My Ramblings - About TTS and Speech Synthesis Technologies

Microsoft Speech Platform 11.0

2011-09-12T00:28:00Z

Microsoft Speech Platform has upgraded to Version 11.0. It includes many improvements on SR and TTS engines in the past year.

There is a MSDN web page for more detail information about the speech platform

http://msdn.microsoft.com/en-us/library/hh361572.asp

In particular, for TTS, there is a new voice "ZiraPro" for en-US.

Question about SupportedAudioFormats of SpeechSynthesizer

2011-01-02T07:30:00Z

Happy new year! It is 2011.

Recently I got a customer question regarding "SupportedAudioFormats". He used the Server Speech Platform 10.2. But he found the SupportedAudioFormats of speechsynthesizer is empty.
This is because of some issue in service voice setup. All the service voices currently is in 8k8bit PCM format.

The voice audio format can be specified like below.
speechSynthesizer. SetOutputToWaveFile(“output.wav”, new SpeechAudioFormatInfo(8000, AudioBitsPerSample.Eight, AudioChannel.mono));

if you set audio format to other format like 16k16bit, the speech synthesizer will do resampling for you.

Where to ask for TTS support?

2010-12-01T13:37:00Z

If you are a developer that is looking for some help on using TTS in your project, there is an alias you can try.
speak AT microsoft.com

It is not a fulltime product support contact but you might get some suggestion to your problem. You can also give your feedback about text to speech engines via this email.

TTS in Microsoft Lync

2010-11-06T05:57:33Z

According to http://www.microsoft.com/en-us/lync/extensible-platform.aspx

"State-of-the-art speech technology now in 26 languages provided by the updated Microsoft Speech Platform, a zero-royalty, enterprise grade server based speech platform that ships with UCMA 3.0 and now adds VoiceXML 2.1 to UCMA needs. "

So you can develop UCMA 3.0 based application with TTS function in 26 language now.

What do you use TTS for?

2010-10-20T14:19:06Z

When we think about improving TTS voice quality, we always need to consider the scenerio that TTS are used in.
For example, if the scenerio is accessiblity, TTS needs to be highly intelligible in all conditions.
If the scenerio is reading turn by turn directions, the key information of routing needs to be highlighted and more prominent.

So if you are developing or have an application that uses TTS now, what do you use it for? What kind of expectation do you have on the TTS quality?

TTS in Windows Phone 7

2010-09-19T23:25:00Z

If you got a chance to play a Windows phone 7, you can try out speech feature in it.

It has three cool speech features: voice search, voice activated dialing and start application by command.

In these scenerios, Zira TTS will use to speak the text back. I hope you will love it.

More news about speech for phone 7.

http://www.microsoft.com/presspass/presskits/speech/videoGallery.aspx

http://www.microsoft.com/presspass/features/2010/aug10/08-03gamechangerfea.mspx

http://www.pcworld.com/businesscenter/article/202560/windows_phone_7_to_excel_on_speech.html

FAQ about The Server Speech Platform

2010-07-29T11:29:00Z

This post will talk about some frequent questions when using TTS functionality in Microsoft Server Speech Platform. I will keep it updated as new questions come up.

For the detailed usage of the API, please go to MSDN.

· Why SpeechSynthesizer Speak doesn't produce sound on my sound card?

Currently the server voice doesn’t speak to sound card because it is primarily designed for server usage. It is a difference against System.Speech.Synthesizer.

There are serveral ways for you to play the audio from SpeechSynthesizer.

1. You can speak to wave file first and play that wave file. See SetOutputToWavefile method of SpeechSynthesizer

2. Use a custom output stream to send the audio bytes to your target device (could be audio device or network socket). See SetOutputAudioStream method.

3. Speak to a memory stream, and then play the stream with System.Media.SoundPlayer

· Why SpeechSynthesizer throws an exception that says "No voice installed on the system or none available with the current security setting"?

It could be caused by following reasons:

1. The speech platform runtime is installed, but the language pack is not installed. Make sure you install the TTS language pack.

2. The speech platform and language pack are installed, but the .net application is compiled as a target platform different from the speech platform. For example, speech platform is x64, but the app is x86.

