Voice Message Systems for Tactical Applications (Canned Speech Approach)

One-way noninteractive voice messages are often used in tactical environments. Examples are surveillance and reconnaissance reports, tactical coordination messages, warnings, and reminders. These messages can be transmitted efficiently in terms of words, phrases, or sentences. We use a speech recognizer to convert speech into text. The resultant data rate is below 100 bs. At the receiver, speech is regenerated by concatenating the stored speech waveforms canned speech corresponding to the received indices. Intelligibility of the resultant speech is high because output speech stems from actual stored speech rather than synthetic speech. If tactical messages are generated by concatenating only words, however, we need to incorporate the sentence-level prosody in the generated speech. For tactical messages, however, prosodic rules are relatively simple because tactical messages are customarily spoken without significant pitch and rhythmic inflections. If tactical messages are generated by concatenating phrases and sentences, resultant speech will sound natural without further speech modification. Another advantage of the canned speech approach is that the spoken language can be translated into any one of the preselected languages at the receiver. The voice message system will play a significant role in future DoD secure voice terminals. Tactical voice message, Extreme-low-data-rate voice messages, Formatted voice messages,