Signal Processing Commons^™

Single Channel Vocal Separation Using Median Filtering And Factorisation Techniques, Derry Fitzgerald, Mikel Gainza

Articles

This paper deals with the problem of the extraction of vocals from single channel audio signals containing both vocals and other instruments, including both pitched instruments and percussion instruments. A novel median filtering-based approach for the extraction of vocal tracks is described, which is simple and efficient to implement. Further improvements in separation quality are then obtained by the application of tensor factorisation techniques to further extract residual instruments from the vocal mix. Finally, a novel use of non-negative partial matrix cofactorisation is demonstrated as a means of further improving separation quality. Here the original single channel mixture is partially …

Interactive Music Archive Access System, Martin Gallagher, Derry Fitzgerald, Dan Barry, Matt Cranitch, Eugene Coyle

Conference papers

The goal of the Interactive Music Archive Access System (IMAAS) project was to develop an interactive music archive access system which was capable of allowing an end-user to easily extract rhythmic, melodic and harmonic musical metadata descriptors from audio, and allow the user to interact with the archive contents in a manner not typically allowed in archive access systems. To this end, the IMAAS system incorporates a range of real-time interaction tools which allow the user to modify the retrieved audio in a number of ways including the ability to isolate individual instruments in stereo mixes, pitch and time-scale modification, …

Locating Tune Changes And Providing A Semantic Labelling Of Sets Of Irish Traditional Tunes, Cillian Kelly, Mikel Gainza, David Dorran, Eugene Coyle

Conference papers

An approach is presented which provides the tune change loactions within a set of Irish traditional turnes. Also provided are semantic labels for each part of each tune within the set. A set in Irish traditional music is a number of individual tunes played segue. Each of the tunes in the set are made up of structural segments called parts. Musical variation is a prominent characteristic of this genre. However, a certain set of notes known as "set accented tones" are considered impervious to musical variation. Chroma information is extracted at "set accented tone" locations within the music. The resulting …

Audio Thumbnail Generation Of Irish Traditional Music, Cillian Kelly, Mikel Gainza, David Dorran, Eugene Coyle

Conference papers

An approach is presented which generates an audio thumbnail of Irish traditional music. An audio thumbnail is consered to be the most representative segment of the music. For popular music, the chorus is considered to be an ideal audio thumbnail, however in Irish Traditional Music there is no chorus. An Irish Traditional tune consists of tow or mor short structural segments called parts. Parts are repeated to extend the tuen, and the tune itself is also repeated once or more in its entirety. To further extend a performance, tunes are concatenated to form a set of tuens. As a result, …

On The Appearance Of A Positive Real Pole In The Results Of Glottal Closed Phase Linear Prediction, Alan O'Cinneide, David Dorran, Mikel Gainza, Eugene Coyle

Conference papers

Often when performing glottal closed phase covariance linear prediction, a positive real pole can appear in the resulting filter transfer function. The commonly adopted approach is to discard this pole, as it does not fit with the usual model of the all-pole vocal tract filter. However, this real pole describes some aspect of the speech signal; this paper provides a novel perspective on its occurrence. This viewpoint has a useful implication to the speech community, especially from the perspective of fitting a glottal pulse to the inverse filtered signal, as the real pole describes the return phase of the glottal …

Towards A Method To Determine The Glottal Formant Parameters Of Voiced Speech Without Time-Domain Reference, Alan O'Cinneide, David Dorran, Mikel Gainza, Eugene Coyle

Conference papers

This paper presents an approach to estimate the glottal formant parameters of the voicing source in the frequency-domain. The method is based on a simplified pole-zero interpretion of the prevalent Liljencrants-Fant (LF) model of glottal flow, and gives approximations for a broad range of pulses shapes. An advantage of the method is that, unlike other methods, it does not rely on time-domain references.

Harmonic/Percussive Separation Using Median Filtering, Derry Fitzgerald

Conference papers

In this paper, we present a fast, simple and effective method to separate the harmonic and percussive parts of a monaural audio signal.The technique involves the use of median filtering on a spectrogram of the audio signal, with median filtering performed across successive frames to suppress percussive events and enhance harmonic components, while median filtering is also performed across frequency bins to enhance percussive events and supress harmonic components. The two resulting median filtered spectrograms are then used to generate masks which are then applied to the original spectrogram to separate the harmonic and percussive parts of the signal. We …

Diffusion And Fractional Diffusion Based Image Processing, Jonathan Blackledge

Conference papers

We consider the background to describing strong scattering in terms of diffusive processes based on the diffusion equation. Intermediate strength scattering is then considered in terms of a fractional diffusion equation which is studied using results from fractional calculus. This approach is justified in terms of the generalization of a random walk model with no statistical bias in the phase to a random walk that has a phase bias and is thus, only ‘partially’ or ‘fractionally’ diffusive. A Green’s function solution to the fractional diffusion equation is studied and a result derived that provides a model for an incoherent image …

