Seminars

Upcoming talks

Speaker: Guilherme Feulo e Micael Antunes
Date and time: Monday, June 25, 2018 - 17:00
Place: Auditório Antonio Gilioli - Bloco A
Abstract: In this seminar we present an algorithm for sensory dissonance modeling through a critical band filter, based on the work of William Sethares and John Pierce. We begin with a historical review of the term dissonance, aiming to define what we understood by sensory dissonance, followed by an explanation of proposed models, from Helmholtz to Vassilakis. We will also present a detailed study of a spectrum generated by the proposed algorithm, and evaluate the limitations imposed by it. At last, we will present musical examples of usage of these techniques for creative processes.

Speaker: Paulo Vitor Itaboraí
Date and time: Monday, August 20, 2018 - 17:00
Place: Antonio Gilioli Auditorium, IME/USP
Abstract: This Seminar will present a broad view of Control Theory in the field of acoustics, aiming to control resonances, partials and decays of an prepared musical instrument in Real Time. In particular, the main discussion will be about guaranteeing stability in a Feedback Control scenario, and possible applications in musical instrument's preparation and design.


(video presentation in portuguese)

Past talks

Speaker: Rodrigo Borges
Date and time: Monday, June 18, 2018 - 17:00
Place: Antonio Gilioli Auditorium, IME/USP
Abstract:

Music recommender systems typically use historical listening information for making personalized recommendations. This approach however keeps high rated songs always as better candidates in a greedy manner. We present a strategy for balancing safe (Exploitation) and novel (Exploration) recommendations in order to prevent suboptimal performance over the long term. The solution proposed is based in a reinforcement learning problem called multi-armed bandit that simulates a situation where someone is playing in several slot machines and needs to optimize his gains. The player starts without any knowledge about the machines as has to choose between the current best machine and new possibilities at each turn. Practical results from the literature are presented as enhancing long term recommendation as well as solving the problem of new items added to the dataset.


(video presentation in portuguese)

Speaker: Fábio Goródscy
Date and time: Monday, June 4, 2018 - 17:00
Place: Auditório Antonio Gilioli, IME/USP
Abstract:

We are presenting strategies for processing audio for query-by-humming, having the goal of matching melodies from MIDI files to the melody hummed in audio files. An application of query-by-humming is defined as interface where an user can hum a melody as he or she remembers and the application brings melodies from a MIDI repertoire that has some degree of similarity, depending on what is expected by the user to be similar, so that the user can discover more information from the melody he hummed.

We are showing a brief revision of key concepts and examples of transcription using algorithms as Melodia and ASyMuT. We discuss metrics for matching these sound representations and show some results using the presented strategies.

Speaker: Marcos Leal (mleal@ime.usp.br)
Date and time: Thursday, May 24, 2018 - 16:00
Place: Antonio Gilioli Auditorium, IME/USP
Abstract: (sorry about the abstract in Portuguese)
Redes Neurais são modelos de aprendizado de máquina extremamente populares na atualidade. Nesse seminário iremos dar uma pincelada histórica nas origens desse modelo, discutir os problemas que o levaram a ser abandonado por anos, apontar as soluções para então discutir as diversas arquiteturas da atualidade com foco em Redes Neurais Convolucionais. A expectativa desse seminário é dar uma visão geral do funcionamento das redes neurais, transmitir o que o processo de convolução espera no contexto de redes neurais e discutir os aspectos onde esse modelo apresenta aplicabilidade no contexto de Computação Musical.

Tags: neural networks, cnn, convolutional neural networks, computer music

Speaker: Giuliano Mega and Daniel Cukier (Playax)
Date and time: Monday, May 21, 2018 - 17:00
Place: Antonio Gilioli Auditorium, IME/USP
Abstract:

Índices de popularidade são instrumentos tradicionalmente utilizados na indústria da música para a comparação de artistas. Apesar da ampla aceitação, pouco se conhece a respeito da metodologia que embasa tais índices, o que os torna instrumentos por vezes difíceis de compreender e/ou criticar.

Neste seminário, nós vamos expor o método e as principais dificuldades enfrentadas durante a construção do índice Playax, um índice de metodologia transparente que agrega sinais de popularidade provenientes de dados de fontes heterogêneas (e.g. rádios, Spotify, Youtube, Instagram, Twitter, etc) em um único número. Em particular, vamos discutir as soluções e relações de compromisso adotadas na reconciliação de sinais contraditórios e que capturam aspectos de dimensões diferentes do fenômeno de popularidade em mídias distintas. A nossa esperança, com este trabalho, é a de iniciar um processo saudável de discussão e crítica que leve eventualmente a um instrumento prático, coerente e transparente para a comparação de artistas e outros tipos de colaboração relacionados a dados de contexto musical.


(video presentation in portuguese)

Speaker: Nicolas Figueiredo
Date and time: Monday, May 7, 2018 - 17:00
Place: Auditório Gilioli - Bloco A
Abstract:

Since 1957, computers have been used for synthesizing and processing audio signal. Some of the first techniques used were additive synthesis, AM/FM synthesis and subtractive synthesis. An alternative to these techniques is physical modeling, that utilizes mathematical descriptions of sound waves and physical components such as strings, tubes and membranes, to create musical signals. This seminar will present the main techniques for physical modeling of instruments, with special attention to waveguides, lumped models and state-space models.


(video presentation in portuguese)

Pages

Subscribe to Upcoming talks Subscribe to Past talks