AES Los Angeles 2014
Paper Session P19
P19 - Signal Processing: Part 3
Sunday, October 12, 1:30 pm — 5:00 pm (Room 308 AB)
Duane Wise, Wholegrain Digital Systems LLC - Boulder, CO, USA
P19-1 Eliminating Transducer Distortion in Acoustic Measurements—Finn Agerkvist, Technical University of Denmark - Kgs. Lyngby, Denmark; Antoni Torras-Rosell, Danish National Metrology Institute - Lyngby, Denmark; Richard McWalter, Technical University of Denmark - Lyngby, Denmark
This paper investigates the influence of nonlinear components that contaminate the linear response of acoustic transducers and presents a method for eliminating the influence of nonlinearities in acoustic measurements. The method is evaluated on simulated as well as experimental data and is shown to perform well even in noisy conditions. The limitations of the Total Harmonic Distortion, THD, measure is discussed and a new distortion measure, Total Distortion Ratio, TDR, which more accurately describes the amount of nonlinear power in the measured signal, is proposed.
Convention Paper 9204
P19-2 Uniformly-Partitioned Convolution with Independent Partitions in Signal and Filter—Frank Wefers, RWTH Aachen University - Aachen, Germany; Michael Vorländer, RWTH Aachen University - Aachen, Germany
Low-latency real-time FIR filtering is often realized using partitioned convolution algorithms, which split the filter impulse responses into a sequence of sub filters and process these sub filters efficiently using frequency-domain methods (e.g., FFT-based convolution). Methods that split both, the signal and the filter, into uniformly-sized sub filters define a fundamental class of algorithms known as uniformly-partitioned convolution techniques. In these methods both operands, signal and filter, are usually partitioned with the same granularity. This contribution introduces uniformly-partitioned algorithms with independent partitions (block lengths) in both operands and regards viable transform sizes resulting from these. The relations of the algorithmic parameters are derived and the performance of the approach is evaluated.
Convention Paper 9205
P19-3 Modeling the Nonlinear Behavior of Operational Amplifiers—Robert-Eric Gaskell, McGill University - Montreal, QC, Canada; GKL Audio Inc. - Montreal, QC, Canada
Due to the gain-bandwidth characteristics of operational amplifiers, their nonlinearities are frequency dependent, showing a rise in distortion at higher frequencies. Depending on the circuit and system implementations, this distortion can be significant to listener perception of sonic character and quality and is therefore relevant to models of op amp-based analog equipment. Power-series models of the harmonic signature of various op amp nonlinearities are developed with and without this frequency dependence. Listening tests are performed to determine the extent to which the distortion characteristic of the model must match that of the real component to create a perceptually similar result.
Convention Paper 9206
P19-4 More Cowbell: A Physically-Informed, Circuit-Bendable, Digital Model of the TR-808 Cowbell—Kurt James Werner, Center for Computer Research in Music and Acoustics (CCRMA) - Stanford, CA, USA; Stanford University; Jonathan S. Abel, Stanford University - Stanford, CA, USA; Julius O. Smith, III, Stanford University - Stanford, CA, USA
We present an analysis of the cowbell voice circuit from the Roland TR-808 Rhythm Composer. A digital model based on this analysis accurately emulates the original. Through the use of physical and behavioral models of each sub-circuit, this model supports accurate emulation of circuit-bent extensions to the voice's original behavior (including architecture-level alterations and component substitution). Some of this behavior is very complicated and is inconvenient or impossible to capture accurately through black box modeling or structured sampling. The band pass filter sub-circuit is treated as a case study of how to apply Mason's gain formula to finding the continuous-time transfer function of an analog circuit.
Convention Paper 9207
P19-5 A Modal Architecture for Artificial Reverberation with Application to Room Acoustics Modeling—Jonathan S. Abel, Stanford University - Stanford, CA, USA; Sean Coffin, Stanford University - Stanford, CA, USA; Kyle Spratt, University of Texas, Austin - Austin, TX, USA
The modal analysis of a room response is considered, and a computational structure employing a modal decomposition is introduced for synthesizing artificial reverberation. The structure employs a collection of resonant filters, each driven by the source signal and their outputs summed. With filter resonance frequencies and dampings tuned to the modal frequencies and decay times of the space, and filter gains set according to the source and listener positions, any number of acoustic spaces and resonant objects may be simulated. Issues of sufficient modal density, computational efficiency and memory use are discussed. Finally, models of measured and analytically derived reverberant systems are presented, including a medium-sized acoustic room and an electro-mechanical spring reverberator.
Convention Paper 9208
P19-6 The Procedural Sound Design of Sim Cell—Leonard J. Paul, School of Video Game Audio - Vancouver, Canada
Synthesis was used to generate all of the audio for the sound design of educational game Sim Cell using the open source language Pure Data . A primary advantage of using Pure Data is that it can be easily embedded into games for iOS, Android, and other platforms. This paper illustrates different examples of how synthesis can be effectively used in video games in contrast to more conventional contemporary audio production methods such as sampling. Synthesis allows for the accurate rendering of high resolution audio easily in addition to very high rates of data compression when compared to sampling.
Convention Paper 9209
P19-7 OBRAMUS: A System for Object-Based Retouch of Amateur Music—Jordi Janer, Universitat Pompeu Fabra - Barcelona, Catalunya, Spain; Stanislaw Gorlow, Sony Computer Science Laboratory - Paris, France; Keita Arimoto, Yamaha Corporation - Iwata, Shizuoka, Japan
In the more recent past, the area of semantic audio has become an object of special attention due to the increase in attractiveness of signal representations that allow manipulations of audio on a symbolic level. The semantics usually refer to audio objects, such as instruments, or musical entities, such as chords or notes. In this paper we present a system for making minor corrections to amateur piano recordings based on a nonnegative matrix factorization. Acting as middleman between the signal and the user, the system enables a simple form of musical recomposition by altering pitch, timbre, onset, and offset of distinct notes. The workflow is iterative, that is the result improves stepwise through user intervention.
Convention Paper 9210