No abstract

Considering music as a sequence of events with multiple complex dependencies, the Long Short-Term Memory (LSTM) architecture has proven very efficient in learning and reproducing musical styles. However, the generation of rhythms requires additional information regarding musical structure and accompanying instruments. In this paper we present DeepDrum, an adaptive Neural Network capable of generating drum rhythms under constraints imposed by Feed-Forward (Conditional) Layers which contain musical parameters along with given instrumentation information (e.g. bass and guitar notes). Results on generated drum sequences are presented indicating that DeepDrum is effective in producing rhythms that resemble the learned style, while at the same time conforming to given constraints that were unknown during the training process.

No abstract

No abstract

Abstract. The Greek Audio Dataset (GAD), is a freely available collection of audio features and metadata for a thousand popular Greek tracks. In this work, the creation process of the dataset is described together with its contents. Following the methodology of existing datasets, the GAD dataset does not include the audio content of the respective data due to intellectual property rights but it includes MIR important features extracted directly from the content in addition to lyrics and manually annotated genre and mood for each audio track. Moreover, for each track a link to available audio content in YouTube is provided in order to support researchers that require the extraction of new featuresets, not included in the GAD. The selection of the features extracted has been based on the Million Song Dataset in order to ensure that researchers do not require new programming interfaces in order to take advantage of the GAD.

No abstract

Although media content is increasingly produced, distributed, and consumed in multiple combinations of modalities, how individual modalities contribute to the perceived emotion of a media item remains poorly understood. In this paper we present MusicVideos (MuVi), a novel dataset for affective multimedia content analysis to study how the auditory and visual modalities contribute to the perceived emotion of media. The data were collected by presenting music videos to participants in three conditions: music, visual, and audiovisual. Participants annotated the music videos for valence and arousal over time, as well as the overall emotion conveyed. We present detailed descriptive statistics for key measures in the dataset and the results of feature importance analyses for each condition. Finally, we propose a novel transfer learning architecture to train Predictive models Augmented with Isolated modality Ratings (PAIR) and demonstrate the potential of isolated modality ratings for enhancing multimodal emotion recognition. Our results suggest that perceptions of arousal are influenced primarily by auditory information, while perceptions of valence are more subjective and can be influenced by both visual and auditory information. The dataset is made publicly available.

No abstract