Synthesize Translations

 

Introduction

The TranslationRecognizer provides transcriptions of spoken input that have been translated, effectively converting spoken language into text. Additionally, you can produce the translation as speech, enabling speech-to-speech translation options. There are two methods available to achieve this.

Event-based Synthesis

When you aim to conduct a 1:1 translation (translating from one source language directly into a single target language), you can utilize event-driven synthesis to capture the translation as an audio stream. To accomplish this, you must indicate the preferred voice for the translated speech in the TranslationConfig. Develop an event handler for the Synthesizing event of the TranslationRecognizer object. Within the event handler, employ the GetAudio() method from the Result parameter to obtain the byte stream of the translated audio.

The exact code for creating an event handler differs depending on the programming language you are using.

Manual Synthesis

Manual synthesis offers a different method to event-based synthesis, eliminating the need for creating an event handler. This technique allows you to produce audio translations for one or multiple target languages.

Manual synthesis of translations is essentially just the combination of two separate operations in which you:

• Use a TranslationRecognizer to translate spoken input into text transcriptions in one or more target languages.

• Iterate through the Translations dictionary in the result of the translation operation, using a SpeechSynthesizer to synthesize an audio stream for each language.

Conclusion

We have successfully learnt about synthesizing translations.