Akteure
28 Einträge
Eintragsliste
OrganisationNewsadoo
Newsadoo (Linz) betreibt eine KI-gestützte Nachrichtenplattform, die Inhalte automatisch sammelt, mittels künstlicher Intelligenz versteht und sortiert und sie personalisiert sowie themenspezifisch zugänglich macht. Der personalisierte Newsfeed wird laufend anhand des Leseverhalt
LinzMedien & AV
OrganisationStoryclash
Storyclash (Linz) betreibt eine KI-Plattform für Creator- und Influencer-Intelligence im Marketing. Die KI durchsucht Millionen von Content-Beiträgen, identifiziert passende Creator, deckt Marken- und Wettbewerbssignale auf und filtert nach Glaubwürdigkeit, Compliance und Offenle
LinzMedien & AV- Forschung
The Binaural Rendering Toolbox. A Virtual Laboratory for Reproducible Research in Psychoacoustics
The Binaural Rendering Toolbox (BRT) is a set of software libraries, applications, and definitions aimed as a virtual laboratory for psychoacoustic experimentation.The BRT is developed in the framework of the SONICOM project 1 and will include the algorithms developed in the 3D Tune-In Toolkit 2 in a new open, extensible architecture.At the core of the BRT Toolbox, a library provides C++ implementations of listener models, source models, and environment models, including a growing collection of portings to different audio frameworks such as PureData, MaxMSP and VST plugins, by means of the Avendish library.In addition, the BRT also includes an application controlled via the Open Sound Control (OSC) protocol.This paper describes the architecture of the BRT, its main features, and its application to reproducible psychoacoustics experiments.The toolbox provides a complete trace of the experiment, including the delivered binaural audio, annotated with the listener and source movements.For this purpose, a new SOFA convention is proposed to store dynamic measurements, facilitating their use in the Auditory Model Toolbox (AMT).
Aktiv: Jänner 2024Medien & AV - Forschung
Deep learning’s shallow gains: a comparative evaluation of algorithms for automatic music generation
Abstract Deep learning methods are recognised as state-of-the-art for many applications of machine learning. Recently, deep learning methods have emerged as a solution to the task of automatic music generation (AMG) using symbolic tokens in a target style, but their superiority over non-deep learning methods has not been demonstrated. Here, we conduct a listening study to comparatively evaluate several music generation systems along six musical dimensions: stylistic success, aesthetic pleasure, repetition or self-reference, melody, harmony, and rhythm. A range of models, both deep learning algorithms and other methods, are used to generate 30-s excerpts in the style of Classical string quartets and classical piano improvisations. Fifty participants with relatively high musical knowledge rate unlabelled samples of computer-generated and human-composed excerpts for the six musical dimensions. We use non-parametric Bayesian hypothesis testing to interpret the results, allowing the possibility of finding meaningful non -differences between systems’ performance. We find that the strongest deep learning method, a reimplemented version of Music Transformer, has equivalent performance to a non-deep learning method, MAIA Markov, demonstrating that to date, deep learning does not outperform other methods for AMG. We also find there still remains a significant gap between any algorithmic method and human-composed excerpts.
Aktiv: März 2023Medien & AV