What is speech technology?
There are three core speech technologies:
- 1) TTS - Text-to-speech converts written text into speech. TTS is also known as synthetic speech or speech synthesis.
- 2) VR - Voice recognition is the ability for a device to recognize speech and have specific actions taken from voice input only. This is also known as ASR or Automatic Speech Recognition.
- 3) Speaker Verification is the ability to verify the identity of a specific user based on the patterns of their speech.
When would you need text-to-speech?
High-quality speech synthesis is needed in any situation where natural-sounding and clear verbal communication is required, but where the delivery of live or pre-recorded human voices is prohibitively expensive, technologically complex, or insufficiently flexible.
Where does Cepstral fit in?
Cepstral is a world leader in providing speech synthesis (TTS) technology and services. The focus of our business is designing, building, and customizing high-quality, natural-sounding synthetic voices.
What makes Cepstral a market leader in synthetic speech?
Cepstral was founded by Dr. Alan Black and Kevin Lenzo, two world-class speech scientists, who were working together at Carnegie Mellon University. Their pioneering work and decades of research in speech synthesis has enabled Cepstral to build world-class products and remain on the cutting edge of voice quality and speech engine performance.
This quality was recently recognized as Cepstral was awarded two of the eight awards at the AVIOS/Speech Developer's Conference.
While Cepstral continues to push the envelope of speech synthesis technology, we also are experienced at bringing commercially viable solutions to the market. These solutions add clear value to our clients' products and enrich the end-user experience.
Cepstral provides voices with an unprecedented combination of quality and size. Our small memory footprint capability makes Cepstral ideal for hand-held or mobile applications where space and processing power is limited.
What if I am using pre-recorded speech?
Synthetic speech offers the ultimate flexibility to say anything without having to spend the time or expense of bringing voice talent back into the studio. Cepstral voices offer the quality and intelligibility of recorded speech from professional voice talent with the flexibility of text-to-speech, so that the quality can be maintained even when the content is constantly changing.
What if I'm already using synthetic speech in my application?
Cepstral's unit selection synthesis represents the third-generation of technology and produces a higher quality, more natural sounding voice than earlier generations of technology (such as formant or diphone synthesis.) Many of Cepstral's clients have licensed our software simply to improve the quality, naturalness, and intelligibility of the voice.
Cepstral can design a voice specifically for your organization to assist in branding and to enrich the user experience. We can provide a whole variety of different types of voices, and we can design a voice that can be used across the whole company (website, phone system, demos, etc.) These voices can be based on a particular voice talent, or can be tailored from one of our existing voices.
What types of voices are available for our clients?
There are three types of voices available: 1) existing voices, 2) tuned voices, and 3) brand voices.
There are a number of current voices available, including male, female, and a child to fit any variety of applications. To listen to samples of these voices, please follow this link.
Cepstral can significantly improve the quality of our existing voices by customizing or tuning them for a particular domain, such as navigation, radiology, or weather. By tuning voices to a specific context, a very high level of comprehensibility is achieved. To hear an example of a high-quality tuned voice, please follow
this link.
Cepstral has invented techniques for the design of personality and brand voices based on a specific voice talent. These voices would allow a company to extend their brand image to automated speech applications throughout the organization. While building a new synthetic voice can be a major undertaking, Cepstral can start with recordings of the voice talent and supplement them with studio time to reduce the costs and effort involved.
How can I hear a demo of one of your voices?
Cepstral has a number of demos available. Please follow this link.
What if I want voices in other languages?
Cepstral has a full complement of North American languages (US English, French Canadian and Americas' Spanish) as well as UK English, German, and Italian. With the exception of Italian, each language has at least one male and one female voice. (Currently, for Italian, we only offer a female voice.) We have also developed limited-domain voices in Egyptian Arabic, Iraqi Arabic, Thai, and Pashto. We are currently developing several additional European and Asian languages. Please contact us if you have questions about the timing or availability of a specific language.
What does Cepstral's relationship with Carnegie Mellon University do for our clients?
Cepstral's close ties with CMU provide a direct connection with some of the top speech synthesis talent in the world. Dr. Alan Black and Kevin Lenzo are recognized leaders in the field and collaborate extensively with colleagues at CMU to pioneer the kind of cutting-edge research that a university environment nurtures. Carnegie Mellon students, faculty, and visiting researchers have enriched the Cepstral family as interns, advisors, collaborators, and employees. CMU is also the hub for a network of small technology companies that Cepstral works with to provide voice and language solutions that extend beyond speech synthesis.
In addition to being focused on bringing the best voice solutions to the market today, Cepstral is committed to continuing its innovation and leadership in the field of speech synthesis.
How much does Cepstral services and software cost?
Cepstral has a variety of services and technologies. In order to provide the optimal pricing, we tailor a cost plan for each individual client. Please contact us for more information.
How can I learn more?
Cepstral offers a two-day Speech Technology Workshop that allows our clients to learn more about speech technology and determine if and how it could improve their business. Most importantly, these workshops are tailored to your specific business and products so the information is extremely relevant.
Once Cepstral has acquired an understanding of our client's products and industry, we will send one of our world-renowned speech scientists to your facility to provide an overview of our findings and recommendations. All of our workshop presenters have an educational background and are equally adept at discussing both business and technical issues.
For more information, feel free to contact us.
I have a question that isn't covered here.
For all other inquiries, please use our Contact Request Form. Please provide as much information as possible. Thank you!
