DECtalk
| Company type | Product |
|---|---|
| Industry | Speech synthesis, Text-to-speech |
| Founded | 1983 |
| Founder | Digital Equipment Corporation |
| Headquarters | United States |
| Products | Speech synthesizer, Text-to-speech system |
DECtalk represents a speech synthesizer and text-to-speech system introduced by Digital Equipment Corporation in 1983, heavily influenced by Dennis Klatt’s research at MIT. His source-filter model, known as KlattTalk or MITalk, formed the foundation of this technology.
Designed mainly for diverse uses ranging from public engagement to assisting individuals with speech impairments, DECtalk converted typed input into artificial speech, supporting communication and public speaking. It interfaced with personal computers through an RS-232C serial connection and was initially sold at approximately $4,000 in 1984.
History
The official announcement of DECtalk occurred in December 1983, with shipments commencing early the following year and increasing significantly by March.
These standalone units interfaced via asynchronous serial ports and included two telephone jacks—one for connecting a phone line, the other for a handset. DECtalk devices could both recognize and generate telephone touch tones, allowing automation of telephonic functions such as call management, email system integration, and phone-based alerts.
Subsequent versions were developed for PCs equipped with ISA bus slots. Software adaptations emerged as well, notably DECtalk Access32, which explored real-time synthesis on standard CPUs. This evolution produced DECtalk software compatible with Digital Unix and Windows NT operating on Alpha and Intel processors. Some iterations exhibited peculiarities, like alveolar stops being pronounced as dental stops or subtle electronic beeps occurring after spoken phrases.
In the early 2000s, ownership of DECtalk’s intellectual property transferred to Force Computers, Inc. Later, in December 2001, Fonix Speech, Inc. (currently SpeechFX, Inc.) acquired the rights and continues to provide DECtalk as a compact text-to-speech solution both in hardware and software formats.
Features
According to The New York Times, the voice quality of DECtalk was described as resembling "a scratchy recording of a person with a lisp," yet generally remained intelligible. The system offered multiple preset voices including Perfect Paul (the default), Beautiful Betty, Huge Harry, Frail Frank, Kit the Kid, Rough Rita, Uppity Ursula, Doctor Dennis, and Whispering Wendy. Users could tailor each voice by modifying parameters such as throat size and crossover frequencies. Additionally, DECtalk supported phonetic spelling enhanced with pitch and duration controls, enabling it to perform singing.
Uses
- During the late 1990s, DECtalk was employed by the US National Weather Service in their initial "Console Replacement System" (CRS) for NOAA Weather Radio. By 2003, it was predominantly supplanted by Speechify from SpeechWorks, although the Perfect Paul voice continued to be used for station identifications until the CRS was replaced by the "Broadcast Message Handler" (BMH) in 2016.
- The Huge Harry voice remains prevalent for ATIS broadcasts at numerous airports, delivering weather updates and operational information.
- One of its early implementations was Med-Fax’s emergency medical information platform (1986), which utilized DECtalk on IBM systems to transmit patient details to hospitals—marking one of the pioneering cross-platform integrations between DEC and IBM environments.
- DECtalk has played a vital role in speech-generating devices for individuals without natural speech capabilities. Most famously, physicist Stephen Hawking relied on a version of DECtalk following his loss of voice due to ALS and tracheotomy procedures. Hawking favored the distinctive voice provided by the "Calltext 5010" device manufactured by SpeechPlus, Inc. in 1988. This synthesizer served as his primary communication tool until his passing in 2018 and was later emulated on a Raspberry Pi by developers from Cambridge and Palo Alto.
In popular culture
- The persona "Dee Klatt" from Chipspeech was inspired by data derived from this technology.
- In the sixth episode of Tales from the Darkside, the voice of Kevin "Mookie" Anderson was produced using a DECtalk system.
- The video game Moonbase Alpha incorporates a software adaptation of DECtalk for its chat function, which led to viral videos highlighting its distinctive speech and singing features.
- The experimental rock group The Space Negros incorporated DECtalk in their 1986 track titled "Robot."
- Kraftwerk’s 1986 song "Music Non Stop," recorded and edited at Kling Klang Studio in Düsseldorf by Florian Schneider and Karl Bartos, prominently included DECtalk.
- The "Perfect Paul" voice from DECtalk has been adapted for the freeware vocal synthesizer UTAU by the contributor UtaUtaUtau.
External links
- —current versions of the DECtalk text-to-speech software
- OpenVMS Notes: DECtalk Model information.
- NMAH: Smithsonian Speech Synthesis History Project: A Brief History of DECtalk