Audio Signal Processing/DSP, Acoustics, Music

The goal of live song identification is to recognize a song based on a short, noisy cell phone recording of a live performance. We propose a system for known-artist live song identification and provide empirical evidence of its feasibility. The proposed system represents audio as a sequence of hashprints, which are binary fingerprints that are derived from applying a set of spectro-temporal filters to a spectro-gram representation. The spectro-temporal filters can be learned in an unsupervised manner on a small amount of data, and can thus tailor its representation to each artist. Matching is performed using a cross-correlation approach with downsampling and rescoring. We evaluate our approach on the Gracenote live song identification benchmark data set, and compare our results to five other base-line systems. Compared to the previous state-of-the-art, the proposed system improves the mean reciprocal rank from .68 to .79, while simultaneously reducing the average runtime per query from 10 seconds down to 0.9 seconds.

Humans tend to organize perceived information into hierarchies and structures, a principle that also applies to music. Even musically untrained listeners unconsciously analyze and segment music with regard to various musical aspects, for example, identifying recurrent themes or detecting temporal boundaries between contrasting musical parts. This paper gives an overview of state-of-the-art methods for computational music structure analysis, where the general goal is to divide an audio recording into temporal segments corresponding to musical parts and to group these segments into musically meaningful categories. There are many different criteria for segmenting and structuring music audio. In particular, one can identify three conceptually different approaches, which we refer to as repetition-based, novelty-based, and homogeneity-based approaches. Furthermore, one has to account for different musical dimensions such as melody, harmony, rhythm, and timbre. In our state-of-the-art report, we address these different issues in the context of music structure analysis, while discussing and categorizing the most relevant and recent articles in this field.

As a result of massive digitization efforts and the world wide web, there is an exploding amount of available digital data describing and representing music at various semantic levels and in diverse formats. For example, in the case of the Beatles songs, there are numerous recordings including an increasing number of cover songs and arrangements as well as MIDI data and other symbolic music representations. The general goal of music synchronization is to align the multiple information sources related to a given piece of music. This becomes a difficult problem when the various representations reveal significant differences in structure and polyphony, while exhibiting various types of artifacts. In this paper, we address the issue of how music synchronization techniques are useful for automatically revealing critical passages with significant difference between the two versions to be aligned. Using the corpus of the Beatles songs as test bed, we analyze the kind of differences occurring in audio and MIDI versions available for the songs.

In recent years Sonar processing systems have migrated from custom hardware to open system architecture with many off-the-shelf components. With the rapid increase in the computing capacity of Graphics Processing Units (GPUs) it is attractive to have computational intensive algorithms run on GPUs. However it is cumbersome to implement such software using the parallel programming techniques required for GPU. HPCGen have developed a cross-compiler that converts high level Matlab compatible source code to executables for GPU or multi-core CPU with no programming effort required. The compiled code can achieve over 60 times performance improvement for algorithms such as Synthetic Aperture Imaging when used on a nVidia GeForce GTX 480 compares to the Intel i7 Quad Core 920 @ 2.8GHz. This paper will present the results of performance improvements of different computational intensive sonar algorithms together with the benefits of using this GPU enabled compiler to greatly reduce the costs in software development and life cycle support of a Sonar processing system while making full use of inexpensive GPU power.

This article discusses Phase Distortion synthesis and its application to arbitrary input signals. The main elements that compose the technique are presented. Its similarities to Phase Modulation are discussed and the equivalence between the two techniques is explored. Two alternative methods of distorting the phase of an arbitrary signal are presented. The first is based on the audio-rate modulation of a first-order allpass filter coefficient. The other method relies on a re-casting of the Phase Modulation equation, which leads to a heterodyned form of waveshaping. The relationship of these implementations to the original technique is explored in detail. Complementing the article, a number of examples are discussed, demonstrating the application of the technique as an interesting digital audio effect.