· What are the difference between desktop speech API, and server speech API?
Here are some differences between desktop speech API and server speech API from usage perspective.
1. They have different GUIDs for COM object.
2. They have different name space. (System.Speech for Desktop, Microsoft.Speech for Server)

Use TTS functionality in Microsoft Speech Platform

2010-06-25T13:53:16Z

In previous post, I introduced the microsoft speech platform is available to download. Here is a short introduction on how to use the TTS functionality provided by the platform. The steps are:

1. Download and install the speech platform runtime and one language pack like en-US as you wish

2. Download the Software Development Kit and install it. the SDK is installed into somewhere like C:\Program Files\Microsoft Speech Platform SDK

3. Launch Visual Studio, go to “Tools” -> “options” -> “Project and Solutions” -> “VC++ Directories” (Here I use VS2008)
Add the path like C:\Program Files\Microsoft Speech Platform SDK\include into the “include files” directories (as top one)
Add the path like C:\Program Files\Microsoft Speech Platform SDK\Lib into the “library files” directories (as top one)

4. Unzip the attached sample and build, you should be able to compile it and run. You can hear the voice is speaking to you. This is actually a simple "hello world" example using the native API (SAPI) to call TTS.

Let me know if this works for you or not. Have fun with TTS!

Bing app with Navigation for Windows Mobile released!

2010-06-10T14:24:04Z

Recently Bing app for Windows Mobile has added free turn by turn directions. TTS is used to read out the address during navigation.

http://www.bing.com/community/blogs/search/archive/2010/05/11/updated-bing-app-for-windows-phone.aspx

If you have a windows phone, have a try with it. How do you feel the TTS voice quality in it? Please comment!

Use Microsoft TTS to read office documents with Office 2010

2010-04-15T03:34:00Z

In Office 2010, there is a new cool feature to read the text aloud with Microsoft TTS.

Here is the post about how to use it. It can use the latest speech platform voices in 26 languages.

http://blogs.technet.com/office2010/archive/2010/04/14/hear-in-your-language-in-office-2010.aspx

Have fun with it!

Microsoft Server Speech Platform 10.1 Released (SR and TTS in 26 languages)

2010-03-30T01:02:00Z

Want to build your speech application with speech recognitions and synthesis?

Install the package below, you can now use SR and TTS in 26 language easily.

For the Runtime, please download the Microsoft Speech Platform - Server Runtime 10.1
For Language Packs, please download the Microsoft Speech Platform - Server Runtime Languages 10.1
For the Software Development Kit, please download the Microsoft Speech Platform - Server SDK 10.1

The 26 languages are:

Catalan	Spain	ca-ES
Chinese (Hong Kong)	China	zh-HK
Chinese (Simplified)	China	zh-CHS
Chinese (Traditional)	Taiwan	zh-TW
Danish	Denmark	da-DK
Dutch	Netherlands	nl-NL
English	Australia	en-AU
English	Canada	en-CA
English	India	en-IN
English	United Kingdom	en-GB
English	United States	en-US
Finnish	Finland	fi-Fl
French	Canada	fr-CA
French	France	fr-FR
German	Germany	de-DE
Italian	Italy	it-IT
Japanese	Japan	ja-JP
Korean	Korean	ko-KR
Norwegian (Bokmal)	Norway	nb-NO
Polish	Poland	pl-PL
Portuguese	Brazil	pt-BR
Portuguese	Portugal	pt-PT
Russian	Russia	ru-RU
Spanish	Spain	es-ES
Spanish	Mexico	es-MX
Swedish	Sweden	sv-SE

Feel free to try the TTS in the 26 languages, post your feedbacks about them here.

Bing Translator with Microsoft TTS functionality now!

2010-03-16T14:23:00Z

Bing Translator (http://www.microsofttranslator.com/Default.aspx) enables TTS output for translation result last week. You now can not only know what the translated results would be but also learn how to read the translated results in the language by clicking the speak button on the top of the result window. Currently TTS are enabled on these target languages: English, German, Spanish, French, Italian, Portuguese, and Russian.

Bing Translator also comes with an API (http://msdn.microsoft.com/en-us/library/ff512420.aspx). Developer can call the API to speak a piece of text and get the TTS wave output stream easily. For a code sample, check out this blog.

For more details about Bing Translator, check their blog.

If you have any feedback on the voices with the translator, please comment.