Inverse Scattering Solutions For Side-Band Signals, Jonathan Blackledge, Timo Hamalainen, Jyrki Joutsensalo

Conference papers

When a signal is recorded that has been physically generated by some scattering process (the interaction of electromagnetic waves with an inhomogeneous dielectric, for example), the `standard model' for the signal (i.e. information content convolved with a characteristic Impulse Response Function) is usually based on a single scattering approximation. An additive noise term is introduced into the model to take into account a range of non-deterministic factors including multiple scattering that, along with electronic noise and other background noise sources, is assumed to be relatively weak. Thus, the standard model is based on a `weak field condition' and the inverse …

Localization Quality Assessment In Source Separation-Based Upmixing Algorithms, Dan Barry, Gavin Kearney

Conference papers

In this paper we explore the source localisation accuracy and perceived spatial distortion of a source separation based upmix algorithm for 2 to 5 channel conversion. Unlike traditional upmixing techniques, source separation based techniques allow individual sources to be separated from the mixture and repositioned independently within the surround sound field. Generally, spectral artefacts and source interference generated during the source separation process are masked when the upmixed sound field is presented in its entirety; however, this can lead to perceived spatial distortion and ambiguous source localisation. Here, we use subjective testing to compare the localisation perceived on a purposely …

Automatic Musical Meter Detection, Mikel Gainza

Conference papers

A method that automatically estimates the metrical structure of a piece of music is presented. The approach is based on the generation of a beat similarity matrix, which provided information about the similarity between any two beats of a piece of music. The repetitive structure of most music is exploited by processing the beat similarity matrix in order to identify similar patterns of beats in different parts of a piece. This principles proves to be equally effective for the detection of both duple and triple meters as awll as complex meters. The use of beat positions and dynamic programming techniques …

Structural Segmentation Of Irish Traditional Music Using Chroma At Set Accented Tone Locations, Cillian Kelly, Mikel Gainza, David Dorran, Eugene Coyle

Conference papers

An approach is presented which provides a structural segmentation of Irish Traditional Music. Chroma information is extracted at certain locations within the music. The resulting chroma vectors are compared to determine similar structural segments. Chroma is only calculated at "set accented tone" locatins within the music. "Set accented tones" are considered to be impervious to melodic variation and are entirely representative of an Irish Traditional tune. Results show that comparing "set accented tones" represented by chroma significantly increases the structural segmentation accuracy that when "set accented tones" are represented by pitch values.

Information Hiding By Stochastic Disfusion And Its Application To Printed Document Authentication, Jonathan Blackledge, Eugene Coyle

Conference papers

The use of image based information exchange has grown rapidly over the years in terms of both e-to-e image storage and transmission and in terms of maintaining paper documents in electronic form. Further, with the dramatic improvements in the quality of COTS (Commercial-O-The-Shelf) printing and scanning devices, the ability to counterfeit electronic and printed documents has become a widespread problem. Consequently, there has been an increasing demand to develop digital watermarking, information hiding and covert encryption methods which can be applied to both electronic and printed images (and documents) for the purposes of authentication, prevent unauthorized copying and, in the …

A Mobile Ecg Monitoring System With Context Collection, Jin Peng Li, Damon Berry, Richard Hayes

Conference Papers

Preventative health management represents a shift from the traditional approach of reactive treatment-based healthcare towards a proactive wellness-management approach where patients are encouraged to stay healthy with expert support when they need it, at any location and any time. This work represents a step along the road towards proactive, preventative healthcare for cardiac patients. It seeks to develop a smart mobile ECG monitoring system that requests and records context information about what is happening around the subject when an arrhythmia event occurs. Context information about the subject’s activities of daily living will, it is hoped, provide an enriched data set …

Inverse Scattering Solutions With Applications To Electromagnetic Signal Processing, Jonathan Blackledge, Timo Hamalainen, Jyrki Joutsensalo

Articles

When a signal is recorded that has been physically generated by some scattering process (the interaction of electromagnetic, acoustic or elastic waves with inhomogeneous materials, for example), the ‘standard model’ for the signal (i.e. information content convolved with a characteristic Impulse Response Function) is usually based on a single scattering approximation. An additive noise term is introduced into the model to take into account a range of non-deterministic factors including multiple scattering that, along with electronic noise and other background noise sources, is assumed to be relatively weak. Thus, the standard model is based on a ‘weak field condition’ and …

On The Use Of The Beta Divergence For Musical Source Separation, Derry Fitzgerald, Matt Cranitch, Eugene Coyle