This paper presents TarsosDSP, a framework for real-time audio analysis and processing. Most libraries and frameworks offer either audio analysis and feature extraction or audio synthesis and processing. TarsosDSP is one of a only a few frameworks that offers both analysis, processing and feature extraction in real-time, a unique feature in the Java ecosystem. The framework contains practical audio processing algorithms, it can be extended easily, and has no external dependencies. Each algorithm is implemented as simple as possible thanks to a straightforward processing pipeline. TarsosDSP's features include a resampling algorithm, onset detectors, a number of pitch estimation algorithms, a time stretch algorithm, a pitch shifting algorithm, and an algorithm to calculate the Constant-Q. The framework also allows simple audio synthesis, some audio effects, and several filters. The Open Source framework is a valuable contribution to the MIR-Community and ideal fit for interactive MIR-applications on Android.

In the 1960s and 1970s, most electrical (or analog) synthesizers were based on a sound synthesis technique called subtractive synthesis, in which a spectrally rich source signal, typically a sawtooth or rectangular pulse wave, is filtered with a time-varying filter. In the 1980s and 1990s this synthesis technique was almost overtaken by other synthesis techniques but in the mid-1990s musicians started again to show interest in the ''warm`` sound of subtractive synthesis. To meet this interest, a Swedish company Clavia introduced in 1995 the NordLead synthesizer that used the subtractive sound synthesis but in digital domain. Furthermore, they introduced the term ''virtual analog`` to represent the digital simulation of the analog audio devices. Since then research on digital signal processing methods for subtractive sound synthesis has increased and more and more interest has been shown to the topic in the past few years. In fact, the last year (2010) was again record-breaking when the number of publication on this topic is considered. In this paper, the history of the publications of virtual analog synthesis research is reviewed. In addition to the publication count over the years, the viewpoints of this paper include the distribution of the methodologies presented in the publications.

In a Musical Industry:
Signal Processing works are tremendously used in Musical industry for Digital Audio
Production where recording, mixing and editing and many Audio FX are implemented so that
the resultant audio file obtained is of a high quality with improved timbre (tonal accuracy)
and other sound parameters. An audio file of an instrumental format (not the commercial mp3 songs) consists melodies,
harmonies, tunes of different instruments been played along with vocals (however vocal is
not considered as of now). For a person who want to learn how a song ( melody or tune ,the
way the wave is varying w.r.t. time) can be played in his instrument, he must know what is
the pitch and the scale ( if necessary chords ) .
A musician who has a good knowledge on frequencies of every note that is being played, can
reproduce by trial and error methods to start from the Root note, then finding chords and
scale then the exact tune. But for a beginner who wants to play the exact music might be a
tough task.
Now, a software in which the required audio file is given , working out the processing steps
inside it generates the frequencies ( pitch ) , next chords, later for scale then itself shows the
notes with respect to time.
Adding audio effects to this project makes it look great for a musician to perform advanced
signal processing steps. Channel separation, tempo changing etc., can be included in the later
versions

In quasi-bandlimited classical waveform oscillators, the aliasing distortion present in a trivially sampled waveform can be reduced in the digital domain by applying a tabulated correction function. This paper presents an approach that applies the correction function in the differentiated domain by synthesizing a bandlimited impulse train (BLIT) that is integrated to obtain the desired bandlimited waveform. The ideal correction function of the BLIT method is infinitely long and in practice needs to be windowed. In order to obtain a good alias-reduction performance, long tables are typically required. It is shown that when a short look-up table is used, a windowed ideal correction function does not provide the best performance in terms of minimizing aliasing audibility. Instead, audibly improved alias-reduction performance can be obtained using a look-up table that has a parametric control over the low-order generations of aliasing. Some practical parametric look-up table designs are discussed in this paper, and their use and alias-reduction performance are exemplified. The look-up table designs discussed in this paper providing the best alias-reduction performance are parametric window functions and least-squares optimized multi-band FIR filter designs.

Convolution reverb is the process used for reverberating a signal by an impulse response of an actual space. The product of this operation is a third signal, containing reverberation of the space where the impulse response is captured. This project is based on creating a convolution reverb plugin on MaxMSP and evaluating the perceptual quality of convolution reverb.
The real-time fast convolution reverb plugin was implemented using multiple frequency delay line non-uniform partitioned convolution method on MaxMSP. An experimental design methodology was introduced in order to evaluate the perceptual quality of convolution reverb. In order to realise, anechoic drum kit music samples were recorded and re-recorded in a chamber to form the control signal. Also, using same equipment and setup, an impulse response was captured in the same chamber to form the test sample. These samples were then used for subjective analysis in a pair-wise categorical preference type listening test. The samples were also analysed in their spectrogram views in order to analyse the quality objectively.
Listening tests were designed so that realism, quality and personal preference categories were present. Null hypothesis was proved, where the realism difference of the two samples resulted in 52% of the participants preferring chamber reverb (control signal). However personal preference category resulted in 61%of the subjects preferring the chamber reverb. This was justified by objective analysis, where the convolution reverb shown to be having a faster decay rate for high frequency bands, thus sounding unnatural.

