DECtalk is a speech synthesizer, originally released as a hardware module in 1984 by Digital Equipment Corporation (DEC).
The DECtalk technology is based on the work of Dennis Klatt, who was a pioneer in speech synthesis throughout the 1960s and 1970s. He developed the phoneme-to-speech part of a system known as MITalk, a project developed at MIT (Massachusetts Institute of Technology), using recordings of his own voice to build a functioning voice model, later to be known as Perfect Paul. Perfect Paul would end up being a model emulating how Klatt talked in his younger days. In 1982, Klatt developed his own speech system nown as Klattalk, which added other voices to give a much wider selection, most of which were modelled after his family. This included Beautiful Betty, modelled after Klatt's wife, and a child voice, Kit the Kid, modelled after his daughter. In 1983, DEC bought an exclusive license to Klattalk to use it in their own products, the first of which was released in 1984 under the name "DECtalk".
When DEC licensed the system and released it as DECtalk, they continued to collaborate with Klatt on the development, until his dead in 1988, whereafter DEC continued development until the late 90s, both as hardware, and later software products. Hardware products included, but were not limited to:
Listen to a sample of DECtalk DTC-01 running DECtalk version 1.8:
Listen to a sample of DECtalk DTC-01 running DECtalk version 2.0:
Listen to a sample of DECtalk PC running DECtalk version 4.2CD:
Listen to a sample of DECtalk Express running DECtalk version 4.2CD:
In 1994, DEC launched DECtalk as a software product running on Windows NT with their own Alpha architecture, and later ported to run on Intel as well. DECtalk software was available as a SDK for developers, though it could be demonstrated for free through a demo program, which allowed a user to type in text and have it read aloud by DECtalk. The voice sounded mostly similar to the DECtalk hardware products, though it had a few differences in the sound and how it behaved, some of which related to the singing.
Listen to a sample of DECtalk 4.2 Software:
The sound of DECtalk software remained mostly the same throughout the nineties, even through DECtalk's acquisition by Compaq in 1998, who released DECtalk version 4.51, and later by SMART Modular Technologies in 1999, who released DECtalk version 4.60. DECtalk 4.60 is probably best remembered for being the version used in DECtalk Access32, a product sold specifically as a speech synthesizer option for the JAWS for Windows screen reader. A slightly modified version of DECtalk 4.60 was also later included in the Window-Eyes screen reader.Listen to a sample of DECtalk 4.60:
In 2000, DECtalk was sold to Force Computers, Inc., who radically altered the sound of DECtalk. Their first and only release, version 4.61, had a much thinner sound than previews versions, and altered the pitch and characteristics of most of the voices. This version was released as software, and was also made available both for the DECtalk PC2 and DECtalk Express.
Listen to a sample of DECtalk 4.61:
In december 2001, DECtalk was sold again, this time to Fonix Corporation, who are the last known owner of DECtalk, under their branch Speech FX, Inc. Fonix made an effort to revert many of the changes made in DECtalk 4.61, to restore a more classic DECtalk sound. However the resulting version, 4.64, ended up being even more removed from the older DECtalk versions from DEC. It had restored a more present sound, but the voice intonation, characteristics and pronunciation was altered to a point of almost not sounding like DECtalk at all. This version of DECtalk is probably best nown for being included in the Moonbase Alpha game's chat client.
Listen to a sample of DECtalk 4.64:
Fonix also released DECtalk 5.0, which was available in many different sounding editions. This was meant to be Fonix's own spin on the DECtalk sound, but is one of the lowest regarded DECtalk version today. This also later released FonixTalk 6.0 and 6.1, which was a continuation of the DECtalk 5.0 sound.Listen to a sample of FonixTalk 6.0:
In 2004, Fonix licensed DECtalk to Access Solutions, to be used in their DECtalk USB unit. The DECtalk USB was a hardware unit, regarded as the successor to the DECtalk Express. It had similar functions, but could connect to a computer both through USB and RS232, and had support for multiple languages. It used a modified version of DECtalk 5.0, altered to sound more like DECtalk 4.2CD from the DECtalk Express.
Listen to a sample of the DECtalk USB:
Fonix went out of business in 2014 and Speech FX in 2020. DECtalk is therefore not being worked on officially anymore, though a small group of private developers are maintaining a project containing working DECtalk builds for modern architecture, known as DECtalk 4.99.
