Download Evaluating Neural Networks Architectures for Spring Reverb Modelling
Reverberation is a key element in spatial audio perception, historically achieved with the use of analogue devices, such as plate and spring reverb, and in the last decades with digital signal processing techniques that have allowed different approaches for Virtual Analogue Modelling (VAM). The electromechanical functioning of the spring reverb makes it a nonlinear system that is difficult to fully emulate in the digital domain with white-box modelling techniques. In this study, we compare five different neural network architectures, including convolutional and recurrent models, to assess their effectiveness in replicating the characteristics of this audio effect. The evaluation is conducted on two datasets at sampling rates of 16 kHz and 48 kHz. This paper specifically focuses on neural audio architectures that offer parametric control, aiming to advance the boundaries of current black-box modelling techniques in the domain of spring reverberation.
Download Feature design for the classification of audio effect units by input/output measurements
Virtual analog modeling is an important field of digital audio signal processing. It allows to recreate the tonal characteristics of real-world sound sources or to impress the specific sound of a certain analog device upon a digital signal on a software basis. Automatic virtual analog modeling using black-box system identification based on input/output (I/O) measurements is an emerging approach, which can be greatly enhanced by specific pre-processing methods suggesting the best-fitting model to be optimized in the actual identification process. In this work, several features based on specific test signals are presented allowing to categorize instrument effect units into classes of effects, like distortion, compression, modulations and similar categories. The categorization of analog effect units is especially challenging due to the wide variety of these effects. For each device, I/O measurements are performed and a set of features is calculated to allow the classification. The features are computed for several effect units to evaluate their applicability using a basic classifier based on pattern matching.
Download Monophonic Pitch Detection by Evaluation of Individually Parameterized Phase Locked Loops
This paper describes a new efficient and sample based monophonic pitch tracking approach using multiple phase locked loops (PLLs). Hereby, distinct subband signals traverse pairs of individually parameterized PLLs. Based on the relation of the instantaneous pitch sample of respective PLLs to one another, relevant features per pitch candidate are derived. These features are combined into pitch candidate scores. Pitch candidates which exhibit the maximum score per sampling instance and exceed a voicing threshold, contribute to the overall pitch track. Evaluations with up to date datasets show that the tracking performance, compared to implementations which use only one PLL has significantly improved and nearly approaches the scores of a state of the art monophonic pitch tracker.
Download Separation of musical notes with highly overlapping partials using phase and temporal constrained complex matric factorization
In note separation of polyphonic music, how to separate the overlapping partials is an important and difficult problem. Fifths and octaves, as the most challenging ones, are, however, usually seen in many cases. Non-negative matrix factorization (NMF) employs the constraints of energy and harmonic ratio to tackle this problem. Recently, complex matrix factorization (CMF) is proposed by combining the phase information in source separation problem. However, temporal magnitude modulation is still serious in the situation of fifths and octaves, when CMF is applied. In this work, we investigate the temporal smoothness model based on CMF approach. The temporal ac-tivation coefficient of a preceding note is constrained when the succeeding notes appear. Compare to the unconstraint CMF, the magnitude modulation are greatly reduced in our computer simulation. Performance indices including sourceto-interference ratio (SIR), source-to-artifacts ratio (SAR), sourceto-distortion ratio (SDR), as well as modulation error ratio (MER) are given.
Download Categories of Perception for Vibrato, Flange, and Stereo Chorus: Mapping Out the Musically Useful Ranges of Modulation Rate and Depth for Delay-Based Effects
Vibrato, Flange, and Stereo Chorus are perhaps the three most often used digital audio effects that are created by smoothly modulating the duration of a delay line at typically sub-audio rates. Common practice is to use a periodic or quasi-periodic modulation control signal with frequency roughly between 2 and 9 Hz, and both the rate and depth of delay modulation are typically adjusted according to the aesthetic criteria of a performer or by an audio production engineer. In order to establish norms for the musically useful range of modulation rate and depth for such delay-based effects, 25 listeners were asked to make categorical judgments regarding their perception of vibrato, flange, and stereo chorus effects. The results map out for these two modulation parameters three perceptual regions for these three related effects: the region in which modulation is too subtle for effective use, the parameter ranges that seem most musically useful, and the region in which it is too extreme for most musical applications. Of particular interest is the observed commonality between these perceptual regions for vibrato, flange, and stereo chorus effects.
