We frequently are asked if it is possible to use TextAloud or other Text To Speech products to allow web pages to talk to your visitors.  Whether to help those who are sight-impaired or to simply add an easier method to hear information available on your site, using Text To Speech is often a valid option for web use. 

There are several different options for making web pages talk, each has advantages and drawbacks, and the correct choice depends on what you want to accomplish.  Two key questions you must answer about your plans are:

Is the text to be spoken static or dynamic?
Static text would mean that the text spoken for specific pages or topics is known ahead of time and doesn't change daily.  Dynamic content may mean that the text changes in real-time, such as when the words spoken depend on user input or information looked up in a database.

Should Speech be generated on the server side or on the visitors computer?
If speech is generated on your web server, you have control over the voice used.  Audio files, typically MP3 or WMA are generated and returned to the user, played typically in their default media player, or an activeX version of the media player imbedded in the web page.  If speech is generated on the client side, the user installs a Text To Speech component, and the user has control over the voice used, the speech is actually generated on the visitors computer.

The examples below discuss common scenarios and their solutions:

Speech from static text using pre-created audio files.
The most common method of adding speech to web pages is creating MP3 or WMA audio files ahead of time.  If you know ahead of time the speech you'd like the user to hear, you can use TextAloud to create WMA or MP3 files.  These files are uploaded to your web server.  You can then:

• Provide users with a direct link to the audio file. The user clicks the link to listen, and their default media player (often Windows Media Player, RealPLayer, or WinAmp) will open and begin to play the audio.

• Use special HTML or scripts to play the audio automatically when a web page is openned.

• Use HTML or scripts to show a java or activeX player within the web page to play the audio.

The important thing to note with this method is that while it is the simplest, the text must be static, as you have to create the audio file manually using TextAloud (and possibly TextAloud's batch file converter) then upload those audio files to your server.

Real-Time generation of audio files on the server.
If your text is more dynamic, you can still generate audio files on the server.  With TextAloud's API installed on your windows server, your web code would access the TextAloud API (supports most programming languages including server-side VBScript), passing text to TextAloud.  TextAloud would generate MP3 or WMA file on the fly, and once the file is created, your web code can return that audio file to the user similar to methods used in the section above.  You can get more info on the TextAloud API at
http://www.nextup.com/api.html

While this method works reasonably well, depending on voice used, time needed to generate each audio file varies.  Only one file creation at a time can take place, so if you have very heavy traffic, users may experience some delay waiting for the audio file creation before they begin to hear speech.

NOTE on using Premium Voices: AT&T Natural Voices cannot be used for web distribution.  The AT&T license agreement prohibits this.  All premium voices have a restriction prohibiting this.  We do sell NeoSpeech voices at
http://www.nextup.com/neospeech.html
Advanced licenses are available for these voices that allow web usage.  Licenses start at $1500 depending on exact details of implementation.  You may use the older Microsoft voices without restriction.

Generating speech on the client computer.
Unlike the above scenarios, there are situations where web pages can cause speech to be generated on the visitors computer directly.  You may have seen Microsoft Agent
http://www.microsoft.com/msagent/default.asp

Using special HTML, if your visitor has installed Microsoft Agent, you can cause a little cartoon character to pop up on their screen and begin speaking the text that you specify.  The downside to this is often the animated character is a terrible distraction and the voices with Agent are not very good.

It is also possible to have the visitor install TextAloud, and your web code can start TextAloud speaking automatically.  This works particularly well in closed group settings, such as cases where all visitors to the site are in a known group.  For example, some e-learning or training settings, or member only sites it may be feasible to have all visitors install TextAloud to enable speech.  Once TextAloud is installed on the client computer, then your web pages can use VBScript to access the TextAloud API
http://www.nextup.com/api.html

Your web pages will then have the ability to pass text to TextAloud, start and stop speech, basically control most functions within TextAloud.  One added benefit of this method is the user has more control over voices used, and can purchase and install premium voices to improve the voice quality.

These examples do not cover all possibilities, so feel free to email us at support@nextup.com to discuss.