347 research outputs found
Modeling Durational Incompressibility
No full textWindmann A, Simko J, Wrede B, Wagner P. Modeling Durational Incompressibility. In: Bimbot F, ed. Speech in Life Sciences and Human Societies. Vol. 2. Red Hook, NY: Curran; 2014: 1375-1379
Algorithmic and perceptual approaches for polytopic recomposition of melodies by multiscale complexity criterion
No full textCette thèse explore sous différents aspects musicologiques, algorithmiques et perceptifs le principe des recompositions musicales dite « polytopiques ». Il s’agit d’un procédé de génération de variations musicales consistant à créer un nouveau contenu en réorganisant l’ordre temporel des motifs musicaux à partir d’un matériau d’origine, tout en préservant certaines propriétés de leurs relations structurelles. Ces relations structurelles sont décrites par un graphe polytopique qui rend compte des similarités entre motifs à différentes échelles de temps, sur la base du modèle « Système & Contraste » (S&C) proposé par (Bimbot et al. 2012, 2016). À travers une démarche créative, cette thèse vise à offrir de nouveaux paradigmes de recomposition musicale, mais également à évaluer la pertinence des hypothèses sousjacentes au modèle polytopique, en ce qu’il rend compte de la structure musicale. Après une présentation détaillée du modèle polytopique et de son utilisation pour recomposer des contenus musicaux, cette thèse évalue la pertinence des recompositions polytopiques par rapport à des recompositions aléatoires, grâce à un test perceptif mené sur un ensemble d’auditeurs (Gillot et al., 2019). Une première série de tests suggère un avantage net des recompositions polytopiques vis-à-vis de recompositions aléatoires comparables en terme de perturbations de l’ordonnancement temporel. Une seconde série de tests étudie différentes variantes d’algorithmes destinées à estimer la délimitation optimale des motifs mélodiques permettant d’obtenir des recompositions polytopiques satisfaisantes vis-à-vis de la surface musicale. Ces algorithmes sont basés sur la minimisation d’un critère de complexité alliant le modèle Système & Contraste et des principes de Transport Optimal. Cet ensemble de méthodes sont non seulement comparées à des annotations humaines, mais elles sont également soumises à une batterie de tests perceptifs sur un panel d’auditeurs, pour évaluer leur acceptabilité musicale. Les résultats indiquent une relative concordance entre les niveaux d’acceptabilité des méthodes testées et des annotations produites, avec toutefois de faibles écarts entre variantes des méthodes automatiques. Ce nouveau concept de recompositions polytopiques peut être utilisé de façon interactive à des fins créatives et un prototype pour applications mobiles a été développé en partenariat avec la PME Kercodex pour la musique au format MIDI.This PhD thesis explores the principles of “polytopic musical recomposition” under musicological, algorithmic and perceptual aspects. Polytopic musical recompositions are defined as a type of musical variations which consists in creating new musical content by rearranging the temporal order of musical segments from an original material, while preserving some of its structural properties. The structural properties are described as a polytopic graph accounting for the similarities between segments at different time scales, according to the “System & Contrast” (S&C) model proposed by (Bimbot et al. 2012, 2016).Based on a creative approach, this work aims at proposing new paradigms for musical re-composition, but also to assess the relevance of the assumptions behind the polytopic model, as of how it accounts for musical structure.After a detailed presentation of the polytopic model and its usability to recompose musical contents, this PhD thesis evaluates the relevance of polytopic recompositions as opposed to random recompositions, by means of a perceptual test carried out on a population of listeners (Gillot et al., 2019). A first series of tests suggests a clear advantage of polytopic recompositions over comparable random recompositions, as regards the time-line disturbances. A second series of tests studies a set of algorithmic variants for estimating the optimal melodic segments for polytopic recompositions, so as to optimize them with respect to the melodic surface. The algorithms are based on the minimization of a complexity criterion relying on the S&C model combined with Optimal Transport principles. The set of algorithmic methods are compared to human annotations, but also evaluated in the context of perceptual tests on a panel of listeners, so as to assess their musical acceptability. Results indicate a relative agreement between the level of acceptability of the algorithmic methods versus human annotations, but they show relatively small differences between variants of automatic methods. The new concept of polytopic recompositions can be used interactively for creative purposes, and a prototype for mobile devices has been developed for music in MIDI format in partnership with the Kercodex SME
System & Contrast : A Polymorphous Model of the Inner Organization of Structural Segments within Music Pieces
No full textInternational audienceThis article introduces a model called “System & Contrast” (S&C), which aims at describing the inner organization of structural segments within music pieces in terms of : (i) a carrier system, i.e. a sequence of morphological elements forming a multi-dimensional network of self-deducible syntagmatic relationships and (ii) a contrast, i.e. a substitutive element, usually the last one, which partly departs from the logic implied by the rest of the system.With a primary focus on pop music, the S&C model provides a framework to describe internal implication patterns in musical segments by encoding similarities and relations between its constitutive elements so as to minimize the complexity of the resulting description. It is applicable at several timescales and to a wide variety of musical dimensions in a polymorphous way, therefore offering an attractive meta-description of different types of musical contents. It has been used as a central component in the creation of a set of annotations for 380 pop songs (Bimbot, Sargent, Deruty, Guichaoua & Vincent, 2014).This article formalizes the S&C model, illustrates how it applies to music and establishes its filiation with Narmour’s Implication-Realization model (Narmour 1990, 1992) and Cognitive Rule-Mapping (Narmour, 2000). It introduces the Minimum Description Length scheme as a productive paradigm to support the estimation of S&C descriptions and sketches several tracks where concepts from the domain of Electrical Engineering and Communication Systems can be paralleled with aspects pertaining to the structural description of music patterns by the S&C model.Strongly based on an Engineering Science viewpoint, the S&C model establishes promising connections between Music Data Processing and Information Retrieval on the one hand, and modern theories in Music Perception and Cognition on the other hand, together with interesting perspectives in other areas in Musicology
Study of electromagnetic calorimeter trigger-primitive performances for CMS data selection
No full textSegmenting DNA sequence into words based on statistical language model
No full textThis paper presents a novel method to segment/decode DNA sequences based on n-gram statistical language model. Firstly, we find the length of most DNA “words” is 12 to 15 bps by analyzing the genomes of 12 model species. The bound of language entropy of DNA sequence is about 1.5674 bits. After building an n-gram biology languages model, we design an unsupervised ‘probability approach to word segmentation’ method to segment the DNA sequences. The benchmark of segmenting method is also proposed. In cross segmenting test, we find different genomes may use the similar language, but belong to different branches, just like the English and French/Latin. We present some possible applications of this method at last
VoiceHome-2 corpus : A corpus dedicated to distant-microphone speech processing in domestic environments
No full textPurpose:
This corpus includes reverberated, noisy speech signals spoken by 12 native French talkers in 4 houses (3 rooms per house) and recorded by an 8-microphone device at various angles and distances and in various noise conditions.
