Balazs Bank

• PhD
• Professor (Associate) at Budapest University of Technology and Economics

Numerous signal processing applications are emerging on mobile computing systems. These applications are subject to responsiveness constraints for user interactivity and, at the same time, must be optimized for energy efficiency. Many current embedded devices are composed of low-power multicore processors that offer a good trade-off between computa...

It is well known that a digital filter transfer function can be converted between the direct form and parallel connections of elementary sections, typically second-order ("biquad") sections. The conversion from direct to parallel form is performed using a partial fraction expansion, which usually requires long division of polynomials when expanding...

This paper proposes a design strategy for MIMO equalization and crosstalk cancellation problems based on fixed-pole IIR filters. The fixed-pole design of parallel filters is generalized to the multichannel case and it is shown that the filter parameters can be estimated by the least-squares equations similarly to the common FIR equalizers, with the...

Discrete-time rational transfer functions are often converted to parallel second-order sections due to better numerical performance compared to direct form infinite impulse response (IIR) implementations. This is usually done by performing partial fraction expansion over the original transfer function. When the order of the numerator polynomial is...

As a result of their complexity and versatility, pianos are arguably one of the most important instruments in Western music. The size, weight, and price of grand pianos as well as their relatively simple control surface (i.e., the keyboard), have led to the development of digital counterparts that mimic the sound of acoustic pianos as closely as po...

Accurate design of a parallel graphic equalizer involves the construction of a complex target frequency response, which is obtained by interpolating smoothly between command gains and by assigning minimum-phase characteristics, followed by a least squares filter design. This work proposes two methods to simplify the computations involved in the des...

Spatial audio rendering techniques using head-related transfer functions (HRTFs) are currently used in many different contexts such as immersive teleconferencing systems, gaming, 3D audio reproduction, etc. Since all these applications usually involve real-time constraints, efficient processing structures for HRTF modeling and interpolation are nec...

The Weighted Least Squares algorithm (WLS) is applied to numerous optimization problems, but requires the use of high computational resources, especially when complex arithmetic is involved. This work aims to accelerate the resolution of a WLS problem by reducing the computational cost (relaying on BLAS/LAPACK routines) and the computational precis...

Most IIR filter design algorithms aim to fit a complex target response, that is, both magnitude and phase. When this is not possible due to low filter order usually a magnitude-only filter design is applied. The magnitude-priority method presented in this paper combines the two approaches in such a way that the filter follows both magnitude and pha...

In the audio signal processing field, multiple IIR filters are required in many applications. As an example, equalizing a Wave Field Synthesis system requires massive filter processing in real time. Graphics Processing Units (GPUs) are well known for their potential in highly parallel data processing. Up to now, the use of the GPUs for implementing...

Multipoint equalization is a useful procedure used to enlarge the zone to be equalized in sound reproduction systems by measuring the room impulse responses in multiple locations and deriving a prototype function capable to represent the real environment. This paper deals with the introduction of a novel prototype func- tion derived from the combin...

Different nonlinear models are exploited to model real-world devices. Among them, an effective technique is based on the combination of orthogonal nonlinear functions and frequency-domain adaptive filtering algorithm for nonlinear system identification. In this paper, first the independence of the model from the orthogonal basis is demonstrated by...

In real-world applications high-order IIR filters are often converted to series or parallel second-order sections to decrease the negative effects of coefficient truncation and roundoff noise. While series biquads are more common, the parallel structure is gaining more interest due to the possibility of full code parallelization. In addition, it is...

This paper presents a combined approach to loudspeaker/room response equalization based on simple in-room measurements. In the first step, the anechoic response of the loudspeaker, which mostly determines localization and timbre perception, is equalized with a low-order non-minimum-phase equalizer. This is actually done using the gated in-room resp...

The direct design of fixed-pole parallel second-order filters is a very effective way of obtaining equalizers having the desirable logarithmic frequency resolution for audio applications. The frequency resolution of the parallel filter design is controlled directly by the choice of pole frequencies, similarly to Kautz filters. This paper reviews an...

Compact spherical loudspeaker arrays are nowadays investigated as sources for directional measurements and audio playback. Their microphone-based calibration is not trivial as it requires an elaborate measurement setup of highly accurate geometry. This contribution shows how measurements of accurate geometric alignment were obtained to the authors'...

For rectangular rooms with symmetric loudspeaker arrangements, full room equalization can be achieved at low frequencies, as demonstrated by previous research. The method is based on generating a plane wave that propagates along the room. However, often the room is not rectangular, and/or a symmetric loudspeaker setup cannot be assured, leading to...

