Titulo Estágio
The Poem's Rythm
Área Tecnológica
Inteligência Artificial
Local do Estágio
CISUC
Enquadramento
Automatic music composition is a research topic that has thrilled generations of computer scientists and hás registered great strides in recent decades. There is in this area a wide range of technologies and algorithms already available which is allowing the development of a wide variety of high quality applications. Some aspects, however, remain without satisfactory solutions, thus offering opportunities for interesting research.
One of the challenging aspects is the construction of systems that are able to autonomously compose both music and lyrics for a song. To find answers to this problem, researchers have sought to develop solutions for the generation of lyrics from music, or vice versa. An example is the system Tra-la-Lyrics [1], developed at DEI, which builds a lyric from a given melody.
In this project we intend to explore techniques for text analysis and generation to compose appropriate rhythmic melodic rhythms, and develop them in an application that demonstrates their potential. The application should involve other techniques that extend the initially explored, particularly by including algorithms for generation of melody to given rhythms, or the inclusion of speech processing techniques as an alternative or complement to text analysis.
[1] H. Gonçalo Oliveira, A. Cardoso, F.C. Pereira , Tra-la-Lyrics: An approach to generate text based on rhythm, in Wiggins & Cardoso editors, Proceedings of the 4th joint workshop on Computational Creativity (IJWCC'07), London
* This reserach may be awarded an initiation research grant given by Project RECARDI
Objetivo
The aims of this Project are:
(a) to identify, study and explore techniques for generating rhythms from songs' lyrics;
(b) to build a prototype application that demonstrates the feasibility of the chosen techniques;
(c) to study, implement and test an extension of the developed prototype, involving additional resources to identify in the course of the Project (eg, generation of melody to rhythm, speech processing techniques, etc.)
Plano de Trabalhos - Semestre 1
A1. (M1-M2) State of the art research - research and document the relevant SoA, publications and software review, for the study of the problem at hand; detailed characterization of the problem to be solved stating specific goals and requirements;
A2. (M3) Design Proposal - conceive a design proposal for the specific context, mapping of assumptions and milestones for the rest of the project; identification of quality attributes, and architectural drivers for the proposed solution, including interaction models and user interface draft
Plano de Trabalhos - Semestre 2
A3. (M4 - M5) Prototyping, first stage- implementation of a proof-of-concept prototype, tuning or corrections based on evaluation results;
A4. (M5 - M7) Prototyping, 2nd stage - extend initial implementation with additional resources;
A5. (M7) Evaluation - definition of evaluation criteria and method; performance of evaluation and analysis of results
A6. (M9 - M10) Statement of Learning - production of paper/thesis with process report and reflection on knowledge production
Condições
This research may be awarded an initiation research grant given by Project RECARDI
The student will integrate the Cognitive and Media Systems Group of CISUC and will use resources available in one of its laboratories.
Orientador
Amilcar Cardoso
amilcar@dei.uc.pt 📩