Programming with latest MS TTS engines

2009-05-03T16:32:00Z

We released 12 TTS voices in 12 languages with UCMA SDK (also called as Microsoft Office Communications Server 2007 R2 SDK) . The Microsoft speech engines bundled with the API support twelve different languages, including English (North America, United Kingdom), French (France, Canada), German, American Spanish, Brazilian Portuguese, Italian, Japanese, Korean and Chinese (Simplified and Traditional). These TTS voices are server version. By saying "server", it means these voice has been optimized to be used in server telephony scenerios with good performance.

This post will describe how to program with the TTS engines inside the SDK.

1. Download the the SDK and language pack

http://www.microsoft.com/downloads/details.aspx?FamilyID=30e14c5a-a42c-4d4e-9513-c4b0b8d21086&displaylang=en

http://www.microsoft.com/downloads/details.aspx?FamilyID=768efa33-6606-4b2b-809a-6c69274621d3&displaylang=en

2. Read MSDN documentation about the Microsoft.Speech namespace

http://msdn.microsoft.com/en-us/library/microsoft.speech.synthesis(office.13).aspx

3. Use C# to program a basic TTS application with these 12 TTS voices.

Add referrence to Microsoft.Speech, then you can program TTS just like I introduced in previous post.

MSDN Speech Launched

2009-05-03T16:25:00Z

MSDN Speech portal is online now. http://msdn.microsoft.com/en-us/speech/default.aspx

From here you can get access to latest Microsoft Speech technologies including Speech API, Speech recognition and Text to speech etc.

Come on and try the newest Speech product.

I will post some tutorials about using the latest TTS engines released with UCMA SDK. Stay tuned.

Tellme deploys Zira voice with MS TTS technology

2009-04-30T13:32:00Z

Tellme today announces that it has deployed a female voice called Zira on their platform, the largest VXML voice platform. The Zira voice is driven by recent MS TTS technology. Tellme customers now can access Zira voice in their applications.

If you are interested in how it sounds, just listen to the clip here.

Text to speech feedback

2009-04-12T14:24:00Z

Dear reader,

If you have any suggestion or feedback when using MS TTS on Vista or Windows 7 or TTS APIs, this blog is a place you can post your comments about:

1. issues you find when using TTS in current products

2. suggestions you have in mind that you want in next version of the product

I promise I will read and response to each comments. :-))

Hello world to TTS (2)

2009-01-27T07:12:00Z

In this post, I will show you how to use TTS with the managed Speech API (System.Speech.dll) in C#. This API is introduced since Windows Vista.

It supports the tranditional SAPI engines and SSML

The steps are as below.

1. create a console C# application.

2. add referrence to System.Speech.dll.

3. input the sample code below:

using System;
using System.Collections.Generic;
using System.Collections.ObjectModel;
using System.Linq;
using System.Text;
using System.Speech.Synthesis;

namespace SystemSpeech
{
    class Program
    {
        static void Speak(int voiceIdx)
        {
            SpeechSynthesizer syn = new SpeechSynthesizer();
            ReadOnlyCollection<InstalledVoice> voices = syn.GetInstalledVoices();
            if (voiceIdx < 0 || voiceIdx >= voices.Count)
            {
                Console.WriteLine("voice index out of range from [0-{0}]", voices.Count);
                return;
            }

            syn.SelectVoice(voices[voiceIdx].VoiceInfo.Name);
            syn.Speak("Hello world");
        }

        static void ListVoice()
        {
            SpeechSynthesizer syn = new SpeechSynthesizer();
            ReadOnlyCollection<InstalledVoice> voices = syn.GetInstalledVoices();
            foreach (InstalledVoice voice in voices)
            {
                Console.WriteLine(voice.VoiceInfo.Description);
            }
        }

        static void Main(string[] args)
        {
            // list voice installed on the system
            ListVoice();

            // speak with the second voices. If you have only one voice, change it to Speak(0)
            Speak(1);
        }
    }
}

Hello world to TTS (1)

2009-01-25T12:11:00Z

In this post, I will show you how easily you can make your first TTS application with C#. This example will use the COM interop to call SAPI functionality because SAPI has implemented a number of automation objects that can be directly used in C# or any other .Net languages.

