Submitted by PAULO ROGERIO ZORZETTO HYPOLITHO null (paulozorzetto@gmail.com) on 2016-11-06T19:59:27Z No. of bitstreams: 1 dissertação de mestrado Paulo Zorzetto 2016.pdf: 7471098 bytes, checksum: 09d11ae1d701810ce164a7fb18725499 (MD5) / Approved for entry into archive by LUIZA DE MENEZES ROMANETTO null (luizaromanetto@hotmail.com) on 2016-11-10T13:56:32Z (GMT) No. of bitstreams: 1 zorzetto_p_me_ia.pdf: 7462080 bytes, checksum: e9ca8bc71dc1a78ff96f76fe1320147b (MD5) / Made available in DSpace on 2016-11-10T13:56:32Z (GMT). No. of bitstreams: 1 zorzetto_p_me_ia.pdf: 7462080 bytes, checksum: e9ca8bc71dc1a78ff96f76fe1320147b (MD5) Previous issue date: 2016-06-30 / Esta pesquisa tem como objetivo geral propor um Modelo Vocal para Percussionistas (MVP), na execução de obras solo em que o uso da voz é requerido, explorando elementos físicos e musicais que fazem parte das práticas diárias do estudante e do profissional da percussão. Para tanto, foram escolhidas três importantes peças do repertório percussivo para a aplicação dos conceitos propostos: ?Corporel, para um percussionista e sons do seu corpo (1985), de Vinko Globokar (1934); To the Earth (1985), de Frederic Rzewski (1938); e URSONATE (1922 - 1932), de Kurt Schwitters (1887 - 1948). Por possuírem características distintas, quanto ao uso da voz, cada uma delas possibilita o entendimento do MVP, como um paradigma teórico amplo o suficiente, para resolver problemas técnicos e interpretativos diversos, a partir da compreensão do contexto em que estão inseridas: a Música-Teatro. Sob as considerações da pesquisadora Claudia Hansen, e do professor Wladimir Mattos, foram estipuladas as bases deste gênero musical, bem como os perfis do percussionista e da percussão que o representam, além daquilo que deve ser levado em conta como sendo a voz e a fala. A partir de três pilares teórico-práticos - a fonoarticulação, a coarticulação e a realização prosódica - as obras escolhidas foram analisadas e preparadas para a performance, levando-se em conta as discussões ocorridas no âmbito da fonética e da fonologia, e as soluções encontradas foram esquematizadas por meio de tabelas, visando a facilitar a compreensão do MVP. Desta forma, esperamos que a proposta do presente trabalho auxilie futuras pesquisas, interpretações e discussões inseridas no universo percussão e voz, acentuando as cores desta tão tênue e simbiótica linha que perpassa os ambientes musical e teatral. / This research aims to propose a Vocal Method for Percussionists (VMP) for playing solo pieces where using the voice is required. The physical and musical elements that are part of the percussionist's daily practice were analyzed. In order to achieve that, three major, well established, percussion pieces were drafted regarding the application of the proposed methods: ?Corporel, for one percussionist using her/his body (1985), by Vinko Globokar (1934); To the Earth (1985), by Frederic Rzewski (1938); and URSONATE (1922 - 1932), by Kurt Schwitters (1887 - 1948). Despite their differences regarding how the voice is used, each one of them enables the reader to assimilate the VMP, as a theoretical paradigm broad enough to figure out a wide range of interpretative and technical issues, fully understanding the context in which they belong: the music-theater. Under Claudia Hansen's and professor Wlad Matos' research analysis, we established the basis of this music genre, as well as the percussionist's and percussion characteristics that define them, and what should be considered as voice and speech. Starting off from three practical-theoretical pillars - articulation of speech, articulatory gesture and the prosodic conduction - the selected pieces were thoroughly analysed, looking for any issues regarding the use of the voice and singling out schematic solutions that would emerge. Thus, we hope that our work may assist future research, interpretation and discussion regarding percussion and voice. Taking into consideration the very gentle and symbiotic confluence within the musical and theatrical environments.

In automatic speech recognition systems (ASRs), training is a critical phase to the system?s success. Communication media, either analog (such as analog landline phones) or digital (VoIP) distort the speaker?s speech signal often in very complex ways: linear distortion occurs in all channels, either in the magnitude or phase spectrum. Non-linear but time-invariant distortion will always appear in all real systems. In digital systems we also have network effects which will produce packet losses and delays and repeated packets. Finally, one cannot really assert what path a signal will take, and so having error or distortion in between is almost a certainty. The channel introduces an acoustical mismatch between the speaker's signal and the trained data in the ASR, which results in poor recognition performance. The approach so far, has been to try to undo the havoc produced by the channels, i.e. compensate for the channel's behavior. In this thesis, we try to characterize the effects of different transmission media and use that as an inexpensive and repeatable way to train ASR systems.