This corpus stands apart from other corpora in the field by the number of rooms and homes considered by the diversity of acoustic conditions recorded and by the facts that it is publicly available at no cost.
Other materials:
Article : N. Bertin, E. Camberlein, R. Lebarbenchon, E. Vincent, S. Sivasankaran, I. Illina and F. Bimbot: VoiceHome-2, an extended corpus for multichannel speech processing in real homes, Speech Communication, Elsevier : North-Holland, 2019, 106, pp.68-78. 〈10.1016/j.specom.2018.11.002〉.
Code baseline to reproduce article's results:
Localization and speech enhancement
Acoustic models and recognition scripts for automatic speech recognition
Related software to execute the baseline:
MBSS Locate (v2.0)
FASST
Documentation:
The corpus documentation is both available into the archive and hereafter by clicking on voiceHome-2_corpus_v1.0_documentation.pdf .
Terms of use
You may exploit the corpus for a non-commercial scientific purpose provided you mention it in any written work or software you derive from its use. Within a published article, paper or report, the corpus must appear in the bibliographical references.
Speaker records diffusion consent
All participants have given an informed and signed consent about public diffusion of recorded sentences.
Contact:
nancy [dot] bertin [at] irisa [dot] f
voiceHome corpus: A corpus dedicated to distant-microphone speech processing in domestic environments
No full textPurpose:
This corpus includes reverberated, noisy speech signals spoken by native French talkers in a lounge and recorded by an 8-microphone device at various angles and distances and in various noise conditions.
Room impulse responses and noise-only signals recorded in various real rooms and homes and baseline speaker localization and enhancement software are also provided.
This corpus stands apart from other corpora in the field by the number of rooms and homes considered and by the fact that it is publicly available at no cost.
Other materials:
Article: N. Bertin, E. Camberlein, E. Vincent, R. Lebarbenchon, S. Peillon, E. Lamandé, S. Sivasankaran, F. Bimbot, I. Illina, A. Tom, S. Fleury and E. Jamet: A French corpus for distant-microphone speech processing in real homes, Interspeech2016, Sep 2016, San Francisco, United States, 2016.
Related software to reproduce article's results:
Multi-Channel BSS Locate (v1.3)
FASST (v2.2.1)
Documentation (in french):
The corpus documentation is both available into the archive and hereafter by clicking on voiceHome_corpus_french_documentation_v1.2.pdf .
Terms of use
You may exploit the corpus for a non-commercial scientific purpose provided you mention it in any written work or software you derive from its use. Within a published article, paper or report, the corpus must appear in the bibliographical references.
Speaker records diffusion consent
All participants have given an informed and signed consent about public diffusion of recorded sentences.
New corpus version available : voiceHome-2 corpus
A new version of the corpus is available : voiceHome-2 corpus web pag
voiceHome-2 corpus - localization and speech enhancement baseline - code
No full textBaseline scripts for localization and speech enhancement for the voiceHome-2 corpus, a corpus dedicated to distant-microphone speech processing in domestic environments that is available at https://zenodo.org/record/1252143 and described in the paper N. Bertin, E. Camberlein, R. Lebarbenchon, E. Vincent, S. Sivasankaran, I. Illina and F. Bimbot: VoiceHome-2, an extended corpus for multichannel speech processing in real homes, Speech Communication, Elsevier : North-Holland, 2019, 106, pp.68-78. ⟨10.1016/j.specom.2018.11.002⟩
Study of electromagnetic calorimeter trigger-primitive performances for CMS data selection
No full textDescription of Chord Progressions by Minimal Transport Graphs Using the System & Contrast Model
No full textInternational audienceIn this paper we model relations between chords by minimal transport and we investigate different types of dependencies within chord sequences. For this purpose we use the “System & Contrast” (S&C) model, designed for the description of music segments at the phrase scale, to infer non-sequential structures called chord progression graphs (CPG). Minimal transport is defined as the shortest displacement of notes, in semitones, between a pair of chords. The paper present three algorithms to find CPGs for chords sequences: one is sequential, and two others are based on the S&C model. The three methods are compared using the perplexity as an efficiency measure. The experiments on a corpus of 45 segments taken from songs of multiple genres, indicate that optimization processes based on the S&C model outperform the sequential model with a decrease in perplexity over 1.0
- …
