The Laryngo Voice solution is developed on the basis of research by the University of West Bohemia in Pilsen and its goal is to help patients suffering from cancer of the larynx. In many cases people with this illness must undergo a complete laryngectomy, an operation in which the entire larynx, including the vocal cords, is removed, leading to the loss of the ability to speak. With the use of Laryngo Voice, however, their voice can be preserved, and even continued to be used in communication, for the creation of a synthetic voice on the basis of written text. The voice is generated on the basis of stored recordings of phonetically rich sentences from the period before the operation. The voice synthesizer is installed on a portable device (i.e. a mobile phone, tablet, notebook) and the patient can adjust the artificially created voice. This modern solution ensures the artificially-generated voice’s great similarity to the patient’s original voice.
The first phase of the synthesis process is a sound recording of the patient’s voice before the operation, which is performed in a specialised recording studio. Home recordings are currently being tested, however. The selected texts that are recorded are carefully chosen to effectively capture a wide range of the given patient’s vocal expression.
After recording the patient’s voice, a control of the spoken text and a phonetic-prosodic analysis are carried out. The recordings are processed, parameterised and segmented. The result is a voice module compatible with these voice synthesis methods:
- statistical-parametric voice synthesis (the speech is generated on the basis of statistical models, with the use of Markov models)
- unit selection (speech originates by the linking of carefully selected short segments of actual speech)
Laryngo Voice is constantly being perfected. We are focusing our current efforts on the full automation of the system, where the entire process, from the recording to the creation of the subsequent synthetic voice, would not require human assistance. The patient would create the program himself, reading the entered text in his synthetic voice, and then he would install it on the selected device. If Laryngo Voice reaches this phase, it would be absolutely unique, in comparison to the competition, for its compete automation and for the speed in which it would be possible to create a quality synthetic voice system and to immediately install it on commonly used portable devices.
Other known projects dealing with the creation of personal synthetic speech systems are predominantly focused on statistical-parametric speech synthesis. An average statistical model created using a large number of different voices is usually adapted for a concrete user on the basis of his voice recordings. The similarity between the resulting output and the voice of the person in question is usually not satisfactory, however.
The goal for the future is also to adapt the current parameters of Laryngo Voice for voice modulators and a “speech prosthesis” that will be part of the patient’s body.
Laryngo Voice and its testing
From the very beginning of the project the University of West Bohemia has been cooperating with the Department of Otorhinolaryngology and Head and Neck Surgery in Motol University Hospital and First Faculty of Medicine, Charles University. Laryngo Voice has been clinically tested. Patients that have been diagnosed with cancer of the larynx are offered the possibility of recording their vocal expression. A personal synthetic speech system has already been created for ten patients.
The response from patients that have tested this solution shows that the system is able to significantly improve the quality of their communication in their personal and professional lives and is very positively received by their loved ones. Laryngo Voice does not only have to be used by patients that have undergone a complete laryngectomy. It can be used by anyone who would like to preserve the voice of any person.