Download Real-time implementation of a source separation algorithm Source separation out of a mix of signals has been under development for many years with different approaches. We use timefrequency representations of two microphone signals to estimate the mixing parameters of the source signals. In order to evaluate the robustness of the algorithm under real-world conditions we built a real-time implementation, which is suitable to detect the sources, their mixing parameters and performs the source separation based on the mixing parameters. Our implementation only needs a few parameters and then works as a stand-alone solution with the opportunity to apply further post-processing or digital audio effects to the source signals.
Download Algorithm for the separation of harmonic sounds with time-frequency smoothness constraint A signal model is described which forces temporal and spectral smoothness of harmonic sounds. Smoothness refers to harmonic partials, the amplitudes of which are slowly-varying as a function of time and frequency. An algorithm is proposed for the estimation of the model parameters. The algorithm is utilized in a sound separation system, the robustness of which is increased by the smoothness constraints.
Download Simulation of sound source motion by time-frequency filtering The generalisation of conventional linear time invariant filter theory from one dimension to the two-dimensional timefrequency (TF) domain provides a powerful tool for the simulation of complex time variable systems. TF filtering is performed as convolution in the TF domain and is based on nonparametric modelling using direct convolution along the time or frequency axis. A method based on short-time Fourier analysis has been developed to produce non-stationary signals with desirable time and frequency characteristics. This method is faster than non-recursive realisations and yields a simple synthesis procedure. The application of filtering in the TF domain for the simulation of sound source motion is presented.
Download Chaotic signal synthesis with real-time control: solving differential equations in PD, MAX/MSP, and JMAX Chaotic signals are useful in two different levels in audio synthesis: as sound material or control structure. Patching languages such as Pd, Max/MSP, and jMAX provide easier mechanisms for generating chaotic structures at control level. We can generate deterministic chaotic signals either by finding numerical solutions to differential equations or by using first return maps. While generating the next sample, both of these methods require calculations with the knowledge of the previous sample. Most signal processing environments for computer music, such as Pd, Max/MSP, and jMAX, transfer audio data among their objects by vectors (blocks). In such environments, finding numerical solutions to differential equations or generating signals based on first return maps, will require writing external objects or setting the block-size to 1. Writing external objects can be time consuming and the real-time control of the calculations have to be embedded in the external object, which will require a recompilation for every change to the mechanism. Setting the block-size to 1 can make writing the patch cumbersome and sometimes very confusing. In this paper we shall present the fexpr∼ object, implemented for Pd, Max/MSP, and jMAX, which can be used for finding numerical solutions to differential equations by simply entering the difference equations as part of the object arguments. The object parameters can then be controlled in real-time using the host patching language. As examples, solutions to Lorenz Equations, Chua’s Oscillators, Duffing’s equation, and the use of first return maps will be presented using the fexpr∼ object.
Download Using nonlinear amplifier simulation in dynamic range controllers Amplifying devices where the gain is automatically controlled by the level of the input signal performs dynamics processing. Non-linear components simulating tube amplifiers can be used in these devices to make musical signal audibly dense [1]. This paper deals with the simulation of tube amplifiers using the power polynomial approximation of transfer characteristic and their use in dynamic range controllers. The influence of various non-linear amplifying devices simulating tube amplifiers on the output signal spectrum of dynamic effects is presented as well.
Download A hybrid approach to timbral consistency in a virtual instrument The aim of this work is to make an instrument that is timbrally consistent over pitch and loudness. This particular work is not attempting to reproduce an existing instrument’s timbre, but to produce a timbrally dynamic virtual instrument that can be designed by the user. In this paper there is a brief introduction to timbre and synthesis methods, followed by a proposal on how to make timbrally consistent virtual instruments out of given timbres.
Download Multichannel audio decorrelation for coding Within digital audio codification, the processing of multichannel signals has become one of the main fields of research. Current work on the subject look for effective ways to exploit the existing redundancy between the different channels in order to reduce the codification binary rate. This work studies the Karhunen-Loeve Transform (KLT) as a method of decorrelating multi-channel signals prior to coding. Results on codification via AAC are reported.
Download Real time spectral expansion for creative and remedial sound transformation In this paper we describe the implementation, use and applications of WaveThresh, a real time Fourier\wavelet spectral expander. Expansion and reverse-expansion of spectral components is offered. In order that analysis methods can be better adapted to the signal we offer a combined wavelet\Fourier mode. This mode separates sinusoids from the rest of the signal (residual) and applies Fourier analysis to the sinusoids and wavelet analysis to the residual.
Download FFT analysis as a creative tool in live performance This paper describes the use of real time spectral analysis to enhance the creative opportunities in improvised live electronics/instrumental performance. FFT analysis allows musicians to observe in performance a visual representation of the spectrum, displaying the spectral characteristics of audio resulting from performance activity and/or computer processing. These characteristics can then be explored during performance, assigning areas of special interest within the spectrum to parameters which in turn control (or at least influence) electronic processing. This creates an effective, easily manipulated but potentially highly complex performance environment, encouraging further interaction between improvising performers, and allowing subtle and complex links to emerge between the timbral features of actual music (result) and the act of performance (cause). We hope to increase awareness of the performance-specific potential of familiar analytical tools, of which FFT is one example, and their unfulfilled creative potential.
Download Recent developments in PWSYNTH PWSynth was originally a visual synthesis language situated in PatchWork. Recently our research team has started a complete rewrite of the system so that it can be adapted to our new programming environment called PWGL. In this paper we present the main differences of the old and new systems. These include switching from C to C++, efficiency issues, interface between PWGL and the synthesis engine, and a novel copy-synth-patch scheme.