In this article the I3Media corpus is presented, a trilingual (Catalan, English, Spanish) speech database of neutral and emotional material collected for analysis and synthesis purposes. The corpus is actually made up of six different subsets of material: a neutral subcorpus, containing emotionless utterances; a ‘dialog' subcorpus, containing typical call center utterances; an ‘emotional' corpus, a set of sentences representative of pure emotional states; a ‘football' subcorpus, including utterances imitating a football broadcasting situation; a ‘SMS' subcorpus, including readings of SMS texts; and a ‘paralinguistic elements' corpus, including recordings of interjections and paralinguistic sounds uttered in isolation. The corpus was read by professional speakers (male, in the case of Spanish and Catalan; female, in the case of the English corpus), carefully selected to meet criteria of language competence, voice quality and acting conditions. It is the result of a collaboration between the Speech Technology Group at Telefónica Investigación y Desarrollo (TID) and the Speech and Language Group at Barcelona Media Centre d'Innovació (BM), as part of the I3Media project.

We present work in progress aiming to build tools for the normalization of User-Generated Content (UGC). As we will see, the task requires the revisiting of the initial steps of NLP processing, since UGC (micro-blog, blog, and, generally, Web 2.0 user texts) presents a number of non-standard communicative and linguistic characteristics, and is in fact much closer to oral and colloquial language than to edited text. We present and characterize a corpus of UGC text in Spanish from three different sources: Twitter, consumer reviews and blogs. We motivate the need for UGC text normalization by analyzing the problems found when processing this type of text through a conventional language processing pipeline, particularly in the tasks of lemmatization and morphosyntactic tagging, and finally we propose a strategy for automatically normalizing UGC using a selector of correct forms on top of a pre-existing spell-checker.