Text To Speech Using Cognitive Service Speech API C#


In this article, we will see in detail about how to create our own Text to Speech Application using Cognitive Services. Cognitive Services are a set of machine learning algorithms to build a rich Artificial Intelligence-enabled application. Hope you all are aware of Artificial Intelligence. We can say, iPhone Siri, Windows 10 Cortana, and automatic robotic cars which run on their own are examples of Artificial Intelligence.
Microsoft Cognitive Services (formerly Project Oxford) is a set of APIs, SDKs, and services available to developers to make their applications more intelligent, engaging, and discoverable. Microsoft Cognitive Services expands on Microsoft’s evolving portfolio of Machine Learning APIs and enables the developers to easily add intelligent features – such as emotion and video detection, facial, speech, and vision recognition, speech and language understanding into our applications. Go here for reference.
We will be using Cognitive Service API to develop our Artificial Intelligent application. Cognitive Service API has 5 main categories as -
  1. Vision
  2. Speech
  3. Language
  4. Knowledge
  5. Search
Vision API
In Vision API, we have Computer Vision API for distilling actionable information from images, Face API to detect, identify, analyze, organize, and tag faces in photos, Content Moderator to Automate image, text, and video moderation, Emotion API Preview to personalize user experiences with emotion recognition and Custom Vision Service Preview for easily customize your own state-of-the-art computer vision models for your unique use case.
Speech API
In Speech API, we have Translator Speech API to Easily conduct real-time speech translation with a simple REST API call, Speaker Recognition API Preview for using speech to identify and authenticate individual speakers, Bing Speech API for converting speech to text and back again to understand user intent, Custom Speech Service PREVIEW to overcome speech recognition barriers like speaking style, background noise, and vocabulary
Language API
In Language API, we have Language Understanding (LUIS) to teach our apps to understand commands from our users, Text Analytics API for easily evaluate sentiment and topics to understand what users want, Bing Spell Check API to detect and correct spelling mistakes in your app, Translator Text API to easily conduct machine translation with a simple REST API call, Web Language Model API PREVIEW to use the power of predictive language models trained on web-scale data, Linguistic Analysis API PREVIEW for simplify complex language concepts and parse text with the Linguistic Analysis API
Knowledge API
In Knowledge API, we have Recommendations API PREVIEW to predict and recommend items your customers want, Academic Knowledge API PREVIEW to tap into the wealth of academic content in the Microsoft Academic Graph, Knowledge Exploration Service PREVIEW enable interactive search experiences over structured data via natural language inputs, QnA Maker API PREVIEW distill information into conversational, easy-to-navigate answers, Entity Linking Intelligence Service API PREVIEW will power your app's data links with named entity recognition and disambiguation, Custom Decision Service  PREVIEW is a cloud-based, contextual decision-making API that sharpens with experience
Search API
In Search API, we have Bing Autosuggest API  give your app intelligent autosuggest options for searches, Bing Image Search API is to search for images and get comprehensive results, Bing News Search API is to Search for the news and gets comprehensive results, Bing Video Search API  is to search for videos and get comprehensive results, Bing Web Search API  is to get enhanced search details from billions of web documents, Bing Custom Search API  is an easy-to-use, ad-free, commercial-grade search tool that lets, Bing Entity Search API  PREVIEW to enrich your experiences by identifying and augmenting entity information from the web you deliver the results you want. Ref: 
In this article, we will be seeing in detail how to use the Bing Speech API to read from the text in multiple languages and also save the audio file for later use using the Bing Speech API Cognitive Services.
  1. First, download and install Visual Studio 2017 from this link.
  2. Register yourself for getting the Cognitive Service API keys. https://azure.microsoft.com/en-gb/services/cognitive-services/
  3. After Register from this link get your API KEY https://azure.microsoft.com/en-us/try/cognitive-services/
How to Get Bing Speech API KEY
To work with Cognitive Services, we need to use the API Key which has been given from our Microsoft web site. Check the prerequisites and follow the steps to register and get the API key. Open this URL https://azure.microsoft.com/en-us/try/cognitive-services/ and make sure as you have already signed in to the site and If not then sign in with your ID.
As we are going to work with Bing Speech API, Select the Speech API and then click on the get API Key for Bing Speech API.
Click on the Get API Key for Bing Speech API.
Sign in with your ID
After login, we can see our Bing Speech API Key to be used in our Code developing our Text to Speech application.
Code Part
We will be using Bing text to speech API for developing our Text to Speech application. In this application, we will be using multi-language text to speech by using the locale of the Bing text to speech API. From this link, you can get all the information about Bing Text to Speech API. https://docs.microsoft.com/en-us/azure/cognitive-services/speech/api-reference-rest/bingvoiceoutput This link also has a simple Console application demo program to explain about how to use the Bing text to speech API, we will be using the “TTSProgram.cs” from the sample solution in our application and this class has all the function to perform the text to speech. You can get the class file from this link https://github.com/Azure-Samples/Cognitive-Speech-TTS/blob/master/Samples-Http/CSharp/TTSProgram.cs In our application we will be creating Windows form Application.
Step 1 - Create a Windows Form Application
After installing all the prerequisites listed above, click Start >> Programs >> Visual Studio 2017 >> Visual Studio 2017, on your desktop.
