Augmented Auditory Representation of e-Texts for Text-to-Speech Systems

Abstract

Abstract.Emerging electronic text formats include hierarchical structure and visualization related information that current Text-to-Speech (TtS) systems ignore. In this paper we present a novel approach for composing detailed auditory representation of e-texts using speech and audio. Furthermore, we provide a scripting language (CAD scripts) for defining specific customizations on the operation of a TtS. CAD scripts can be assigned as well to specific text meta-data to enable their discrete auditory representation. This approach can form a mean for a detailed exchange of functionality across different TtS implementations. Moreover, it can be hosted to current TtS systems with minor (or major) modifications. Finally, we briefly present the implementation of DEMOSTHeNES Composer for augmented auditory generation of meta-text using the above methodology.

A Methodology for the Extraction of Reader's Emotional State Triggered from Text Typography

“…In Figure 8 a proposed real-time system that automatically produces emotional annotation to documents and conveys the visual elements into acoustic modality using expressive speech synthesis is presented. Future work includes the use of an e-TSA composer (Xydas & Kouroupetroglou, 2001a, 2001b (Xydas et al, 2005) on the DEMOSTHeNES Text-to-Speech platform (Xydas & Kouroupetroglou, 2001c) and models for expressive speech synthesis as proposed by Schröder (Schröder, 2006), for acoustic rendition of emotionally annotated documents. …”

Auditory Accessibility of Metadata in Books: A Design for All Approach

“…As we described above, the book is formatted in LogicML. Following previous works on Document-to-Audio conversion [30], the semantics meta-data can be acoustically represented by specific auditory elements like (a) alternative text insertion in the document's text stream, (b) modifications in the prosody, (c) switching voices and (d) inserting other sounds like earcons and auditory icons in the waveform stream, according to the class of meta-data provided in the e-book. The user can be trained to recognize and to combine speech and sounds with specific commands and events.…”

“…3) [29] [30]. It will be mapped in specific acoustic elements, as mentioned above, producing a new annotated document and auditory synthesizer will implement the mapping (the output files can be e.g.…”

Modeling Improved Prosody Generation from High-Level Linguistically Annotated Corpora

“…Concept-to-Speech (CtS) systems (i.e. a Natural Language Generation -NLG -system coupled with a TtS system [9]) can provide linguistic information which can be used in prosody modeling [10], [11].…”