Download Time-Domain Chroma Extraction
In this paper, a novel chroma extraction technique called TimeDomain Chroma Extraction (TDCE) is introduced. In comparison to many other known schemes, the calculation of a time-frequency representation is unnecessary since the TDCE is a pure sample-bysample technique. It mainly consists of a pitch tracking module that is implemented with a phase-locked loop (PLL). A set of 24 bandpass filters over two octaves is designed with the F 0 output of the pitch tracker to estimate a chroma vector. To verify the performance of the TDCE, a simple chord recognition algorithm is applied to the chroma output. The experimental results show that this novel time-domain chroma extraction technique yields good results while requiring only minor complexity and thus, enables the extraction of musical features in real-time on low-cost DSP platforms.
Download A Csound Opcode for a Triode Stage of a Vacuum Tube Amplifier
The Csound audio programming language adheres to the inputoutput paradigm and provides a large number of specialized commands (called opcodes) for processing output signals from input signals. Therefore it is not directly suitable for component modeling of analog circuitry. This contribution describes an attempt to virtual analog modeling and presents a Csound opcode for a triode stage of a vacuum tube amplifier. Externally it communicates with other opcodes via input and output signals at the sample rate. Internally it uses an established wave digital filter model of a standard triode. The opcode is available as library module.
Download Audio Nonlinear Modeling through Hyperbolic Tangent Functionals
In the present contribution we present the preliminary results of a black box nonlinear system (NLS) modeling. The NLS is composed by a nonlinear sigmoid-type input-output relationship (NLTF) followed by a linear system (LTI), as in a Hammerstein nonlinear system. Here, the used NLTF is derived from a deformation of the Hyperbolic Tangent power expansion. The advantage of using the hyperbolic tangent function is that nonlinearity depends on the linear and cubic terms that measure curvature (and thus nonlinearity) of the transfer function. The hyperbolic tangent model is extended to other types of nonlinear systems by expanding the nonlinear system in linear and increasingly nonlinear contributions, where the expansion parameters are deformed to enhance or suppress specific nonlinear modes of the expansion. Simulations were performed using Matlab 2012a. The preliminary results show fairly good agreement between the system obtained by parametric inference and a reference system, with mean square error (MSE)=0.035.
Download Balancing Error and Latency of Black-Box Models for Audio Effects Using Hardware-Aware Neural Architecture Search
In this paper, we address automating and systematizing the process of finding black-box models for virtual analogue audio effects with an optimal balance between error and latency. We introduce a multi-objective optimization approach based on hardware-aware neural architecture search which allows specifying the optimization balance of model error and latency according to the requirements of the application. By using a regularized evolutionary algorithm, it is able to navigate through a huge search space systematically. Additionally, we propose a search space for modelling non-linear dynamic audio effects consisting of over 41 trillion different WaveNet-style architectures. We evaluate its performance and usefulness by yielding highly effective architectures, either up to 18× faster or with a test loss of up to 56% less than the best performing models of the related work, while still showing a favourable trade-off. We can conclude that hardware-aware neural architecture search is a valuable tool that can help researchers and engineers developing virtual analogue models by automating the architecture design and saving time by avoiding manual search and evaluation through trial-and-error.
Download Pure Data External for Reactive HMM-Based Speech and Singing Synthesis
In this paper, we present the recent progress in the M AGE project. M AGE is a library for reactive HMM-based speech and singing synthesis. Here, it is integrated as a Pure Data external, called mage~, which provides reactive voice quality, prosody and identity manipulation combined with contextual control. mage~ brings together the high-quality, natural and expressive speech of HMMbased speech synthesis with high flexibility and reactive control over the speech production level. Such an object provides a basis for further research in gesturally-controlled speech synthesis. It is an object that can “listen” and reactively adjust itself to its environment. Further in this work, based on mage~ we create different interfaces and controllers in order to explore the realtime, expressive and interactive nature of speech.