Emotional Speech Synthesis and Recognition

Emotional speech synthesis

I have been working on algorithms for emotional speech synthesis. The objective was to manipulate the prosody of computer generated speech signals so that a human listener can perceive different kinds of emotions or attitudes, such as happiness, sadness or anger. The algorithms that I developped were inspired by psychoacoustic studies but in no way tryed to reproduce precisely the way humans modulate their prosody to express emotions. Rather, I developed operators for prosodic deformation which are analogous to the deformation of faces in Walt Disney pictures used to express visually the emotions of characters. In brief, there was little science in this project, but a lot of fun!

These algorithms are described in:

Oudeyer P-Y. (2003) The production and recognition of emotions in speech: features and algorithmsInternational Journal in Human-Computer Studies , 59(1-2), pp. 157–183, special issue on Affective Computing. Bibtex

You can hear yourself some examples of results in the following baby-like meaningless babbling sounds:

Sad :

Angry :


Also, I developped means to control continuously the degree of emotion as well as the age of the voice, as seen on the following screenshot of our software :


happy, but not so much   normally happy    very happy




sad, but not so much  normally sad   very sad




angry, but not so much     normally angry   very angry




I applyed these algorithm to meaningful speech, and in particular to japanese speech.
Here is a screenshot of the corresponding software, in which you can see lower level controls than on the previous software) :


Related project:

The Maïdo and Gurby Experiment is a particular setup showing how autonomous creatures can coordinate socially to build a shared repertoire of syllables. Social coordination is achieved through the modulation of prosody in babbling sounds, which allows creatures to convey basic attitudes and emotionn.