Speech Languages reference

This reference lists the supported languages for:

• Text-to-speech
• Speech Capture and Speech Transcription

Text-to-speech languages

For text-to-speech conversion, we offer more than 100 languages and accents. For each supported languages, we may offer different voices from a gender perspective as well as from a quality perspective (standard versus neural).

Standard voices

• Traditionally, these are based on concatenative synthesis, where speech is generated by stringing together short recorded audio samples of a human speaker. This method relies on a vast library of spoken words, syllables, or phonemes.
• The speech produced is clear and intelligible, but it can sometimes lack the natural fluidity and intonation of human speech. It may sound more robotic or synthetic, particularly in how it handles intonation, stress, and rhythm.
• Standard voices have no usage charges

Neural voices

• These utilize advanced machine learning models, particularly neural networks, to synthesize speech. This method involves training the model on large datasets of recorded speech, enabling it to learn and mimic the nuances of human speech.
• They offer a more natural, lifelike speech quality, closely mimicking the subtleties of human expression, emotion, and accent. The flow of speech is smoother, and it's better at handling complex intonations and varying speech patterns.
• Neural voices are charged by character

SSML support

SSML stands for Speech Synthesis Markup Language. It serves as a powerful tool to finely tune the text-to-speech synthesis process. With SSML, you gain the ability to infuse synthesized speech with natural-sounding inflections, emphasis, pauses, and other speech characteristics. This control allows you to tailor the output to match your unique requirements and create more engaging and lifelike voice experiences for your users.

Infobip supports SSML for :

• Google and Amazon Polly standard voices
• Microsoft Azure neural voices

Microsoft Azure standard voices do not support SSML.

Here is the list of supported tags that can be used across voice API products:

 
    <speak> - Identifying SSML-enhanced text
    <break /> - Adding a pause 
    <say‑as> - Controlling how special types of words are spoken
    <p> - Adding a pause between paragraphs
    <s> - Adding a pause between sentences
    <emphasis> - Emphasizing words
    <sub> - Pronouncing acronyms and abbreviations
    <phoneme> - Using phonetic pronunciation
    <prosody> - Controlling volume, speaking rate, and pitch
 

Google, Amazon, and Microsoft offer other SSML tags, such as par, audio, seq, and so on, but they are not supported in this current implementation.

This is an example of an SSML-enhanced tag text used in a Voice Message API: SSML in a Voice Message API

 
    "text": "<speak>Hello. This is a test message sent from <emphasis level=\"strong\">Infobip Voice Message API</emphasis>. Your confirmation code is <say-as interpret-as=\"spell-out\">12345</say-as>.</speak>"
 

For more information about the complete syntax of these SSML tags, see:

• For Google standard voices: https://cloud.google.com/text-to-speech/docs/ssml?hl=en (opens in a new tab)
• For Amazon Polly voices: https://docs.aws.amazon.com/polly/latest/dg/supportedtags.html (opens in a new tab)
• For Microsoft Azure neural voices: https://learn.microsoft.com/en-us/azure/ai-services/speech-service/speech-synthesis-markup (opens in a new tab)

Arabic

Bengali

Bulgarian

Catalan

Chinese (Simplified)

Croatian

Czech

Danish

Dutch

English

Filipino

Finish

French

German

Greek

Gujarati

Hebrew

Hindi

Hungarian

Icelandic

Indonesian

Italian

Japanese

Kannada

Korean

Malay

Malayam

Norwegian

Polish

Portuguese

Romanian

Russian

Slovak

Slovenian

Spanish

Swedish

Tamil

Telugu

Thai

Turkish

Ukrainian

Vietnamese

Welsh

The languages and voices marked as beta are in a pre-production stage. Infobip does not have the technical support obligation for these languages.

Speech recognition languages

Check out the table to see which abbreviation you need to use when selecting a specific language in the API request.