This paper presents SonaGraph, a framework and an application for a simplified but efficient harmonic spectrum analyzer suitable for assisted and algorithmic composition. The model is inspired by the analog Sonagraph and relies on a constant-Q bandpass filter bank. First, the historical Sonagraph is introduced, then, starting from it, a simplified ("cartoonified") model is discussed. An implementation in SuperCollider is presented that includes various utilities (interactive GUIs, music notation generation, graphic export , data communication). A comparison of results in relation to other tools for assisted composition is presented. Finally, some musical examples are discussed, that make use of spectral data from SonaGraph to generate, retrieve and display music information.

Mediante  el  análisis  de  señales  acústicas  producto  de  embarcaciones  sobre  el  agua  tomadas  en campo,  se  desarrolló  un  algoritmo  para  la  detección,  clasificación  y  localización  de  dichas embarcaciones.  Además  se  tuvieron  en  cuenta  condiciones  de  operación  de  las  embarcaciones como lo son: la carga, la velocidad y condiciones propias del medio de propagación como lo es el ruido  ambiente.  Las  mediciones  se  realizaron  en  condiciones  controladas  con  embarcaciones  de prueba  en  una  piscina  olímpica,  en  donde  variables  como  las  dimensiones,  la  temperatura  y  la profundidad de esta, fueron determinadas. Además se contó con grabaciones realizadas en campo (Cartagena, Colombia), con las cuales se verificó el funcionamiento del algoritmo con señales de embarcaciones reales y diferentes señales de ruido ambiente.Para  los  registros  con  diferentes  condiciones  de  operación  de  las  embarcaciones  de  prueba,  se realizaron tomas a diferentes velocidades y con diferentes cargas (peso), para luego ser ingresados en  el  algoritmo.  Por  último,  el  desarrollo  de  este  se  basó  en  métodos  paramétricos  para  la estimación de densidad espectral de potencia (PSD) para señales aleatorias como lo son las señales de las diferentes embarcaciones analizadas en este trabajo.

This research proposes new approaches for computational identification of Persian musical modes. This involves constructing a database of audio musical files and developing computer algorithms to perform a musical analysis of the samples. Essential features, the spectral average, chroma, and pitch histograms, and the use of symbolic data, are discussed and compared. A tonic detection algorithm is developed to align the feature vectors and to make the mode recognition methods independent of changes in tonality. Subsequently, a geometric distance measure, such as the Manhattan distance, which is preferred, and cross correlation, or a machine learning method (the Gaussian Mixture Models), is used to gauge similarity between a signal and a set of templates that are constructed in the training phase, in which data-driven patterns are made for each dastgàh (Persian mode). The effects of the following parameters are considered and assessed: the amount of training data; the parts of the frequency range to be used for training; down sampling; tone resolution (12-TET, 24-TET, 48-TET and 53-TET); the effect of using overlapping or non overlapping frames; and silence and high-energy suppression in pre-processing. The santur (hammered string instrument), which is extensively used in the musical database samples, is described and its physical properties are characterised; the pitch and harmonic deviations characteristic of it are measured; and the inharmonicity factor of the instrument is calculated for the first time.

The results are applicable to Persian music and to other closely related musical traditions of the Mediterranean and the Near East. This approach enables content-based analyses of, and content based searches of, musical archives. Potential applications of this research include: music information retrieval, audio snippet (thumbnailing), music archiving and access to archival content, audio compression and coding, associating of images with audio content, music transcription, music synthesis, music editors, music instruction, automatic music accompaniment, and setting new standards and symbols for musical notation.

