Paper Archive

In the digital simulation of non-linear audio effect circuits, the arising non-linear equation system generally poses the main challenge for a computationally cheap implementation. As the computational complexity grows super-linearly with the number of equations, it is beneficial to decompose the equation system into several smaller systems, if possible. In this paper we therefore develop an approach to determine such a decomposition automatically. We limit ourselves to cases where an exact decomposition is possible, however, and do not consider approximate decompositions.

Download

This paper proposes an image-guided HRTF selection procedure that exploits the relation between features of the pinna shape and HRTF notches. Using a 2D image of a subject’s pinna, the procedure selects from a database the HRTF set that best fits the anthropometry of that subject. The proposed procedure is designed to be quickly applied and easy to use for a user without previous knowledge on binaural audio technologies. The entire process is evaluated by means of an auditory model for sound localization in the mid-sagittal plane available from previous literature. Using virtual subjects from a HRTF database, a virtual experiment is implemented to assess the vertical localization performance of the database subjects when they are provided with HRTF sets selected by the proposed procedure. Results report a statistically significant improvement in predictions of localization performance for selected HRTFs compared to KEMAR HRTF which is a commercial standard in many binaural audio solutions; moreover, the proposed analysis provides useful indications to refine the perceptually-motivated metrics that guides the selection.

Download

This paper presents a comparison framework to study the relative benefits of the typical trapezoidal method with the lesser-used implicit midpoint method for the simulation of audio lumped systems at a fixed rate. We provide preliminary tools for understanding the behavior and error associated with each method in connection with typical analysis approaches. We also show implementation strategies for those methods, including how an implicit midpoint method solution can be generated from a trapezoidal method solution and vice versa. Finally, we present some empirical analysis of the behavior of each method for a simple diode clipper circuit and provide an approach to help interpret their relative performance and how to pick the more appropriate method depending on desirable properties. The presented tools are also intended as a general approach to interpret the performance of discretization approaches at large in the context of fixed-rate simulation.

Download

In this paper, we introduce a function designed specifically for sparse audio representations. A progression in the selection of dictionary elements (atoms) to sparsely represent audio has occurred: starting with symmetric atoms, then to damped sinusoid and hybrid atoms, and finally to the re-appropriation of the gammatone (GT) and formantwave-function (FOF) into atoms. These asymmetric atoms have already shown promise in sparse decomposition applications, where they prove to be highly correlated with natural sounds and musical audio, but since neither was originally designed for this application their utility remains limited. An in-depth comparison of each existing function was conducted based on application specific criteria. A directed design process was completed to create a new atom, the ramped exponentially damped sinusoid (REDS), that satisfies all desired properties: the REDS can adapt to a wide range of audio signal features and has good mathematical properties that enable efficient sparse decompositions and synthesis. Moreover, the REDS is proven to be approximately equal to the previous functions under some common conditions.

Download

In this paper, we present novel strategies for stationary/transient signal separation in audio signals in order to exploit the basic observation that stationary components are sparse in frequency and persistent over time whereas transients are sparse in time and persistent across frequency. We utilize a multi-resolution STFT approach which allows to define structured shrinkage operators to tune into the characteristic spectrotemporal shapes of the stationary and transient signal layers. Structure is incorporated by considering the energy of time-frequency neighbourhoods or modulation spectrum regions instead of individual STFT coefficients, and shrinkage operators are employed in a dual-layered Iterated Shrinkage/Thresholding Algorithm (ISTA) framework. We further propose a novel iterative scheme, Iterative Cross-Shrinkage (ICS). In experiments using artificial test signals, ICS clearly outperforms the dual-layered ISTA and yields particularly good results in conjunction with a dynamic update of the shrinkage thresholds. The application of the novel algorithms to recordings from acoustic musical instruments provides perceptually convincing separation of transients.

Download

Applause is the result of many individuals rhythmically clapping their hands. Applause recordings exhibit a certain temporal, timbral and spatial structure: claps originating from a distinct direction (i.e, from a particular person) usually have a similar timbre and occur in a quasi-periodic repetition. Traditional upmix approaches for blind mono-to-stereo upmix do not consider these properties and may therefore produce an output with suboptimal perceptual quality to be attributed to a lack of plausibility. In this paper, we propose a blind upmixing approach of applause-like signals which aims at preserving the natural structure of applause signals by incorporating periodicity and timbral similarity of claps into the upmix process and therefore supporting plausibility of the artificially generated spatial scene. The proposed upmix approach is evaluated by means of a subjective preference listening test.

Download

Research on perception of music production practices is mainly concerned with the emulation of sound engineering tasks through lab-based experiments and custom software, sometimes with unskilled subjects. This can improve the level of control, but the validity, transferability, and relevance of the results may suffer from this artificial context. This paper presents a dataset consisting of mixes gathered in a real-life, ecologically valid setting, and perceptual evaluation thereof, which can be used to expand knowledge on the mixing process. With 180 mixes including parameter settings, close to 5000 preference ratings and free-form descriptions, and a diverse range of contributors from five different countries, the data offers many opportunities for music production analysis, some of which are explored here. In particular, more experienced subjects were found to be more negative and more specific in their assessments of mixes, and to increasingly agree with each other.

Download

The same music composition can be performed in different ways, and the differences in performance aspects can strongly change the expression and character of the music. Experienced musicians tend to have their own performance style, which reflects their personality, attitudes and beliefs. In this paper, we present a datadriven analysis of the performance style of two master violinists, Jascha Heifetz and David Fyodorovich Oistrakh to find out their differences. Specifically, from 26 gramophone recordings of each of these two violinists, we compute features characterizing performance aspects including articulation, energy, and vibrato, and then compare their style in terms of the accents and legato groups of the music. Based on our findings, we propose algorithms to synthesize violin audio solo recordings of these two masters from scores, for music compositions that we either have or have not observed in the analysis stage. To our best knowledge, this study represents the first attempt that computationally analyzes and synthesizes the playing style of master violinists.

Download