Automatic refinement of an expressive speech corpus assembling subjective perception and automatic classification

Abstract

tomatic refinement of an expressive speech corpus assembling subjective perception and automatic classification. Speech Communication, Elsevier : North-Holland, 2009, 51 (9) This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT AbstractThis paper presents an automatic system able to enhance expressiveness in speech corpora recorded from acted or stimulated speech. The system is trained with the results of a subjective evaluation carried out on a reduced set of the original corpus. Once the system has been trained, it is able to check the complete corpus and perform an automatic pruning of the unclear utterances, i.e. with expressive styles which are different from the intended corpus. The content which most closely matches the subjective classification remains in the resulting corpus. An expressive speech corpus in Spanish, designed and recorded for speech synthesis purposes, has been used to test the presented proposal. The automatic refinement has been applied to the whole corpus and the result has been validated with a second subjective test.

Spectral Properties and Prosodic Parameters of Emotional Speech in Czech and Slovak

“…The result of final multiple comparison of ANOVA also confirms good correlation between particular emotions. In the next future, we plan to use results of ANOVA and hypothesis test for creation of the database of values for emotional speech classifier based on statistical evaluation approach (Iriondo et al 2009), or it can be used for identification of speaker emotional states or in realtime emotion recognition systems (Attasi & Smékal 2008).…”

Evaluation of influence of spectral and prosodic features on GMM classification of Czech and Slovak emotional speech

“…Suprasegmental features comprise statistical values of parameters describing prosody by duration, fundamental frequency, and energy. Included in this category is also a separate group of features constituting the voice quality parameters: jitter, shimmer [11], Hammarberg index [12], LiljencrantsFant features [13], and spectral tilt [14]. All mentioned speech identification systems and classifiers are usually based on statistical approach, using the discriminative or artificial neural networks [15,16], hidden Markov models (HMM) [17], or Gaussian mixture models (GMM) [18,19].…”

Reliable Pitch Marking of Affective Speech at Peaks or Valleys Using Restricted Dynamic Programming

“…The generation and, hence, the labeling of speech databases is one of the hot topics in affective US-TTS synthesis, e.g., [10]- [12] (see [13] for a review of previous emotional speech data collections). In this context, if an affective (or emotional) US-TTS system is considered for the generation of a talkinghead, the complexity of the corpus generation and annotation process increases.…”