In this thesis, the design of a music synthesizer for systems suffering from limitations in computing power and memory capacity is presented. First, different possible synthesis techniques are reviewed and their applicability in computationally efficient music synthesis is discussed. In practice, the applicable techniques are limited to additive and source-filter synthesis, and, in special cases, to frequency modulation, wavetable and sampling synthesis.

Next, the design of the structures of the applicable techniques are presented in detail, and properties and design issues of these structures are discussed. A major implementation problem is raised in digital source-filter synthesis, where the use of classic waveforms, such as sawtooth wave, as the source signal is challenging due to aliasing caused by waveform discontinuities. Methods for existing bandlimited waveform synthesis are reviewed, and a new approach using polynomial bandlimited step function is presented in detail with design rules for the applicable polynomials. The approach is also tested with two different third-order polynomials. They reduce aliasing more at high frequencies, but at low frequencies their performance is worse than with the first-order polynomial. In addition, some commonly used sound effect algorithms are reviewed with respect to their applicability in computationally efficient music synthesis.

In many cases the sound synthesis system must be capable of producing music consisting of various different sounds ranging from real acoustic instruments to electronic instruments and sounds from nature. Therefore, the music synthesis system requires careful sound design. In this thesis, sound design rules for imitation of various sounds using the computationally efficient synthesis techniques are presented. In addition, the effects of the parameter variation for the design of sound variants are presented.

Music is a ubiquitous and vital part of the lives of billions of people worldwide. Musical creations and performances are among the most complex and intricate of our cultural artifacts, and the emotional power of music can touch us in surprising and profound ways. In view of the rapid and sustained growth of digital music sharing and distribution, the development of computational methods to help users find and organize music information has become an important field of research in both industry and academia. The Dagstuhl Seminar 16092 was devoted to a research area known as music structure analysis, where the general objective is to uncover patterns and relationships that govern the organization of notes, events, and sounds in music. Gathering researchers from different fields, we critically reviewed the state of the art for computational approaches to music structure analysis in order to identify the main limitations of existing methodologies. This triggered interdisciplinary discussions that leveraged insights from fields as disparate as psychology, music theory, composition, signal processing, machine learning, and information sciences to address the specific challenges of understanding structural information in music. Finally, we explored novel applications of these technologies in music and multimedia retrieval, content creation, musicology, education, and human-computer interaction. In this report, we give an overview of the various contributions and results of the seminar. We start with an executive summary, which describes the main topics, goals, and group activities. Then, we present a list of abstracts giving a more detailed overview of the participants' contributions as well as of the ideas and results discussed in the group meetings of our seminar. Seminar February 28–March 4, 2016 – http://www.dagstuhl.de/16092 1998 ACM Subject Classification H.5.5 Sound and Music Computing In this executive summary, we start with a short introduction to computational music structure analysis and then summarize the main topics and questions raised in this seminar.

A prominent categorization of Indian classical music is the Hindustani and Carnatic traditions. The distinction is geographical with the two styles having evolved under distinctly different historical and cultural influences. Both styles are grounded in the melodic and rhythmic framework of raga and tala. The styles differ along dimensions such as structure of a performance, aesthetics, voice production and instrumentation. In this work we explore methods to distinguish the two prominent styles by analysing the extracted melodic contour from audio segments of vocal performances. The assumption that style distinctions are represented in the melodic contour is validated via human listening tests. The presented features can be applied to achieve automatic classification of style.

Trivially sampled geometric waveforms such as the rectangular pulse wave used in subtractive sound synthesis suffer from aliasing caused by the discontinuities in the waveform or its derivative. Several algorithms for the reduction of aliasing distortion have been suggested, providing either complete removal or great suppression of aliasing. Some antialiasing oscillators utilize variable fractional delay filters as an essential part of the algorithm. In this paper, these oscillators are reviewed with an emphasis on motivating the use of the fractional delay filters.

This paper describes a sound synthesis technique that modulates the coefficients of allpass filter chains using audio-rate frequencies. It was found that modulating a single allpass filter section produces a feedback AM--like spectrum, and that its bandwidth is extended and further processed by non-sinusoidal FM when the sections are cascaded. The cascade length parameter provides dynamic bandwidth control to prevent upper range aliasing artifacts, and the amount of spectral content within that band can be controlled using a modulation index parameter. The technique is capable of synthesizing rich and evolving timbres, including those resembling classic virtual analog waveforms. It can also be used as an audio effect with pitch-tracked input sources. Soft- ware and sound examples are available at http://www.acoustics.hut.fi/publications/papers/dafx09-cm/.

