[LIS 9732] text-to-speech and speech synthesis

A brief introduction to text-to-speech.

Transcript

1. text-to-speech (tts) [and, its good pal, speech synthesis]
2. has this ever happened to you? you’ve got some textual information... but you need to hear it! who has time for reading? there’s got to be a better way!
3. text-to-speech has got you covered! tts synthesis is a process of speech generation using a computer. say goodbye to “microsoft sam”! modern tts technology sounds far less creepy and can relay information in a more lifelike and natural way. relate to computerized systems in a whole new way!
4. how’s it work? basically, tts synthesis is an attempt to mimic the physical process of speech through articulatory models of the human vocal tract or terminal analogue synthesis models of etcetera... (Taylor, 2009) more simply, it is a simulation of the concatenation of natural-sounding diphones. (Taylor, 2009)
5. what’re some applications of tts? tts is actually pretty vital because: offers a voice to those who can’t speak. accessibility of information for the visually impaired. call-centre automation (ugh). allows for hands-free “reading”.
6. typical user? Photos poached shamelessly from Wikipedia.
7. but, mostly: it’s everyone, all the time because: tts is a low- bandwidth way to transmit spoken text. (Nass & Brave, 2005)
8. © Apple assistive technologies for visual or speech impairments (or the very lazy/busy) dialogue systems and conversational agents when paired with speech recognition or chat bots. where does it fit in the nlp landscape?
9. hey, thanks References Nass, C., & Brave, S. (2007). Wired for Speech: How Voice Activates and Advances the Human-Computer Relationship. The MIT Press. Taylor, P. A. (2009). Text-to-speech synthesis. Cambridge, U.K. ; New York: Cambridge University Press.