Azure TTS API: A Developer's Guide to Text-to-Speech

1Go to your Speech Service resource in the Azure portal.
2
3In the "Resource Management" section, click on "Keys and Endpoint".
4
5You will find two keys (KEY 1 and KEY 2) and the Endpoint URL. Keep these values safe, as they are required to access the Azure TTS API.
6

1import azure.cognitiveservices.speech as speechsdk
2
3def synthesize_speech(text):
4    speech_config = speechsdk.SpeechConfig(subscription="YOUR_SPEECH_KEY", region="YOUR_SPEECH_REGION")
5
6    # Set the voice name.  
7    speech_config.speech_synthesis_voice_name='en-US-JennyNeural' # Example Voice
8
9    audio_config = speechsdk.audio.AudioOutputConfig(use_default_speaker=True)
10
11    # Creates a speech synthesizer using the speech configuration and audio config
12    speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)
13
14    # Get text from the console and synthesize to the default speaker
15    speech_synthesis_result = speech_synthesizer.speak_text_async(text).get()
16
17    if speech_synthesis_result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
18        print("Speech synthesized for text [{}]".format(text))
19    elif speech_synthesis_result.reason == speechsdk.ResultReason.Canceled:
20        cancellation_details = speech_synthesis_result.cancellation_details
21        print("Speech synthesis canceled: {}".format(cancellation_details.reason))
22        if cancellation_details.reason == speechsdk.CancellationReason.Error:
23            print("Error details: {}".format(cancellation_details.error_details))
24
25if __name__ == '__main__':
26    text_to_speak = "Hello, world! This is Azure Text-to-Speech."
27    synthesize_speech(text_to_speak)
28

1import azure.cognitiveservices.speech as speechsdk
2
3def synthesize_speech_with_params(text):
4    speech_config = speechsdk.SpeechConfig(subscription="YOUR_SPEECH_KEY", region="YOUR_SPEECH_REGION")
5
6    # Set the voice name.
7    speech_config.speech_synthesis_voice_name = 'en-US-JennyNeural'
8
9    # Configure speech synthesis output format.
10    speech_config.speech_synthesis_output_format = speechsdk.SpeechSynthesisOutputFormat.Audio16Khz128KBitRateMonoMp3
11
12    audio_config = speechsdk.audio.AudioOutputConfig(use_default_speaker=True)
13
14    speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)
15
16    speech_synthesis_result = speech_synthesizer.speak_text_async(text).get()
17
18    if speech_synthesis_result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
19        print("Speech synthesized for text [{}]".format(text))
20    elif speech_synthesis_result.reason == speechsdk.ResultReason.Canceled:
21        cancellation_details = speech_synthesis_result.cancellation_details
22        print("Speech synthesis canceled: {}".format(cancellation_details.reason))
23        if cancellation_details.reason == speechsdk.CancellationReason.Error:
24            print("Error details: {}".format(cancellation_details.error_details))
25
26
27if __name__ == '__main__':
28    text_to_speak = "Hello, world! This is Azure Text-to-Speech with custom parameters."
29    synthesize_speech_with_params(text_to_speak)
30

1<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="http://www.w3.org/2001/mstts" xml:lang="en-US">
2    <voice name="en-US-JennyNeural">
3        Here are <emphasis level="strong">three</emphasis> things to know about <break time="300ms"/> Azure TTS.
4        <mstts:expressive-as style="cheerful">
5            It's really amazing!
6        </mstts:expressive-as>
7    </voice>
8</speak>
9

1const sdk = require("microsoft-cognitiveservices-speech-sdk");
2
3async function synthesizeSpeech(text) {
4    const speechConfig = sdk.SpeechConfig.fromSubscription("YOUR_SPEECH_KEY", "YOUR_SPEECH_REGION");
5    speechConfig.speechSynthesisVoiceName = "en-US-JennyNeural";
6
7    const audioConfig = sdk.AudioConfig.fromAudioFileOutput("output.wav");
8
9    const synthesizer = new sdk.SpeechSynthesizer(speechConfig, audioConfig);
10
11    synthesizer.speakTextAsync(text,
12        function (result) {
13        if (result.reason === sdk.ResultReason.SynthesizingAudioCompleted) {
14            console.log("Speech synthesized to [" + outputFile + "]");
15        } else if (result.reason === sdk.ResultReason.Canceled) {
16            console.log("Speech synthesis canceled: " + result.errorDetails);
17        }
18        synthesizer.close();
19        synthesizer = null;
20    },
21        function (error) {
22        console.log(error);
23        synthesizer.close();
24        synthesizer = null;
25    });
26}
27
28synthesizeSpeech("Hello, world! This is Azure Text-to-Speech in JavaScript.");
29

1using Microsoft.CognitiveServices.Speech;
2
3public async Task SynthesizeSpeechAsync(string text)
4{
5    var config = SpeechConfig.FromSubscription("YOUR_SPEECH_KEY", "YOUR_SPEECH_REGION");
6    config.SpeechSynthesisVoiceName = "en-US-JennyNeural";
7
8    using (var synthesizer = new SpeechSynthesizer(config))
9    {
10        var result = await synthesizer.SpeakTextAsync(text);
11
12        if (result.Reason == ResultReason.SynthesizingAudioCompleted)
13        {
14            Console.WriteLine($"Speech synthesized to speaker for text "{text}"");
15        }
16        else if (result.Reason == ResultReason.Canceled)
17        {
18            var cancellation = SpeechSynthesisCancellationDetails.FromResult(result);
19            Console.WriteLine($"CANCELED: Reason={cancellation.Reason}");
20
21            if (cancellation.Reason == CancellationReason.Error)
22            {
23                Console.WriteLine($"CANCELED: ErrorCode={cancellation.ErrorCode}");
24                Console.WriteLine($"CANCELED: ErrorDetails={cancellation.ErrorDetails}");
25                Console.WriteLine($"CANCELED: Did you set the speech resource key and region values?");
26            }
27        }
28    }
29}
30