The rapidly growing corpus of digital audio material requires novel retrieval strategies for exploring large music collections. Traditional retrieval strategies rely on metadata that describe the actual audio content in words. In the case that such textual descriptions are not available, one requires content-based retrieval strategies which only utilize the raw audio material. In this contribution, we discuss content-based retrieval strategies that follow the query-by-example paradigm: given an audio query, the task is to retrieve all documents that are somehow similar or related to the query from a music collection. Such strategies can be loosely classified according to their specificity, which refers to the degree of similarity between the query and the database documents. Here, high specificity refers to a strict notion of similarity, whereas low specificity to a rather vague one. Furthermore, we introduce a second classification principle based on gran-ularity, where one distinguishes between fragment-level and document-level retrieval. Using a classification scheme based on specificity and granularity, we identify various classes of retrieval scenarios, which comprise audio identification, audio matching, and version identification. For these three important classes, we give an overview of representative state-of-the-art approaches, which also illustrate the sometimes subtle but crucial differences between the retrieval scenarios. Finally, we give an outlook on a user-oriented retrieval system, which combines the various retrieval strategies in a unified framework. 1998 ACM Subject Classification H.5.5 Sound and Music Computing, J.5 Arts and Humanities– Music, H.5.1 Multimedia Information Systems, I.5 Pattern Recognition

Chroma-based audio features, which closely correlate to the aspect of harmony, are a well-established tool in processing and analyzing music data. There are many ways of computing and enhancing chroma features, which results in a large number of chroma variants with different properties. In this paper, we present a chroma toolbox [13], which contains MATLAB implementations for extracting various types of recently proposed pitch-based and chroma-based audio features. Providing the MATLAB implementations on a well-documented website under a GNU-GPL license, our aim is to foster research in music information retrieval. As another goal, we want to raise awareness that there is no single chroma variant that works best in all applications. To this end, we discuss two example applications showing that the final music analysis result may crucially depend on the initial feature design step.

Recordings for acoustic research should ideally be made in a lossless format. However, in some cases pre-existing data may be available in a lossy format such as mp3, prompting the question in how far this compromises the accuracy of acoustic measurements. In order to determine whether this is the case, we compressed 10 recordings of read speech in different compression rates (16-320 kbps), and reconverted them to wav in order to examine the effect of compression on commonly used suprasegmental measures of fundamental frequency (f0), pitch range and level. Results suggest that at compression rates between 56 and 320 kbps, measures of f0 and most measures of pitch range and level remain reliable, with mean errors below 2% and often better than that. The skewness of the distribution of f0 measurements , however, shows much greater measurement errors, with mean errors of 6.9%-7.6% at compression rates between 96 kbps and 320 kbps, and 44.8% at 16 kbps. We conclude that mp3 compressed recordings can be subjected to the acoustic measurements tested here. Nevertheless, the indeterminacy added by mp3 compression needs to be taken into account when interpreting measurements.

El presente informe se centra en el desarrollo e implementación de un equipo electrónico, basado en DSP (Digital Signal Processing), que permite generar sonido musical en forma interactiva (tiempo real) a través del recientemente difundido método de Síntesis por Modelado Físico. Esta técnica consiste en obtener un modelo físico-matemático de los elementos principales del instrumento (excitador, resonador y radiador) y de su interacción mutua, describiéndolos a partir de ecuaciones diferenciales y otros elementos matemáticos. Conseguido el modelo físico, se implementa un algoritmo de resolución por métodos numéricos que produce el sonido final por simulación en tiempo real, a partir de un conjunto de parámetros que definen el instrumento (propiedades geométricas y de los materiales, propiedades dinámicas, coeficientes de roce viscoso, etc.), y de señales de control generadas por el intérprete. Dichas señales ingresan al modelo en forma de acciones mecánicas que modifican el estado dinámico de determinadas partes del mismo.

