SpeechActs was a research prototype that integrates third-party speech recognition and synthesis with telephony, natural language processing capabilities, and other tools for creating speech applications. The system was designed by the Speech Applications Group at Sun Microsystems Laboratories for traveling professionals who require access to on-line information while they are away from their computer. SpeechActs' conversational speech interface is easier to use, requires less learning, and offers more direct access to information than keypad input for telephone-based interaction. Without having to train the system, a user can telephone SpeechActs from an airport, a hotel room, or a colleague's office and speak requests naturally without having to memorize commands. For example, SpeechActs will understand phrases such as "I'd like mail please" or "What's on my calendar the day after tomorrow."

SpeechActs provides natural, speech-only interfaces to a suite of integrated applications. Two of these are speech interfaces to Sun's graphical deskset tools, Mail Tool and Calendar Manager. Other applications provide speech access to dynamic data feeds for weather forecasts, stock quotations, currency exchange data, and international time.

SpeechActs Applications (includes recorded demos) This page includes descriptions and recorded demos of applications in the SpeechActs suite, including Mail, Calendar, Stock Quotes, Weather, International Time, and Currency Exchange.

SpeechActs Architecture SpeechActs is a prototype testbed for developing spoken language applications. This page describes the architecture of the SpeechActs Framework.

Speech User Interface Issues (includes recorded demos) A major challenge in creating speech applications is designing a usable interface. This page examines a number of speech interface issues including the use of conversational devices, tailoring feedback, and designing for errors. Recorded demos illustrate many of the issues.

Publications A number of papers were published on the SpeechActs architecture and user interface research. Many of those papers are provided online.

SpeechActs Examples

For those who want a quick taste of what it is like to interact with SpeechActs applications, here are three short examples. Many more recordings can be found by following the links above.

The SpeechActs mail application allows users to call up their workstation and have their electronic mail messages sorted into meaningful groups, and selected messages read aloud in priority order.

Mail example (43 seconds audio), formats: au, wav, real-audio, text

The SpeechActs Application Framework supports the use of speech recognizers and speech synthesizers as plug-in components. A Discourse Manager keeps track of the conversational context and performs tasks such as interpreting the meaning of pronouns, translating relative dates into specific ones, disambiguating user names, and storing common information so that it can be shared by the applications. Listen to the flexible ways users can refer to dates in the Calendar.

Calendar example (27 seconds audio), formats: au, wav, real-audio, text

Also, notice how the Discourse Manager keeps track of the current place as the user switches from the international time application to the currency exchange application.

Context example (30 seconds audio), formats: au, wav, real-audio, text

These examples were created on a Sun SPARCstation using SunXTL telephony software, the Hark speech recognizer from BBN and the TruVoice synthesizer from Centigram (now Lernout & Hauspie's TruVoice).