«Máquina de Ouver» is being developed by a research team from the Faculty of Sciences and Technology of the University of Coimbra, Portugal. This research work is trying to find a plausible solution for the visual representation of speech expressiveness through typographic composition and manipulation.
This system takes a sound recording of speech and its corresponding transcription as the input, analyses sound features, speech speed variation, and pauses and then maps these values to typographic variables such as size, weight, leading, etc.
The only manual stage of this process is the initial one, where the transcription and annotation of the text need to be done priorly to the sound analysis. The creation of posters or video artifacts is automated through InDesign and After Effects scripting.
With the supervision of Penousal Machado, Ana Boavida and Pedro Martins, this work is part of João Couceiro e Castro's Master Thesis in Design and Multimedia project. The use of poetry performances as input is due to its underlying expressiveness, thus making possible to get more dynamic and interesting results.
To generate the artifacts present in this submission we used excerpts from a performance by Mário Viegas of the Armindo Mendes de Carvalho poem "Cantiga do Ais", retrieved from the RTP Archives.
All the material submitted to SELECTED 2019 represents some of the practical results achieved so far.