String and membrane vibrations cannot be considered as linear above a certain amplitude due to the variation in string or membrane tension. A relevant special case is when the tension is spatially constant and varies in time only in dependence of the overall string length or membrane surface. The most apparent perceptual effect of this tension modu...

In audio, often specialized filter design methods are used that take into account the logarithmic frequency resolution of hearing. A notable side-effect of these quasi-logarithmic frequency design methods is a high-frequency attenuation for non-minimumphase targets due to the frequency-dependent windowing effect of the filter design. This paper pre...

The use of second-order parallel filters with pre-defined pole locations has been recently proposed for equalization and transfer function modeling. This letter presents an improved method for obtaining the pole positions of the parallel filter. The steps of the new method are the following: first, the target frequency response is smoothed to the r...

Recently, the fixed-pole design of second-order parallel filters has been introduced to accomplish arbitrary (e.g., logarithmic) frequency resolution for transfer function modeling and equalization. The frequency resolution is set by the pole frequencies, and the resulting filter response corresponds to the smoothed (moving-average filtered) versio...

Many audio systems show some form of nonlinear behavior that has to be taken into account in modeling. For this, often a black-box model is identified, coming from the generality and simplicity of the approach. One such model is the polynomial Hammerstein model, which uses parallel branches that have a polynomial-type nonlinearity and a linear filt...

In traditional warped FIR and IIR filters, the frequency-warping profile is adjusted by a single free parameter, leading to a less flexible allocation of frequency resolution. As an example, it is not possible to achieve a truly logarithmic frequency resolution, which would be often desired in audio applications. In this paper a new approach is pre...

An exploratory experiment was carried out in which subjects with different musical skills were asked to play a digital piano keyboard, first by following a specific key sequence and style of execution, and then performing freely. Judgments of perceived sound quality were recorded in three different settings, including standard use of the digital pi...

This paper investigates the audibility of longitudinal components in piano string vibrations with listening tests. The recorded fortissimo sounds of two grand and one upright pianos have been resynthesized with and without longitudinal components and used in ABX type listening tests. Results suggest that the longitudinal components are audible up t...

This paper presents a real-time piano synthesizer where both the transverse and longitudinal motion of the string is modeled by modal synthesis, resulting in a coherent and highly parallel model structure. The paper applies recent developments in piano modeling and focuses on the issues related to practical implementation (e.g., numerical stability...

Above a certain amplitude, membrane vibration becomes nonlinear due to the variation of surface tension. This leads to audible pitch glides, that greatly contribute to the characteristic timbre of tom-tom drums of the classical drum set, and many other percussion instruments. Therefore, there is a strong motivation to take the tension modulation ef...

In physics-based sound synthesis, it is in general possible to incorporate a mechanical or acoustical admittanceimpedance in the form of a digital filter. Examples include modeling of the termination of a string or a tube. However, when digital filters are fitted to measured admittance or impedance data, care has to be taken that the resulting filt...

In audio, equalizer design should take into account the frequency resolution of the auditory system. In this paper, this is accomplished by the fixed-pole design of parallel second-order filters. The design process has two steps: first, the poles of the filter are set according to the desired frequency resolution. Then, the feedforward coefficients...

This paper 1 presents a direct design technique for par-allel second-order sections based on a perceptually mo-tivated logarithmic scale, with application to instrument body modeling. Traditional FIR and IIR design techniques work on a linear frequency scale, which is usually not optimal for audio applications. Warped filters and Kautz filters are...

Longitudinal vibration of piano strings greatly contributes to the distinctive character of low piano notes. In this paper a simplified modal model is developed, which describes the generation of phantom partials and longitudinal free modes jointly. The model is based on the simplification that the coupling from the transverse vibration to the long...

H-1117 Budapest, Magyar tudósok krt. 2., {bank,sujbert}@mit.bme.hu Sound synthesis algorithms modeling the linear behavior of strings are well developed. However, some musical instruments require the modeling of such nonlinear phenomena as the appearance of longitudinal string modes, phantom partials, or mode coupling and pitch glide due to tension...

In this paper signal-based and physics-based sound synthesis methods are described, with a particular emphasis on our own results achieved in the recent years. The applications of these methods are given for the case of organ, piano, and violin synthesis. The two techniques are compared based on these case studies, showing that in some cases the ph...

this paper a robust and efficient method is presented for restoration of nonlinearly distorted movie soundtracks. The method is based on the a priori knowledge of the shape of the nonlinear function, which is assumed to be a static nonlinearity. The original undistorted signal is modeled by a set of harmonically related sinusoids. This signal is le...