La cathédrale de Noyon en Picardie, possède sous le maître-autel, une salle souterraine de petite dimension (15 m²) dont les murs nord, ouest et sud sont couverts de soixante-quatre poteries dites « acoustiques ». Cet espace, nommé caveau phonocamptique par les érudits du XIX e siècle, communiquait avec la cathédrale par un conduit étroit de section carrée, autrefois ouvert au centre du choeur des chantres. Nous avons mesuré les dimensions géométriques de l'ensemble des pots ainsi que leurs fréquences de résonances. Il apparaît que deux types de poteries ont été insérés, avec deux fréquences accordées à la quarte (environ 108 Hz et 141 Hz en moyenne) et réparties dans les murs selon un schéma présentant une certaine régularité. Les fréquences de résonance, ainsi que leur accord, sont assez caractéristiques de ce qu'on l'on observe par ailleurs en France [1,2]. Il semble donc que l'objectif de ces pots placés dans le caveau, était bien acoustique au sens large. Cependant, contrairement à l'interprétation acoustique habituelle, il est fort peu probable que ces poteries aient été insérées pour des questions d'absorption dans cet espace, dont la finalité reste encore inconnue. En outre, l'étude des sources écrites étudiées jusqu'alors ne permet pas d'établir que ce caveau faisait l'objet de cérémonies particulières. Les rares textes concernant d'autres édifices ne mentionnent que la possibilité d'amplifier la voix, ce qui, nous le savons aujourd'hui, n'est pas possible lorsque les pots sont placés en haut des murs ou dans les voûtes. La question à Noyon, est de savoir si les concepteurs de ce caveau ont cherché à rapprocher les pots du choeur des chantres en les plaçant dans le caveau, afin d'amplifier leur voix par cet artifice.

In this paper, we propose a new method suitable for the automatic analysis of microtiming played by drummers in jazz recordings. Specifically, we aim to estimate the drum-mers' swing ratio in excerpts of jazz recordings taken from the Weimar Jazz Database. A first approach is based on automatic detection of ride cymbal (RC) onsets and evaluation of relative time intervals between them. However, small errors in the onset detection propagate considerably into the swing ratio estimates. As our main technical contribution , we propose to use the log-lag autocorrelation function (LLACF) as a mid-level representation for estimating swing ratios, circumventing the error-prone detection of RC onsets. In our experiments, the LLACF-based swing ratio estimates prove to be more reliable than the ones based on RC onset detection. Therefore, the LLACF seems to be the method of choice to process large amounts of jazz recordings. Finally, we indicate some implications of our method for microtiming studies in jazz research.

This work investigates a novel approach for harmonic content estimation, which is based on quantized Fourier series. This algorithm is a suitable candidate for real time applications involving a massive computation of harmonic content of signals, especially due to its low computational complexity. In particular, this new tool can be useful for analysing musical sounds.

Recently, Musical Notes separation of Musical signals source has been a focused research topic in digital signal processing particularly, in Music signal processing. The proposed Non-Negative decomposition method for the separation of musical notes from input recorded .wav file. The method is easy to execute, mathematically efficient. The proposed methodology contains the input .wav file is first transformed into time-frequency representation which is responsible for the generation of spectrogram which includes generation of windows with the help of sampling techniques. Afterwards, transformation of spectrogram into matrix form containing all non-negative elements, that why it is termed non-negative decomposition. Then by means of filtering or clustering technique it is possible to separate notes of musical sources and to recover the musical notes of respective musical source. Musical Notes separation has a lot of applications including speech recognition, Music up-mixing ,music transcription, remixing and content creation, karaoke systems ,Audio post-production and re-mastering, Spatial audio and upmixing, Denoising (Separate noise speech , tools for remixing for disc jockeys or producers, (Remove background music from music, Remove bleed from other instruments) , instrument-wise equalizing and preprocessing in music analysis tasks.

Many different musical applications, including remote soni-fication, sound installation, augmented reality, and distributed/telematic music performance, make use of high speed Internet connections between different performance spaces. Most of the technical literature on this subject focuses on system latency, but there are also significant contributions from the acoustics of all rooms connected: specifically, smaller auxiliary rooms will tend to introduce spectral coloration, and the "main" larger volume will send more reverberation to the off-site performers. Measurements taken in two linked networked sites used in telematic performance show that both of these issues are present. Some improvements are suggested, including physical room alterations and equalization methods using signal processing.

