Download Perceptual Linear Filters: Low-Order ARMA Approximation for Sound Synthesis This paper deals with the approximation of a given frequency response by a low-order linear ARMA filter (Auto-Regressive Moving Average). The aim of this work is the audio synthesis, then to improve the perceptual quality, a criterion based on human listening is defined and minimized. Two complementary approaches are proposed here for solving this non-linear and non-convex problem: first, a weighted version of the Iterative Prefiltering, second, an adaptation of the Gauss-Newton method. This algorithm is adapted to guarantee the causality/stability of the obtained filter, and eventually its minimum phase property. The benefit of the new method is illustrated and evaluated.
Download Granular analysis/synthesis of percussive drilling sounds This paper deals with the automatic and robust analysis, and the realistic and low-cost synthesis of percussive drilling like sounds. The two contributions are: a non-supervised removal of quasistationary background noise based on the Non-negative Matrix Factorization, and a granular method for analysis/synthesis of this drilling sounds. These two points are appropriate to the acoustical properties of percussive drilling sounds, and can be extended to other sounds with similar characteristics. The context of this work is the training of operators of working machines using simulators. Additionally, an implementation is explained.
Download Flexible Real-Time Reverberation Synthesis With Accurate Parameter Control Reverberation is one of the most important effects used in audio
production. Although nowadays numerous real-time implementations of artificial reverberation algorithms are available, many of
them depend on a database of recorded or pre-synthesized room
impulse responses, which are convolved with the input signal. Implementations that use an algorithmic approach are more flexible
but do not let the users have full control over the produced sound,
allowing only a few selected parameters to be altered. The realtime implementation of an artificial reverberation synthesizer presented in this study introduces an audio plugin based on a feedback delay network (FDN), which lets the user have full and detailed insight into the produced reverb. It allows for control of
reverberation time in ten octave bands, simultaneously allowing
adjusting the feedback matrix type and delay-line lengths. The
proposed plugin explores various FDN setups, showing that the
lowest useful order for high-quality sound is 16, and that in the
case of a Householder matrix the implementation strongly affects
the resulting reverberation. Experimenting with delay lengths and
distribution demonstrates that choosing too wide or too narrow a
length range is disadvantageous to the synthesized sound quality.
The study also discusses CPU usage for different FDN orders and
plugin states.
Download Efficient Modeling and Synthesis of Bell-like Sounds This paper describes two different techniques that can be used to model and synthesize bell-like sounds. The first one is a sourcefilter model based on frequency-zooming ARMA (pole-zero) modeling techniques. The frequency-zooming approach is powerful also in modal analysis of bell sound behavior. The second technique is based on a digital waveguide with a single loop filter that is designed to generate inharmonic partials by including one or more second-order allpass sections in the loop filter, possibly augmented with one or a few parallel resonators. A small handbell with inharmonic partials was recorded and used as a target of modeling and synthesis. Sound examples are found in http://www.acoustics.hut.fi/demos/dafx02/.
Download Sound synthesis using an allpass filter chain with audio‐rate coefficient modulation This paper describes a sound synthesis technique that modulates the coefficients of allpass filter chains using audio-rate frequencies. It was found that modulating a single allpass filter section produces a feedback AM–like spectrum, and that its bandwidth is extended and further processed by non-sinusoidal FM when the sections are cascaded. The cascade length parameter provides dynamic bandwidth control to prevent upper range aliasing artifacts, and the amount of spectral content within that band can be controlled using a modulation index parameter. The technique is capable of synthesizing rich and evolving timbres, including those resembling classic virtual analog waveforms. It can also be used as an audio effect with pitch-tracked input sources. Software and sound examples are available at http://www.acoustics.hut.fi/publications/papers/dafx09-cm/
Download Barberpole Phasing and Flanging Illusions Various ways to implement infinitely rising or falling spectral notches, also known as the barberpole phaser and flanging illusions, are described and studied. The first method is inspired by the Shepard-Risset illusion, and is based on a series of several cascaded notch filters moving in frequency one octave apart from each other. The second method, called a synchronized dual flanger, realizes the desired effect in an innovative and economic way using two cascaded time-varying comb filters and cross-fading between them. The third method is based on the use of single-sideband modulation, also known as frequency shifting. The proposed techniques effectively reproduce the illusion of endlessly moving spectral notches, particularly at slow modulation speeds and for input signals with a rich frequency spectrum. These effects can be programmed in real time and implemented as part of a digital audio processing system.