Today, DECtalk is probably best known for being incorrectly attributed to being the voice used by Stephen Hawking. It has also become somewhat famous for it's ability to sing, which many people have used in text-to-speech creations such as songs, skits and anything inbetween. The most used DECtalk version for these projects is version 4.3, which is generally agreed upon to be the best software version of DECtalk, sounding the closest to the hardware products.
Listen to a sample of DECtalk 4.3:
The focus of this archive is speech synthesizer creations, especially those that use DECtalk. The archive is available either as a Resilio Sync folder, or as a website at dectalk.nu, either of which will allow anyone to download the archive. However, if new files are added to the Resilio Sync folder, they will not be shared automatically with everybody else. Read on to find out how you can contribute to the archive.
The primary focus of the software part of the archive is to include different versions of the DECtalk speech synthesizer, however it is also open to other, primarily older, software speech synthesizers, as long as they meet the following criteria:
If you have a speech synthesizer you think belongs in the archive, or if you notice a speech synthesizer in the archive not meeting these criteria, please get in touch.
If you would like to download the entire archive, I'd recommend connecting to the Resilio Sync folder, to help keep the archive alive by automatically distributing it to other people through Resilio Sync. However, if you do not want to / are unable to use Resilio Sync for one reason or another, you can download the archive as a ZIP file. Please note however, that the ZIP file might not be updated as frequently as the archive itself. You can see the latest updates in the archive in the changelog.
Last ZIP update: 08/03/2025
WARNING!
Some files in this archive contain content that might be deemed unsuitable by various people. Since this content is relevant to the topic of speech synthesizer creations, it has been added to the archive. You have been warned. If you find this offensive, you have my appologies, but the content will not be removed unless requested by it's creator. Listen and/or read at your own risk.
If you would like to contribute to the archive, contact Aksel Christoffersen by mail, or through the AudioGames.net Forum. If it's your first time contributing your own creation(s) to the archive, please specify which name you would like your content to be listed under. It could be your real name, or a selected username. If no name is specified, the content will be saved in the "Miscellaneous" folder.
If your content is in .wav format, I would recommend you convert it to .flac format instead. FLAC (Free Lossless Audio Codec) is a so called "lossless" audio format, which compresses an audio file to a smaller size, about 60 percent of the original size, without losing any of the original audio. Since FLAC also preserves the timestamp of the original file, no data is lost at all, and virtually all operating systems can play FLAC at this time. If you would like to keep your content in .wav format, or if you do not know how to convert to .flac, please let me know, and I'll be happy to do it for you. All existing archive content in .wav format will remain unchanged.
If the size of your content is small, under 10 MB, you can send it as attachments to your e-mail. If not, you might want to use a file sharing site. My reccomendation would be Dropbox, where you can copy a link to a file or folder in your Dropbox folder, and send it in an e-mail for easy sharing. If you don't already have a Dropbox account, you can sign up for a free one here and get 2.5 GB of space for free.
The purpose of this archive is to get the content into the hands of anyone who wants it. As such, you are always permitted to distribute anything from this archive to anyone who might want it, as long as you give credit to whomever created the content you're distributing, and do not try to pass it off as your own work. If an individual intend to keep the content around for a longer period of time, I'd strongly adwise to connect to the Resilio Sync folder, to help keep the archive alive by automatically distributing it to other people through Resilio Sync. Alternatively, you can always give the url to the website, dectalk.nu, if the individual in question is simply looking to get a small number of files.
Note: this guide describes installation of Resilio Sync on Windows, with specific instructions for screen reader users, and has not been tested with other operating systems.
To my knowledge, the songs created with synthesizers are not violations of copyright. Certainly, the skits created with speech synthesizers do not violate anything. However, if something does violate copyright or gives information relevant to breaking copyright law, I reserve the right to remove it at any time. This would, for example, include links to pirated software or content, so please don't post anything like that.
In this archive you will find at least one type of every known official version of the DECtalk speech synthesizer, that being either as a speak window, software for DECtalk hardware units, or some other type of software. However, this archive will not contain any unofficial or modified version of DECtalk, due in part to the, at the moment, unclear state regarding the ownership of DECtalk. This rule does however have the following exceptions:
If you're looking for ready-to-go modern builds of DECtalk, you can find them here, and you can find everything about the current DECtalk development here. Here is a list of every version of DECtalk currently in the archive, with links to the apropriate file(s) for each version.
Note: Some of these versions are also available as NVDA Add-ons. Check the NVDA Add-ons folder for more information.
Here you can find some other resources relating to DECtalk speech synthesis.
I sincerely hope you'll enjoy the content of this archive. If you have any questions, requests, comments or the like, feel free to drop a message.