News Archive (1999-2012) | 2013-current at LinuxGizmos | Current Tech News Portal |    About   

Web-based toolset creates custom synthesized voices

Feb 21, 2007 — by LinuxDevices Staff — from the LinuxDevices Archive — 1 views

Speech synthesis specialist Cepstral LLC has announced a toolset aimed at helping users synthesize their own voices. The Web 2.0-based VoiceForge tools are said to enable users to capture or “bank” their own voices, after which the company's lightweight, embeddable text-to-speech software can synthesize their voice saying anything at all.

Cepestral hopes VoiceForge will enable users to go beyond the safe but boring voice profiles that come standard with text-to-speech applications, creating unique, interesting, and personal synthesized voices for specific vertical applications in entertainment, telephony, navigation, and education. Synthesizing ethnic or accented voices appears to be possible with the technology, for example.

The Web-based VoiceForge toolsuite is initially used to record “audio prompts” into a custom text-to-speech (TTS) voice database. Once created, the database can be plugged into Cepstral's “core engine” text-to-speech application, a lightweight software module that will be licensed to application and device developers.

Cepstral says the core engine runs on a wide range of platforms, from cellphones to “large distributed systems.” It supports Linux, Windows, and Windows CE, among other operating systems.

According to Cepstral, the web-based VoiceForge tools can create custom voices rapidly and inexpensively. Additionally, users retain “all” IP (intellectual property) rights to their voice creations. Suggested applications include voice prompts, reading documents and Web pages, VoIP announcements, SMS-to-voice, text-to-podcast, and custom ring tones.

Cepstral CEO Craig Campbell stated, “As an industry, we make voices that are safe, but not necessarily exciting. With VoiceForge, clients can now create unique high-quality TTS voices that keep pace with consumer and business demand for branded, celebrity, ethnic, and even cartoon personalities. To cite but one example of the need to improve voice diversity, there are currently no African American TTS voices available.”

Little information about the just-announced VoiceForge tools appears on the company's website. However, the site does offer an interesting online demonstration of the company's TTS application, as well as trial downloads of single-voice TTS applications for Linux, Windows, and Mac.

Availability

Specifics of VoiceForge pricing and availability were not disclosed. It appears that companies can license the tools for use in-house, or can obtain contract services through designated third-party suppliers.

In the 1990s, Cepstral's founders released Festival, an open-source TTS engine, according to the company.


 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.



Comments are closed.