very loud
</prosody>
<prosody volume="silent">
very quiet
</prosody>
</text>