I have the following code:
using (var synthesizer = new SpeechSynthesizer(speechConfig, null))
{
// Subscribes to viseme received event
synthesizer.VisemeReceived += (s, e) =>
{
console.writeLine(e.Animation);
};
var result = await synthesizer.SpeakSsmlAsync(ssml);
}
My ssml variable is as follows (My resource is on the "eastus" region and I'm using the voice "en-US-JaneNeural"):
<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="http://www.w3.org/2001/mstts" xml:lang="en-US">
<voice name="en-US-JaneNeural">
<mstts:viseme type="FacialExpression"/>
</voice>
</speak>
The SpeechSynthesisVisemeEventArgs.Animation is always an empty string. I have tried different regions for the resource and different voices. I have also made sure that the <mstts:viseme type="FacialExpression"/> is in the ssml request, but the animation is always empty.
Is there any other setting I have to change or how can I get the BlendShapes data?