Constarium
← Search

Data · dataset · 2026

Research data on hybrid composition between rock and northeastern Brazilian music supported by artificial intelligence

Listed in REDU - Unicamp Institutional Research Data Repository

The dataset associated with this research was produced throughout a Research-Creation process aimed at investigating the interaction between the composer and artificial intelligence systems in musical composition.

Description

Rather than focusing on the specific software employed, the research documents the compositional processes established through the interaction between human creative decision-making and AI-assisted music generation.

The dataset includes the prompts used during the experiments, the audio files generated by the AI system, musical transcriptions, spectral analyses produced with Sonic Visualiser, intermediate compositional materials, and the final musical works developed throughout the research. It also contains analytical records comparing original audio tracks and separated stems, allowing the investigation of timbral characteristics and compositional procedures.

Read the rest (2 more)

The research methodology combined artistic practice and qualitative analysis. Compositional experiments were carried out using artificial intelligence as a creative partner, followed by a process of listening, critical evaluation, selection, editing, and further musical development by the researcher. Throughout the study, all compositional decisions and interactions with the AI system were documented in order to make the creative process as transparent as possible.

The dataset is primarily qualitative in nature and was produced to document the artistic research process that led to the final compositions and analytical discussions presented in the dissertation. As supplementary material, the repository also includes the prompts employed during the AI-assisted experiments and their corresponding audio results, contributing to the transparency and reproducibility of the methodological procedures adopted in this research.

Links

Where it is published

Catalogue records · 1

Topics

Stated by source
Arts and Humanities
Inferred from text
Audio 75%
Provenance · 1 source records, 11 field assertions
SourceKeyLast seenRaw
REDU - Unicamp Institutional Research Data Repositorydoi:10.25824/redu/3LJTOH9 d agoJSON v1
FieldAssertionExtractorEvidence
concepts[field].anzsrc:group:4602mapping · redu unicamp brvocabulary-mapper@1.0.0keywords['Artificial intelligence']
concepts[field].dataverse_subject:arts-and-humanitiessource · redu unicamp brconnector:redu_unicamp_br@1.0.0/subjects
concepts[field].local:field:computer-science-aimapping · redu unicamp brconnector:redu_unicamp_br@1.0.0/subjects
concepts[field].local:field:humanitiesmapping · redu unicamp brconnector:redu_unicamp_br@1.0.0/subjects
concepts[modality].local:modality:audioenrichment · redu unicamp brkeyword-concept-rules@1.0.0title+description (75%)
created_datesource · redu unicamp brconnector:redu_unicamp_br@1.0.0
descriptionsource · redu unicamp brconnector:redu_unicamp_br@1.0.0/description
publication_datesource · redu unicamp brconnector:redu_unicamp_br@1.0.0
titlesource · redu unicamp brconnector:redu_unicamp_br@1.0.0/name
updated_datesource · redu unicamp brconnector:redu_unicamp_br@1.0.0
version_labelsource · redu unicamp brconnector:redu_unicamp_br@1.0.0