Conference papers

Non-negative Tensor Factorisation based methods have found use in the context of musical sound source separation. These techniques require the use of a suitable cost function to determine the optimal factorisation, and most work has focused on the use of the generalised Kullback-Liebler divergence, and more recently the Itakura-Saito divergence. These divergences can be regarded as limiting cases of the parameterised Beta divergence. This paper looks at the use of the Beta Divergence in the context of musical source separation with a view to determining an optimal value of Beta for this problem. This is considered for both magnitude and …

Musical Sound Source Separation Using Extended Tensor Decompositions, Derry Fitzgerald

Conference papers

Recently, tensor decompositions have found use in sound source separation. In particular, non-negative tensor decompositions have received a lot of attention due to their ability to decompose audio spectrograms into meaningful ”parts” such as individual notes. Extensions to the basic non-negative tensor factorisation framework allow the incorporation of additional constraints, such as shift-invariance in both frequency and time. This enables the factorisations to capture more complex structures than individual notes, such as individual sources playing different pitches and time-evolving instrument timbres. Further music specific constraints such as harmonicity and source-filter modeling have been shown to improve separation performance for musical …

Using Tensor Factorisation Models To Separate Drums From Polyphonic Music, Derry Fitzgerald, Matt Cranitch, Eugene Coyle

Conference papers

This paper describes the use of Non-negative Tensor Factorisation models for the separation of drums from polyphonic audio. Improved separation of the drums is achieved through the incorporation of Gamma Chain priors into the Non-negative Tensor Factorisation framework. In contrast to many previous approaches, the method used in this paper requires little or no pre-training or use of drum templates. The utility of the technique is shown on real-world audio examples.

Imaging Reconstruction For Light Scattering From A Tenuous Random Medium, Jonathan Blackledge

Conference papers

We consider the basis for describing strong scattering in terms of diffusive processes based on the diffusion equation. Intermediate strength scattering is then considered in terms of a fractional diffusion equation which is studied using results from fractional calculus. This approach is justified in terms of the generalization of a random walk model with no statistical bias in the phase to a random walk that has a phase bias and is thus, only `partially' or `fractionally' diffusive. A Green's function solution to the fractional diffusion equation is studied and a result derived that provides a model for an incoherent image …

Hand-Held Flyback Driven Coaxial Dielectric Barrier Discharge: Development And Characterization, Victor J. Law, Vladimir Milosavljevic, Neil O’Connor, James F. Lalor, Steven Daniels

Articles

The development of a handheld single and triple chamber atmospheric pressure coaxial dielectric barrier discharge driven by Flyback circuitry for helium and argon discharges is described. The Flyback uses external metal-oxide-semiconductor field-effect transistor power switching technology and the transformer operates in the continuous current mode to convert a continuous dc power of 10–33 W to generate a 1.2–1.6 kV 3.5 μs pulse. An argon discharge breakdown voltage of ∼768 V is measured. With a 50 kHz, pulse repetition rate and an argon flow rate of 0.5–10 argon slm (slm denotes standard liters per minute), the electrical power density deposited in …

A Real-Time Framework For Video Time And Pitch Scale Modification, Ivan Damnjanovic, Dan Barry, David Dorran, Josh Reiss

Conference papers

A framework is presented which addresses the issues related to the real-time implementation of synchronised video and audio time-scale and pitch-scale modification algorithms. It allows for seamless real-time transition between continually varying, independent time-scale and pitch-scale parameters arising as a result of manual or automatic intervention. We illuminate the problems which arise in a real-time context as well as provide novel solutions to prevent artefacts, minimise latency, and improve synchronisation. The time and pitch scaling approach is based on a modified phase vocoder with optional phase locking and an integrated transient detector which enables high quality transient preservation in real-time. …

Real-Time Plasma Controlled Chemistry In A Two-Frequency, Confined Plasma Etcher, Vladimir Milosavljevic, Albert R. Ellingboe, Cezar Gaman, John V. Ringwood

Articles

The physics issues of developing model-based control of plasma etching are presented. A novel methodology for incorporating real-time model-based control of plasma processing systems is developed. The methodology is developed for control of two dependent variables (ion flux and chemical densities) by two independent controls (27 MHz power and O2flow). A phenomenological physics model of the nonlinear coupling between the independent controls and the dependent variables of the plasma is presented. By using a design of experiment, the functional dependencies of the response surface are determined. In conjunction with the physical model, the dependencies are used to deconvolve the sensor …

The Annotation Of Traditional Irish Dance Music Using Matt2 And Tansey, Bryan Duggan, Brendan O'Shea, Mikel Gainza, Padraig Cunningham