This paper describes preliminary research results on enhancement of loudspeaker impulse response measurement in small room. Proposed enhancement is a combination of MLS (Maximum Length Sequence) algorithm and Delay-Sum Beamforming in order to measure frequency response of loudspeakers even in very small and poorly suppressed rooms. Application of the beamforming algorithm allows the extension measure of impulse response by damping reflections from walls, floor and ceiling.

The digital revolution has brought about a massive increase in the availability and distribution of music-related documents of various modalities comprising textual, audio, as well as visual material. Therefore, the development of techniques and tools for organizing, structuring, retrieving, navigating, and presenting music-related data has become a major strand of research—the field is often referred to as Music Information Retrieval (MIR). Major challenges arise because of the richness and diversity of music in form and content leading to novel and exciting research problems. In this article, we give an overview of new developments in the MIR field with a focus on content-based music analysis tasks including audio retrieval, music synchronization, structure analysis, and performance analysis.

This thesis presents an interdisciplinary research on reed woodwind instruments and human voice, focusing on acoustical, physiological and perceptual aspects of sound generation.
The wind instruments studies concentrate on breathing and blowing under realistic conditions and provide a deeper insight on required aerodynamical input parameters. The variation of blowing pressure with loudness and fundamental frequency was measured in professional players of oboe, bassoon, clarinet, and alto saxophone and was found to be quite systematic, though differing between the instruments. Airflow for sustained tones was measured by indirect spirometry, together with blowing pressure and sound pressure level, using extreme reeds, one soft and one hard. Recordings were made in an ordinary room as well as in a calibrated reverberant chamber. Also, tones with an intense vibrato were analysed for the oboe, the saxophone and the bassoon. The results revealed wide variations in blowing pressure, suggesting that a rhythmic modulation of the contraction of expiratory muscles was a main factor, and relatively small variation in fundamental frequency. The players’ perception of self-produced static lung pressures typically used in performances was analysed in a psychophysical experiment, that revealed a quasi-linear relationship between perceived and produced pressures. The respiratory movements during playing were measured by a non-invasive technique, respiratory inductive pletysmography, which offered acceptably reliable data. The results revealed significant participation of the rib cage in all players and also of the abdominal wall in several players. Also, the impact of the continuous changes of O2 and CO2 gases in the pulmonary air exhaled during performance on the fundamental frequency was predicted from theory and compared with experimental data. The effect, smaller than that of temperature variation, still would represent a factor of potential relevance to wind instrument intonation. In addition, the sound production characteristics of a particular type of phonation, perceptually judged as similar to that used in Tibetan chant, were studied by high-speed imaging. Also, it was examined using acoustical and physiological methods. The results revealed a synchronised co-oscillation of the vocal and ventricular folds, which yields a lowering of fundamental frequency due to multiplication of the vocal fold period.

The development of automated methods for revealing the repetitive structure of a given music recording is of central importance in music information retrieval. In this paper , we present a novel scape plot representation that allows for visualizing repetitive structures of the entire music recording in a hierarchical, compact, and intuitive way. In a scape plot, each point corresponds to an audio segment identified by its center and length. As our main contribution , we assign to each point a color value so that two segment properties become apparent. Firstly, we use the lightness component of the color to indicate the repetitive-ness of the encoded segment, where we revert to a recently introduced fitness measure. Secondly, we use the hue component of the color to reveal the relations between different segments. To this end, we introduce a novel grouping procedure that automatically maps related segments to similar hue values. By discussing a number of popular and classical music examples, we illustrate the potential and visual appeal of our representation and also indicate limitations.