Download Exposure Bias and State Matching in Recurrent Neural Network Virtual Analog Models Virtual analog (VA) modeling using neural networks (NNs) has
great potential for rapidly producing high-fidelity models. Recurrent neural networks (RNNs) are especially appealing for VA due
to their connection with discrete nodal analysis. Furthermore, VA
models based on NNs can be trained efficiently by directly exposing them to the circuit states in a gray-box fashion. However,
exposure to ground truth information during training can leave the
models susceptible to error accumulation in a free-running mode,
also known as “exposure bias” in machine learning literature. This
paper presents a unified framework for treating the previously
proposed state trajectory network (STN) and gated recurrent unit
(GRU) networks as special cases of discrete nodal analysis. We
propose a novel circuit state-matching mechanism for the GRU
and experimentally compare the previously mentioned networks
for their performance in state matching, during training, and in exposure bias, during inference. Experimental results from modeling
a diode clipper show that all the tested models exhibit some exposure bias, which can be mitigated by truncated backpropagation
through time. Furthermore, the proposed state matching mechanism improves the GRU modeling performance of an overdrive
pedal and a phaser pedal, especially in the presence of external
modulation, apparent in a phaser circuit.
Download Improved Reverberation Time Control for Feedback Delay Networks Artificial reverberation algorithms generally imitate the frequency-dependent decay of sound in a room quite inaccurately. Previous research suggests that a 5% error in the reverberation time (T60) can be audible. In this work, we propose to use an accurate graphic equalizer as the attenuation filter in a Feedback Delay Network reverberator. We use a modified octave graphic equalizer with a cascade structure and insert a high-shelf filter to control the gain at the high end of the audio range. One such equalizer is placed at the end of each delay line of the Feedback Delay Network. The gains of the equalizer are optimized using a new weighting function that acknowledges nonlinear error propagation from filter magnitude response to reverberation time values. Our experiments show that in real-world cases, the target T60 curve can be reproduced in a perceptually accurate manner at standard octave center frequencies. However, for an extreme test case in which the T60 varies dramatically between neighboring octave bands, the error still exceeds the limit of the just noticeable difference but is smaller than that obtained with previous methods. This work leads to more realistic artificial reverberation.
Download Modeling of the Carbon Microphone Nonlinearity for a Vintage Telephone Sound Effect The telephone sound effect is widely used in music, television and the film industry. This paper presents a digital model of the carbon microphone nonlinearity which can be used to produce a vintage telephone sound effect. The model is constructed based on measurements taken from a real carbon microphone. The proposed model is a modified version of the sandwich model previously used for nonlinear telephone handset modeling. Each distortion component can be modeled individually based on the desired features. The computational efficiency can be increased by lumping the spectral processing of the individual distortion components together. The model incorporates a filtered noise source to model the self-induced noise generated by the carbon microphones. The model has also an input level depended noise generator for additional sound quality degradation. The proposed model can be used in various ways in the digital modeling of the vintage telephone sound.
Download Interpolation of long gaps in audio signals using the warped Burg's method This paper addresses the reconstruction of missing samples in audio signals via model-based interpolation schemes. We demonstrate through examples that employing a frequency-warped version of Burg’s method is advantageous for interpolation of long duration signal gaps. Our experiments show that using frequencywarping to focus modeling on low frequencies allows reducing the order of the autoregressive models without degrading the quality of the reconstructed signal. Thus a better balance between qualitative performance and computational complexity can be achieved.