We describe a novel Query-by-Example (QBE) approach in Music Information Retrieval, which allows a user to customize query examples by directly modifying the volume of different instrument parts. The underlying hypothesis is that the musical genre shifts (changes) in relation to the volume balance of different instruments. On the basis of this hypothesis, we aim to clarify the relationship between the change of the volume balance of a query and the shift in the musical genre of retrieved similar pieces, and thus help instruct a user in generating alternative queries without choosing other pieces. Our QBE system first separates all instrument parts from the audio signal of a piece with the help of its musical score, and then lets a user remix those parts to change acoustic features that represent musical mood of the piece. The distribution of those features is modeled by the Gaussian Mixture Model for each musical piece, and the Earth Movers Distance between mixtures of different pieces is used as the degree of their mood similarity. Experimental results showed that the shift was actually caused by the volume change of vocal, guitar, and drums.

Download

The suitability of optimal path finding methods for vocal melody extraction in polyphonic music is well recognized since they combine local pitch strength and temporal smoothness considerations in a global sense. However, when such single-F0 tracking systems are applied to sound mixtures in which pitched accompaniment is of comparable strength to the singing voice, they suffer from irrecoverable degradations. In this study we investigate the use of an optimal path finding method that is allowed to dynamically track multiple F0 paths, specifically two, through the F0 candidate space. It is shown that when such a system is applied to typical polyphonic mixtures with vocal solo the melodic information is indeed retrieved. Audio examples are available at http://www.ee.iitb.ac.in/daplab/DualF0TrackingResults_DAFx

Download

This paper investigates numerical simulation of a string coupled transversely to a resonant body. Starting from a complete nite difference formulation, a second model is derived in which the body is represented in modal form. The main advantage of this hybrid form is that the body model is scalable, i.e. the computational complexity can be adjusted to the available processing power. Numerical results are calculated and discussed for simplied models in the form of string-string coupling and string-plate coupling.

Download

This paper presents a new approach to modal synthesis for rendering sounds of virtual objects. We propose a generic method for modal analysis that preserves sound variety across the surface of an object, at different scales of resolution and for a variety of complex geometries. The technique performs automatic voxelization of a surface model and automatic tuning of the parameters of hexahedral finite elements, based on the distribution of material in each cell. The voxelization is performed using a sparse regular grid embedding of the object, which easily permits the construction of plausible lower resolution approximations of the modal model. With our approach, we can compute the audible impulse response of a variety of objects. Our solution is robust and can handle non-manifold geometries that include both volumetric and surface parts, such as those used in games, training simulations, and other interactive virtual environment.

Download

This extended abstract summarizes DAFx-related developments at CCRMA over the past year or so.

Download

This contribution combines techniques for sound synthesis and spatial reproduction for a joint synthesis of the sound production and sound propagation properties of virtual string instruments. The generated sound field is reproduced on a massive multichannel loudspeaker system using wave field synthesis techniques. From physical descriptions of string vibrations and sound waves by partial differential equations follows an algorithmic procedure for synthesis-driven wave field reproduction. Its processing steps are derived by mathematical analysis and signal processing principles. Three different building blocks are addressed: The simulation of string vibrations, a model for the radiation pattern of the generated acoustical waves, and the determination of the driving signals for the multichannel loudspeaker array. The proposed method allows the spatial reproduction of synthetic spatial sound without the need for pre-recorded or pre-synthesized source tracks.

Download

This paper presents a study on algorithms for headphone-free binaural synthesis using a dedicated loudspeaker configuration. Both algorithms under investigation improve the properties of the binaural synthesis performance of the array. Firstly, beam-forming provides sound radiation localized at two freely adjustable, narrow target spots. Adjusting both spots to the locations of the listener’s ears achieves a good basis. Secondly, an additional interaural crosstalk canceler improves the overall result.

Download

This paper addresses the problem of rendering a virtual source through loudspeaker arrays. The orientation of the virtual source and its aperture determine its radial beampattern. The methodology we present here imposes that the wavefield in a predetermined listening area best approximates the desired wavefield in the least squares sense. With respect to the traditional techniques the number of constraints is much higher than the number of loudspeakers. As a consequence, the loudspeaker coefficient vector is the solution of an over-determined equation system. Moreover this system may be ill-conditioned. In order to solve these issues, we resort to a least squares inversion combined with a Singular Value Decomposition (SVD) to attenuate the problem of ill-conditioning. Some experimental results show the feasibility and the issues of this methodology.

Download

A recently introduced structure to implement a continuously smooth spectral delay, based on a cascade of first-order allpass filters and an equalizing filter, is described and the properties of this spectral delay filter are reviewed. A new amplitude envelope equalizing filter for the spectral delay filter is proposed and the properties of structures utilizing feedback and/or time-varying filter coefficients are discussed. In addition, the stability conditions for the feedback and the time-varying structures are derived. A spectral delay filter can be used for synthesizing chirp-like sounds or for modifying the timbre of arbitrary audio signals. Sound examples on the use of the spectral delay filters utilizing the structures discussed in this paper can be found at http://www.acoustics.hut. fi/publications/papers/dafx09-sdf/.

Download

We present a real-time implementation of SMS synthesis in Pure Data. This instrument focuses on interaction with the ability to continuously synthesize any frame position within an SMS sound representation, in any order, thereby freeing time from other parameters such as frequency or spectral shape. The instrument can be controlled expressively with a Wacom Tablet that offers both coupled and absolute controls with good precision. A prototype graphical interface in python is presented that helps to interact with the SMS data through visualization. In this system, any sound sample with interesting spectral features turns into a playable instrument. The processing functionality originates in the SMS C code written almost 20 years ago, now re-factored into the open source library, libsms, also wrapped into a python module. A set of externals for Pure Data, called smspd, was made using this library to facilitate on-the-fly analysis, flexible modifications, and interactive synthesis. We discuss new transformations are introduced based on the possibilities of this system and ideas for higher-level, feature based transformations that benefit from the interactivity of this system.

Download