The Agent Tool is a complex composition and performance
environment that aords the scripting of abstract Agents
of varying complexity to control elements of synthesis and
sound manipulation. The micro-threaded architecture of
the system means that all scripts are written from the per-
spective of an Agent/performer. Consequently, complex
compositions, gestures and performances can be crafted in
a simple and ecient manner. It is designed to be an open-
ended framework whereby all data can be emitted via Open
Sound Control (OSC) for external processing if required.
User interaction with the system can come in a variety
of forms. These include, but are not limited to graphical
manipulation, scripting, real time video input and external
control via OSC. The system was initially designed as an
environment to allow dynamic and ecient graphic sound
synthesis through extraction of data from static imagery,
video playback or real time video input. The open scripting
engine consequently allows the system to perform direct au-
dication of image stimuli or conversely allow complex soni-
cations to take place. The Agent Tool is a cross-platform
package that runs on various Linux distributions, Mac OSX
and Windows operating systems. This paper seeks to dis-
cuss the Agent based functionality the system oers and
consequently the composition and interaction design that
the system aords.

Discrete-time modelling strategies of analogue Moog sawtooth oscillator waveforms are presented. Two alternative approaches suitable for real-time implementation are proposed, one modelling the analogue waveform in time-domain using phase distortion synthesis and another matching the spectrum of an existing antialiasing sawtooth oscillator to the corresponding analogue spectrum using a first-order IIR post-equalising filter. A parameter estimation procedure for both approaches is explained and performed. Performance evaluation using polynomial fits for the estimated parameters is carried out and good matches between the model outputs and recorded waveforms are obtained. The best match of the tested algorithms is produced by the phase distortion model and by post-equalising the fourth-order B-spline bandlimited step function sawtooth oscillator.

A novel approach to implement the distortion effect is introduced. The proposed approach is based on time-varying phase distortion of the input signal, and it is implemented using a coefficient modulated first-order allpass filter. This new technique provides control over the distorted band with a proper choice of the modulating signal. By choosing a modulating signal that applies the phase distortion only for low frequencies, the aliasing often generated by conventional distortion effects, which modify the signal amplitude, can be greatly reduced. Modulation signals that produce distortion effects applicable for electric guitar playing are also discussed. Sound examples on the use of the filter can be found at http://www.acoustics.hut.fi/~jpekonen/Papers/dafx08/.

Sheet music and audio recordings represent and describe music on different semantic levels. Sheet music describes abstract high-level parameters such as notes, keys, measures, or repeats in a visual form. Because of its explicitness and compactness, most musicologists discuss and analyze the meaning of music on the basis of sheet music. On the contrary, most people enjoy music by listening to audio recordings, which represent music in an acoustic form. In particular, the nuances and subtleties of musical performances, which are generally not written down in the score, make the music come alive. In this paper, we address the problem of bridging the gap between the sheet music domain and the audio domain. In particular, we discuss aspects on music representations, music synchronization, and optical music recognition, while indicating various strategies and open research problems.

The extraction of local tempo and pulse information from audio recordings constitutes a challenging task, in particular for music with significant tempo variations.  Furthermore, the existence of various pulse levels such as measure, tactus,  and  tatum  often  makes  the  determination  of  absolute tempo problematic. In this demo, we present the tempogram
toolbox, which contains MATLAB implementations for extracting various types of recently proposed tempo and pulse-related audio representations [2, 3].  These representations are particularly designed to reveal useful information even for  music  with  weak  note  onset  information  and  changing  tempo.  The  toolbox  is  provided  under  a  GNU-GPL license at

www.mpi-inf.mpg.de/resources/MIR/tempogramtoolbox

The general goal of music synchronization is to automatically align the multiple information sources such as audio recordings , MIDI files, or digitized sheet music related to a given musical work. In computing such alignments, one typically has to face a delicate tradeoff between robustness and accuracy. In this paper, we introduce novel audio features that combine the high temporal accuracy of onset features with the robustness of chroma features. We show how previous synchronization methods can be extended to make use of these new features. We report on experiments based on polyphonic Western music demonstrating the improvements of our proposed synchronization framework.

This paper will consider wave (amplitude) and phase signal shaping techniques for the digital emulation of distortion effect processing. We examine how to determine the Wave- and Phase-shaping functions with harmonic amplitude and phase data. Three distortion effects units are used to provide test data. The action of the Wave- and Phase-shaping functions derived for these effects is demonstrated with the assistance of a super-resolution frequency-domain analysis technique.

This paper describes a method to randomly generate vectors of symbol probabilities so that the corresponding discrete memoryless source has a prescribed entropy. One application is to Monte Carlo simulation of the performance of noiseless variable length source coding.