This paper reviews recent developments in physics-based synthesis of piano. The paper considers the main components of the instrument, that is, the hammer, the string, and the soundboard.Modeling techniques are discussed for each of these elements, together with implementation strategies. Attention is focused on numerical issues, and each implement...

This paper reviews recent developments in physics-based synthesis of piano. The paper considers the main components of the instrument, that is, the hammer, the string, and the soundboard. Modeling techniques are discussed for each of these elements, together with implementation strategies. Attention is focused on numerical issues, and each implemen...

A robust loss filter design method is presented for digital waveguide string models, which can be used with high filter orders. The method aims at minimizing the decay time error in partials of the synthetic tone. This is achieved by a new weighting function based on the first-order Taylor series approximation of the decay time errors. Smoothing of...

In this paper a novel method is presented for the physics-based sound synthesis of the piano, based on digital waveguides. The approach combines the advantages of the commuted synthesis technique and the methods using a nonlinear hammer model. The interaction force of the hammer-string contact is computed by an auxiliary digital waveguide connected...

In this paper a robust and efficient method is presented for restoration of nonlinearly distorted movie soundtracks. The method is based on the a priori knowledge of the shape of the nonlinear function, which is assumed to be a static nonlinearity. The original undistorted signal is modeled by a set of harmonically related sinusoids. This signal is...

Real-time sound synthesis by physical modeling requires accurate design of each model block, together with special care on efficiency, computability, and complexity issues. This paper reviews the case of the piano: implementation of the complete model is discussed, from the sound generation mechanism to radiation issues and coupling-pedal effects....

The presentwork is about the synthesis of piano sound based on the grounds of physical principles. For that, #rst the acoustical properties of the piano havetobe

The paper discusses a crucial part of the piano model, the hammer model – digital waveguide interaction. The discontinuity problem arising when feeding the interaction force into the digital wave-guide is investigated, and a solution for its avoidance is proposed. The stability problem of the hammer model is overcome by a novel multi-rate implement...

www.mit.bme.hu In the recent years, digital waveguide modeling of musical instruments has proven to be an e ective tool for sound synthesis purposes, but some practical questions still have remained unanswered. In this paper a new equivalent structure of the digital waveguide for string synthesis is presented. This structure can be used for highly...

Recently, the fixed-pole design of parallel second-order filters has been proposed to accomplish arbitrary fre-quency resolution similarly to Kautz filters, at 2/3 of their computational cost. This paper relates the parallel filter to the complex smoothing of transfer functions. Complex smoothing is a well-established method for limiting the freque...

In physics-based sound synthesis, it is generally possible to in-corporate a mechanical or acoustical immittance (admittance or impedance) in the form of a digital filter. Examples include mod-eling of the termination of a string or a tube. However, when dig-ital filters are fitted to measured immittance data, care has to be taken that the resultin...

This paper presents a sustain-pedal effect simulation algorithm for piano synthesis, by using parallel second-order filters. A ro-bust two-step filter design procedure, based on frequency-zooming ARMA modeling and least squares fit, is applied to calibrate the algorithm from impulse responses of the soundboard and the string register. The model tak...

Above a certain amplitude, the string vibration becomes nonlin-ear due to the variation of tension. An important special case is when the tension varies with time but spatially uniform along the string. The most important effect of this tension modulation is the exponential decay of the pitch (pitch glide). In the case of nonrigid string terminatio...

This study is motivated by the physical modeling of the longitu-dinal string vibrations in the piano. Informal listening tests show that the longitudinal vibrations play an important role in the at-tack of the sound, and are responsible for the metallic character of low notes. First, a simple mathematical model is developed for qualitative understa...

In this paper a novel modeling method is presented for beating and two-stage decay. Here, one digital waveg-uide is used for each note and some resonators are run in parallel to simulate the beating and two-stage decay of those partials, where these phenomena are most promi-nent. The resonator bank is implemented by using the multi-rate approach, r...

Distortion is a desirable effect for sound coloration in electric guitar amplifiers and effect processors. At high sound levels, par-ticularly at low frequencies, the loudspeakers used in classic style cabinets are also a source of distortion. This paper presents a case study of measurements and digital modeling of a typical guitar loudspeaker as a...

ABSTRACT In this paper a mixed-paradigm piano model is presented. The ma- jor development,is the ability of modeling,longitudinal string vi- brations. Longitudinal string motion is the reason for the metallic sound of low piano notes, therefore its modeling greatly improves the perceptual quality of synthesized piano sound. In this novel approach,t...