Thank you for your interest in LinguaProfessio’s™ Human Voice Engine™. We hope that this web site will answer many of your questions about what our products can do for you.

We offer three different versions of the Human Voice Engine™; each one fined tuned for a specific task.

What is LinguaProfessio’s™ Human Voice Engine™?

Our Human Voice Engine™ (HVE™) products are designed from the ground up to be the most intelligible text-to-speech engine available.

Most speech synthesizers contain a limited set of phonetic recordings. In an effort to approximate speech, such systems must string together these tiny verbal fragments. Sadly, the result is often difficult to understand.

On the other hand, our Human Voice Engine™ is built upon a foundation of nearly 100,000 studio-quality recordings. The lion’s share of this massive library is packed with whole words and complete phrases, not just word parts.

So while most synthetic speech systems must join many small recordings to reproduce even simple sentences, the LinguaProfessio™ advantage is its ability to reproduce speech using larger, easier to understand recordings.

Also, our patented and patent-pending technology doesn’t play operating system favorites. Our software is at home on any Microsoft OS back to Windows '98 SE, UNIX, Linux, Solaris and Macintosh OS X. This means that we can integrate into your facility with a minimum of fuss.

LinguaProfessio’s™ HVE™ software makes few demands on your I.T. staff. Plain text goes in and almost instantly industry standard digital audio comes out. It’s that simple.


The Early Synthesis Model

For decades, universities and large corporations have endeavored to electronically reproduce human speech. In the early days it was thought that the only way for a machine to generate the unlimited vocabulary required for everyday speech was to use some form of synthesis.

Synthesized speech systems re-construct from tiny “modeled data” sound fragments. When building spoken audio, a synthesizer will string together these tiny snippets of “modeled” voice information. Next, an electronic vocal chord simulator forces the sound through a synthetic mouth, tongue, nostrils, and lips.

After years of effort this “synthesis by rule” method has actually started to sound reasonably good. And of course modern synthesizers certainly are a major improvement over the early days when the robotic voices were very hard to understand and often mispronounced a great many words.

The LinguaProfessio™ Model

As good as some synthesizers have become, most people continue to believe that there is nothing better than a real human voice!

We at LinguaProfessio™ agree.

Our model is entirely different. Rather than using a comparatively small selection of sound fragments that are re-combined to create synthetic speech, our Human Voice Engine™ draws upon a huge library of complete word and phrase recordings.

This approach isn’t exactly new; it has been tried before. It never took long, however, for developers to realize the sheer magnitude of the task in front of them. The investment in time and money that must be made to record and edit a voice library deep enough to be even minimally functional always seemed to stall these past efforts while they were still in their infancy. It is for this reason that virtually all of these preceding attempts have failed. In fact, the published consensus in many textbooks is that it simply can't be done. It is suggested that budgetary considerations alone will prevent the creation of a library that has sufficient scope to avoid sounding unnatural and unclear. But the primary reason this method will fail, it was thought, is because the number of human performance variables is overwhelming!

A lot of variables? You bet! Overwhelming? We don’t think so. From the beginning we knew what we were facing and how it could best be achieved. Our R&D staff is made up of people with years of experience in the fields of sound recording and editing, computer generated speech and linguistics. We developed patent-pending technologies to streamline and control the voice recording and editing process. And we developed a patent-pending concatenation model that is unlike anything that has come before.

The Depth of the Human Voice Engine™

The LinguaProfessio™ Human Voice Engine™ can speak, at will, the vast majority of commonly used English words and thousands upon thousands of the less commonly used English words. Additionally, we’ve incorporated thousands of everyday phrases into our library.

Of course this isn’t the end of it. When a text-to-speech request includes a word that isn’t a part of our library, our software uses sophisticated routines to construct the target word using small recorded fragments. This part of the process works much like most other speech synthesizers. Our advantages, however, are that even these small voice snippets were recorded by the same professional announcer who recorded everything else in the library. Also, we resort to this method only when absolutely necessary, while virtually every other product on the market uses this approach as the default behavior.

Expanded Possibilities

The ways in which LinguaProfessio’s™ Human Voice Engine™ can be used are almost limitless. It is even possible to embed HVE™ directly into hardware for inclusion in almost any kind of device: in car navigation systems, eMail readers, automated kiosks, virtual agents, just about anything your imagination can conceive.


LinguaProfessio’s™ Products Will Help You Communicate With Your Customers

It doesn’t matter how smooth a speech synthesizer sounds if you can’t understand what it is saying. In making the Human Voice Engine™, our primary goal was to make speech that is easy to understand. This is a target that we feel we have hit. Now we hope that you’ll allow our efforts to work for you.