Juan José Burred

Dynamic Spectral Envelope Modeling for Timbre Analysis of Musical Instrument Sounds

We present a computational model of musical instrument sounds that focuses on capturing the dynamic behavior of the spectral envelope. A set of spectro-temporal envelopes belonging to different notes of each instrument are extracted by means of sinusoidal modeling and subsequent frequency interpolation, before being subjected to principal component analysis. The prototypical evolution of the envelopes in the obtained reduced-dimensional space is modeled as a nonstationary Gaussian Process. This results in a compact representation in the form of a set of prototype curves in feature space, or equivalently of prototype spectro-temporal envelopes in the time-frequency domain. Finally, the obtained models are successfully evaluated in the context of two music content analysis tasks: classification of instrument samples and detection of instruments in monaural polyphonic mixtures.

Adaptation of source-specific dictionaries in Non-Negative Matrix Factorization for source separation

This paper concerns the adaptation of spectrum dictionaries in audio source separation with supervised learning. Supposing that samples of the audio sources to separate are available, a filter adaptation in the frequency domain is proposed in the context of Non-Negative Matrix Factorization with the Itakura-Saito divergence. The algorithm is able to retrieve the acoustical filter applied to the sources with a good accuracy, and demonstrates significantly higher performances on separation tasks when compared with the non-adaptive model.

On the Use of Auditory Representations for Sparsity-Based Sound Source Separation

Sparsity-based source separation algorithms often rely on a transformation into a sparse domain to improve mixture disjointness and therefore facilitate separation. To this end, the most commonly used time-frequency representation has been the short time Fourier transform (STFT). The purpose of this paper is to study the use of auditory-based representations instead of the STFT. We first evaluate the STFT disjointness properties for the case of speech and music signals, and show that auditory representations based on the equal rectangular bandwidth (ERB) and Bark frequency scales can improve the disjointness of the transformed mixtures

Audio event detection based on layered symbolic sequence representations

We introduce a novel application of genetic motif discovery in symbolic sequence representations of sound for audio event detection. Sounds are represented as a set of parallel symbolic sequences, each symbol representing a spectral shape, and each layer indicating the contribution weights of each spectral shape to the sound. Such layered symbolic representations are input to a genetic motif discovery algorithm that detects and clusters recurrent and structurally salient sound events in an unsupervised and query less manner. The found motifs can be interpreted as statistical temporal models of spectral evolution. The system is successfully evaluated in two tasks: environmental sound event detection, and drum onset detection.

Polyphonic musical instrument recognition based on a dynamic model of the spectral envelope

We propose a new method for detecting the musical instruments that are present in single-channel mixtures. Such a task is of interest for audio and multimedia content analysis and indexing applications. The approach is based on grouping sinusoidal trajectories according to common onsets, and comparing each groups overall amplitude evolution with a set of pre-trained probabilistic templates describing the temporal evolution of the spectral envelopes of a given set of instruments. Classification is based on either an Euclidean or a probabilistic definition of timbral similarity, both of which are compared with respect to detection accuracy.

Genetic motif discovery applied to audio analysis

Motif discovery algorithms are used in bioinformatics to find relevant patterns in genetic sequences. In this paper, the application of such methods to audio analysis is proposed. In the presented system, sounds are first transformed into a sequence of discrete states, corresponding to characteristic spectral shapes. The resulting sequences are then subjected to the MEME algorithm for motif discovery, which estimates a structured statistical model for each found motif. The system is evaluated in two tasks: the discovery of repetitive patterns in a large sound database, and the detection of specific audio events in an audio stream. Both tasks are unsupervised and demonstrate the viability of the approach.

Speech-guided source separation using a pitch-adaptive guide signal model

In this paper, we present a new method to perform underdetermined audio source separation using a spoken or sung reference signal to inform the separation process. This method explicitly models possible differences between the spoken reference and the target signal, such as pitch differences and time lag. We show that the proposed algorithm outperforms state-of-the art methods.

Audio Content Analysis

Since the introduction of digital audio more than 30 years ago, computers and signal processors have been capable of storing, modifying and transmitting sound signals. Before the advent of the Internet, compression technologies and digital telephony, such systems were aimed at the highest possible reproduction quality from physical media, or constrained to very specialised voice recognition or security systems. The first set of widespread techniques aimed at the extraction of semantics from audio were automatic speech recognition (ASR) systems. In the last couple of years, largescale, online distribution of high-quality audio has become a reality, widening the type of sounds to be analysed to music and any other kind of sounds, and shifting computers to the central position on the user side of the audio distribution chain. This has mainly been motivated by the advances in audio compression algorithms, especially the enormously successful MP3, and in network technologies. Audio content analysis (ACA), i.e. the automatic extraction of semantic information from sounds, arose naturally from the need to efficiently manage the growing collections of data and to enhance man–machine communication. ACA typically delivers a set of numerical measures from the audio signals, called audio features, that offer a compact and representative description. Such measures are usually called low-level features to denote that they represent a low level of abstraction. Although the classification is not strict, it is possible to consider as low-level features the measures most directly tied to the shape of the signal in the time or spectral domain, and which are mostly applicable to any kind of audio. Midand high-level features provide information more easily processed and usable by humans like phonemes, words or prosody in the case of speech or melody, harmony and structure in the case of music. To ensure interoperability, both lowand mid-level features can be

Geometric multichannel common signal separation with application to music and effects extraction from film soundtracks

We address the task of separation of music and effects from dialogs in film or television soundtracks. This is of interest for film studios wanting to release films in new, previously unavailable languages when the original separated music and effects track is not available. For this purpose, we propose several methods for common signal extraction from a set of soundtracks in different languages, which are multichannel extensions of previous methods for center signal extraction from stereo signals. The proposed methods are simple, effective, and have an intuitive geometrical interpretation. Experiments show that the proposed methods improve the results provided by our previously proposed methods based on basic filtering techniques.

Sample Orchestrator : gestion par le contenu d'échantillons sonores

RESUME. Nous presentons les principaux travaux menes dans le projet Sample Orchestrator, destine au developpement de fonctions innovantes de manipulation d’echantillons sonores. Celles-ci se fondent sur des etudes consacrees a la description des sons, c’est-a-dire a la formalisation de structures de donnees pertinentes pour caracteriser le contenu et l’organisation des sons. Ces travaux ont ete appliques a l’indexation automatique des sons, ainsi qu’a la realisation d’applications inedites pour la creation musicale - synthese sonore interactive par corpus et aide informatisee a l’orchestration. Le projet a aussi comporte un important volet consacre au traitement de haute qualite des sons, a travers plusieurs perfectionnements du modele de vocodeur de phase ‐ traitement par modele sinusoidal dans le domaine spectral et calcul automatique des parametres d’analyse.