Conference papers

Current estimates put the canon of traditional Irish dance tunes at least 7,000 compositions. Given this diversity, a common problem faced by musicians and ethnomusicologists is identifying tunes from recordings. This is evident even in the number of commercial recordings whose title is gan aimn (without name). This work attempts to solve this problem by developing a Content Based Music Information Retrieval (CBMIR) System adapted to the characteristics of traditional Irish music. A system is presented called MATT2 (Machine Annotation of Traditional Tunes) whose primary goal is to annotate recordings of traditional Irish dance music with useful meta-data including tune …

Linear Prediction: The Problem, Its Solution And Application To Speech, Alan O'Cinneide, David Dorran, Mikel Gainza

Conference papers

Linear prediction is a signal processing technique that is used extensively in the analysis of speech signals and, as it is so heavily referred to in speech processing literature, a certain level of familiarity with the topic is typically required by all speech processing engineers. This paper aims to provide a well-rounded introduction to linear prediction, and so doing, facilitate the understanding of the technique. Linear prediction and its mathematical derivation will be described, with a specific focus on applying the technique to speech signals. It is noted, however, that although progress in linear prediction has been driven primarily by …

A Surface Inspection Machine Vision System That Includes Fractal Texture Analysis, Jonathan Blackledge, Dmitry Dubovitskiy

Articles

The detection, recognition and classification of features in a digital image is an important component of quality control systems in production and process engineering and industrial systems monitoring, in general. In this paper, a new pattern recognition system is presented that has been designed for the specific task of monitoring the quality of sheet-steel production in a rolling mill. The system is based on using both the Euclidean and Fractal geometric properties of an imaged object to develop training data that is used in conjunction with a supervised learning procedure based on the application of a fuzzy inference engine. Thus, …

Object Detection And Classification With Applications To Skin Cancer Screening, Jonathan Blackledge, Dmitryi Dubovitskiy

Articles

This paper discusses a new approach to the processes of object detection, recognition and classification in a digital image. The classification method is based on the application of a set of features which include fractal parameters such as the Lacunarity and Fractal Dimension. Thus, the approach used, incorporates the characterisation of an object in terms of its texture.

The principal issues associated with object recognition are presented which includes two novel fast segmentation algorithms for which C++ code is provided. The self-learning procedure for designing a decision making engine using fuzzy logic and membership function theory is also presented and …

Drop Photometer Design., Kevin Tiernan

Doctoral

The tensiograph drop analyser, operating in the mode of ultraviolet-visual absorption photometer, was re-engineered to improve measurement accuracy. As well as instrumentation re-design, the other objective of the research was to analyse tensiograph signals to explore how their features are defined by drop properties. Using fibre optic technology, the tensiograph emits a light beam into a growing pendant drop whereupon it undergoes a number of internal reflections. A portion of the reflected light energy is coupled to a photodiode producing a proportional current that is electronically processed and transferred to a computer for analysis. The signal produced, known as the …

Analysis Of Financial Data Using Non-Negative Matrix Factorization, Ruairí De Fréin, Konstantinos Drakakis, Scott Rickard, Andrzej Cichocki

Articles

We apply Non-negative Matrix Factorization (NMF) to the prob-lem of identifying underlying trends in stock market data. NMF is arecent and very successful tool for data analysis including image andaudio processing; we use it here to decompose a mixture a data, thedaily closing prices of the 30 stocks which make up the Dow Jones In-dustrial Average, into its constitute parts, the underlying trends which govern the financial marketplace. We demonstrate how to impose ap-propriate sparsity and smoothness constraints on the components of thedecomposition. Also, we describe how the method clusters stocks to-gether in performance-based groupings which can be used for …

Portfolio Diversification Using Subspace Factorizations, Ruairí De Fréin, Konstantinos Drakakis, Scott Rickard

Conference papers

Successful investment management relies on allocating assets so as to beat the stock market. Asset classes are affected by different market dynamics or latent trends. These interactions are crucial to the successful allocation of monies. The seminal work on portfolio management by Markowitz prompts the adroit investment manager to consider the correlation between the assets in his portfolio and to vary his selection so as to optimize his riskreturn profile. The factor model, a popular model for the return generating process has been used for portfolio construction and assumes that there is a low rank representation of the stocks. In …

Structural Segmentation Using Set Accented Tones, Cillian Kelly, Mikel Gainza, David Dorran, Eugene Coyle

Conference papers

An approach which efficiently segments Irish Traditional Music into its constituent structural segments is presented. The complexity of the segmentation process is greatly increased due to melodic variation existent within this music type. In order to deal with these variations, a novel method using ‘set accented tones’ is introduced. The premise is that these tones are less susceptible to variation than all other tones. Thus, the location of the accented tones is estimated and pitch information is extracted at these specific locations. Following this, a vector containing the pitch values is used to extract similar patterns using heuristics specific to …