Suppose you have Windows Vista (or Windows 7 Beta ) and Vistual Studio installed. Here is a step by step guide:

1. Create the solution by clicking the menu: File -> New -> Project, select the "Visual C#" -> "Windows" -> "Console Application",

choose a project name "SAPIInterop" . A basic C# project will be created automatically

2. Add the referrence to SAPI.dll by clicking "Project" -> "Add Referrence" -> "COM". In the list, find and select the "Microsoft Speech Object Library". A "SpeechLib" referrence will appear in the project referrences list.

3. Now we can add some piece of code to make the first "hello world" TTS application.

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using SpeechLib;
using System.Diagnostics;

namespace SAPIInterop
{
    class Program
    {
        static void Speak(int voiceIdx)
        {
            SpVoice spVoice = new SpVoice();
            ISpeechObjectTokens tokens = spVoice.GetVoices(null, null);
            if (voiceIdx < 0 || voiceIdx >= tokens.Count)
            {
                Console.WriteLine("voice index out of range from [0-{0}]", tokens.Count);
                return;
            }

            spVoice.Voice = tokens.Item(voiceIdx);
            spVoice.Rate = 0;
            spVoice.Speak("Hello world", SpeechVoiceSpeakFlags.SVSFDefault);
        }

        static void ListVoice()
        {
            SpObjectTokenCategory tokenCat = new SpObjectTokenCategory();
            tokenCat.SetId(SpeechLib.SpeechStringConstants.SpeechCategoryVoices, false);
            ISpeechObjectTokens tokens = tokenCat.EnumerateTokens(null, null);
            foreach (SpObjectToken item in tokens)
            {
                Console.WriteLine(item.GetDescription(0));
            }
        }

        static void Main(string[] args)
        {
            // list voices installed on the system
            ListVoice();
            // speak with the second voices. If you have only one voice, change it to Speak(0)
            Speak(1);
        }
    }
}

Try Windows 7

2009-01-25T10:03:00Z

Windows 7 beta is ready to test drive now.

http://www.microsoft.com/windows/windows-7/beta-download.aspx

I have installed it on two of my computers. So far, I feel good about it. And I am trying to install it on my home computer.

If you want to use TTS, the TTS voice inside the beta version is the "Microsoft Anna". You can have a try with it as well.

Time to learn silverlight

2009-01-25T06:52:00Z

Sliverlight is becoming wider adoption than years before. I have been viewing showcases on the sliverlight website today.

http://silverlight.net/Showcase/

In general, the UI looks rich and flexible by using this technology.

Some sites I have tried:

Renlifang: http://renlifang.msra.cn/news.aspx from MSRA

Games: http://itouchfuture.cn/lab/powerarrow/

And It also opens new opportunity for ISV to create Sliverlight UI control.

So do you use Sliverlight now?

If you are are developer, when designing a new web UI application, using Sliverlight is worth consideration.

TTS introduction

2009-01-14T17:34:00Z

TTS means "Text to speech" which is the technology to let computer speak to you.

In short, you can input the text to computer, then a computer algorithm (so called TTS engine) will analyze the text, determine the pronunciation, synthesize the sound with some mathematical models. The output of a TTS engine is usually kind of sound data like PCM wave data etc.

On Windows platform, TTS functionality can be called via Speech API. There are two Speech APIs as below:

- System.Speech: managed

- SAPI: unmanaged

To get Speech API SDK, here is the good source to start

http://blogs.msdn.com/chuckop/archive/2008/02/26/speech-content-in-the-windows-sdk.aspx

I will discuss some aspects on how to use TTS via Speech API in future posts.

The user scenerios of TTS are mainly related to human machine interaction. Together with SR (Speech Recognition), Speech technology provides the foundmentals for next generation of natual user interface.

Dear readers, do you use TTS in your application or scenerio? How do you think about it? I'd like to hear comments or story from you.

Dragon book

2009-01-10T18:54:00Z

I should have read this book when I prepared my entrance test for master degree. It is a really good book about compiler design.

But it was about 8 years agao, I have forgotten the details in the book. :(

En, I plan to add it into my computer book collection.

China-pub: http://www.china-pub.com/STATIC07/0811/jsj_longshu_081124.asp

Amazon: http://www.amazon.com/Principles-Compiler-Addison-Wesley-information-processing/dp/0201000229

hi, there

2009-01-10T18:26:00Z

This is my first post