Click New >> Project. Select Visual C# >Select Windows Classic Desktop >> Select Windows Forms App and select your project folder and give your application name and click Ok to create your Windows Form application.
After creating the project now let’s add the “TTSProgram.cs” in our project. Add an Existing Item and select the “TTSProgram.cs” from the attached zip file
Step 2 – Add Controls to your form
In this demo application, I have added 2 Combobox, 2 textboxes, and one button. In the Combox I have added the Locale and Service name mapping for multi-language text to speech recording. Its good to see more than 30 languages can be used for the locale, You can get the complete list of language can be used with Locale and Service name mapping from this link
Here we will be using 3 Language as English, Tamil, and Korean language. In Locale Combobox, we have added the item as “en-US, ko-KR, ta-IN “ and Service name mapping in another combo box item as “Microsoft Server Speech Text to Speech Voice (en-US, ZiraRUS), Microsoft Server Speech Text to Speech Voice (ko-KR, HeamiRUS), Microsoft Server Speech Text to Speech Voice (ta-IN, Valluvar)”
Our Form design look like this
Step 3 – Button Click Event
In the Button Click event. We add our API Key in Authentication section and check for the key is valid and if the API Key is valid then we create an object for synthesizing all the Authentication and Synthesis class has been used from the “TTSProgram.cs” class. Here we have created to 2 events one is to play the Audio after reading the Text and another event to display the error message, then we call the Cortana.Speak method and pass user entered textbox text along with Locale and service name mapping to speak in the user-selected language.
  1. private void btnSpeak_Click(object sender, EventArgs e)  
  2.         {  
  3.             txtstatus.Text = "Starting Authtentication";  
  4.             string accessToken;   
  5.             Authentication auth = new Authentication("AddYourAPIKEYHere");  
  6.             try  
  7.             {  
  8.                 accessToken = auth.GetAccessToken();  
  9.                 txtstatus.Text = "Token: {0} " + accessToken;  
  10.             }  
  11.             catch (Exception ex)  
  12.             {  
  13.                 txtstatus.Text = "Failed authentication.";  
  15.                 txtstatus.Text = ex.Message;  
  16.                 return;  
  17.             }  
  19.             txtstatus.Text = "Starting TTSSample request code execution.";  
  20.             string requestUri = "https://speech.platform.bing.com/synthesize";  
  21.             var cortana = new Synthesize();  
  22.             cortana.OnAudioAvailable += PlayAudio;  
  23.             cortana.OnError += ErrorHandler;    
  24.             cortana.Speak(CancellationToken.None, new Synthesize.InputOptions()  
  25.             {  
  26.                 RequestUri = new Uri(requestUri),  
  27.                 Text = txtSpeak.Text,  
  28.                 VoiceType = Gender.Female,   
  29.                 Locale = cboLocale.SelectedItem.ToString(),   
  30.                 VoiceName = cboServiceName.SelectedItem.ToString(),    
  31.                 OutputFormat = AudioOutputFormat.Riff16Khz16BitMonoPcm,  
  32.                 AuthorizationToken = "Bearer " + accessToken,  
  33.             }).Wait();  
  34.         }  
PlayAudio Event
This event will be triggered when there is a response to reading the text as audio is available. In this method, we get the Audio stream and first we save it in our root folder. Instead of saving the audio, you can also directly play the audio using SoundPlayer class.
  1. private  void PlayAudio(object sender, GenericEventArgs<Stream> args)  
  2.         {   
  3.             Stream readStream = args.EventData;  
  5.             try  
  6.             {  
  7.                 string saveTo = Path.GetDirectoryName(Application.ExecutablePath) + @"\SaveMP3File";  //Folder to Save  
  8.                 if (!Directory.Exists(saveTo))  
  9.                 {  
  10.                     Directory.CreateDirectory(saveTo);  
  11.                 }  
  12.                 string filename = saveTo + @"\Shanu" + DateTime.Now.ToString("yyyyMMddHHmmss") + ".mp3"//Save the speech as mp3 file in root folder  
  14.                 FileStream writeStream = File.Create(filename);  
  16.                 int Length = 256;  
  17.                 Byte[] buffer = new Byte[Length];  
  18.                 int bytestoRead = readStream.Read(buffer, 0, Length);  
  19.                 while (bytestoRead > 0)  
  20.                 {  
  21.                     writeStream.Write(buffer, 0, bytestoRead);  
  22.                     bytestoRead = readStream.Read(buffer, 0, Length);  
  23.                 }  
  25.                 readStream.Close();  
  26.                 writeStream.Close();  
  27.                 SoundPlayer player = new System.Media.SoundPlayer(filename);  
  28.                 player.PlaySync();  
  30.             }  
  31.             catch (Exception EX)  
  32.             {  
  33.                 txtstatus.Text = EX.Message;  
  34.             }  
  35.             args.EventData.Dispose();  
  37.         }  
Step 4 – Build and Run the Application
Text to Speech in the English Language
We have selected the Locale as “en-US” and entered text to save as speech audio. When we click on the button audio file will be created in our root folder.
Text to Speech in the Tamil Language
We have selected the Locale as “ta-IN” and entered text to save as speech audio. When we click on the button audio file will be created in our root folder with the Tamil Language as speech.
Text to Speech in the Korean Language
We have selected the Locale as “ko-KR” and entered text to save as speech audio. When we click on the button audio file will be created in our root folder with the Korean Language as speech.
We can also directly play the audio from the saved mp3 format file in our root folder. We can see as now we have 3 audio files as English, Tamil and in the Korean Language.


Hope you like this article. We will be seeing more articles related to Microsoft Cognitive Services in upcoming days.

Similar Articles