AES Milan 2018
Signal Processing Track Event Details
Wednesday, May 23, 09:15 — 10:15 (Scala 3)
Workshop: W01 - Designing for High Speech Intelligibility
Chair:Dirk Noy, WSDG - Basel, Switzerland
Obtaining a high speech intelligibility in large, voluminous spaces such as stadia or airports can be challenging; however, due to the requirements for audience information and entertainment as well as the criteria for emergency announcements the audio systems play a relevant role in safely operating any large public gathering place. This presentation outlines concepts of designing audio systems for high speech intelligibility. Further topics include the challenges introduced by non-optimal, reverberant room acoustics and their possibly detrimental effects towards Speech Intelligibility. A range of solution strategies will be presented based both on room acoustical and electro acoustical approaches. The presentation will conclude with a number of case studies, including Zurich Airport Checkin 2 Remodeling and KKL Lucerne Concert Hall Electroacoustics Refurbishment.
Wednesday, May 23, 09:30 — 11:00 (Scala 1)
Tutorial: T02 - Reusing and Prototyping to Accelerate Innovation in Audio Signal Processing
Presenters:Gabriele Bunkheila, MathWorks - Cambridge, UK
Jonas Rutstrom, MathWorks - Sollentuna, Sweden
Voice assistants are shifting consumer expectations on performance and capabilities of audio devices and human-machine interfaces. As new products are driven to deliver increasingly complex features, successful manufacturers and IP providers need to reuse more design assets, deliver original innovation more efficiently, and prototype more quickly than ever before. In this session you will learn about different techniques to integrate existing code and IP into early simulations of algorithms and system designs, ranging from embeddable code to cloud-based services. You will also be exposed to quick prototyping workflows, including methods for running in real-time and validating ideas on live real-world signals. The presentation will go through practical worked examples using MATLAB, while discussing some early-stage challenges in the design of voice-driven connected devices.
Wednesday, May 23, 10:15 — 11:15 (Scala 3)
Tutorial: T03 - Perceptually Motivated Filter Design with Applications to Loudspeaker-Room Equalization
Presenter:Balázs Bank, Budapest University of Technology and Economics - Budapest, Hungary
Digital filters are often used to model or equalize acoustic or electroacoustic transfer functions. Applications include headphone, loudspeaker, and room equalization, or modeling the radiation of musical instruments for sound synthesis. As the final judge of quality is the human ear, filter design should take into account the quasi-logarithmic frequency resolution of the auditory system. This tutorial presents various approaches for achieving this goal, including warped FIR and IIR, Kautz, and fixed-pole parallel filters, and discusses their differences and similarities. It also shows their relation to fractional-octave smoothing, a method used for displaying transfer functions. With a better allocation of frequency resolution, these methods require a significantly lower computational power compared to straightforward FIR and IIR designs at a given sound quality.
![]() | This session is presented in association with the AES Technical Committee on Loudspeakers and Headphones and AES Technical Committee on Signal Processing |
Wednesday, May 23, 11:15 — 12:45 (Scala 3)
Workshop: W03 - Audio Repurposing Using Source Separation
Chair:Philip Coleman, University of Surrey - Guildford, Surrey, UK
Panelists:
Estefanía Cano Cerón, Fraunhofer Institute for Digital Media Technology (IDMT) - Ilmenau, Germany
Chungeun Kim, University of Surrey - Guildford, Surrey, UK
Jon Francombe, BBC Research and Development - Salford, UK
Jouni Paulus, Fraunhofer IIS - Erlangen, Germany; International Audio Laboratories Erlangen - Erlangen, Germany
Source separation tries to extract sound objects from an existing mixture. In reality, perfect separation is not achievable; the sound quality of single extracted sources is often heavily degraded. Fortunately, if the separated sources are recombined with small alterations in level or position, the degradations are often masked. This workshop discusses using source separation to enable repurposing the original audio content: speech intelligibility can be improved for broadcast listeners or cochlear implant users; sound objects can be extracted from a recording to enable object-based transmission and rendering. It is important to be able to assess the sound quality of the remix and the extent to which audio remixing is possible. We will highlight possible evaluation methods involving listeners and state-of-the-art algorithms.
![]() | This session is presented in association with the AES Technical Committee on Semantic Audio Analysis |
Wednesday, May 23, 11:15 — 12:15 (Scala 1)
Tutorial: T04 - Build a Synth for Android
Presenter:Don Turner, Developer Advocate, Android Audio Framework - UK
With 2 billion users Android is the world's most popular operating system, and it can be a great platform for musical creativity. In this session Don Turner (Developer Advocate for the Android Audio Framework) will build a synthesizer app from scratch* on Android. He'll demonstrate methods for obtaining the best performance from the widest range of devices, and how to take advantage of the new breed of low latency Android "pro audio" devices. The app will be written in C and C++ using the Android NDK APIs.
*Some DSP code may be copy/pasted
![]() | This session is presented in association with the AES Technical Committee on Audio for Games |
Wednesday, May 23, 16:30 — 18:00 (Scala 3)
Tutorial: T08 - Modern Sampling: It’s Not About the Sampling; It’s About the Reconstruction!
Presenter:Jamie Angus, University of Salford - Salford, Greater Manchester, UK; JASA Consultancy - York, UK
Sampling, and sample rate conversion, are critical processes in digital audio. The analogue signal must be sampled, so that it can be quantized into a digital word. If these processes go wrong, the original signal will be irretrievably damaged!
1. Does sampling affect the audio?
2. Can we reconstruct audio after sampling?
3. Does sampling affect the timing, or distort the music?
4. Can modern sampling techniques improve things?
This tutorial will look at the modern theories of sampling, and explain, in a non-mathematical way, how these modern techniques can improve the sampling and reconstruction of audio.
Using audio examples, it will show that sampled audio, when properly reconstructed, preserves all of the original signal. Because it’s not the sampling but the reconstruction that matters!
![]() | This session is presented in association with the AES Technical Committee on High Resolution Audio |
Thursday, May 24, 09:00 — 10:30 (Scala 3)
Workshop: W09 - The State of the Art in Sound Synthesis and Procedural Audio
Chair:Joshua D. Reiss, Queen Mary University of London - London, UK
Panelists:
Stefan Bilbao, University of Edinburgh - Edinburgh, UK
Davide Rocchesso, University of Palermo - Palermo, Italy
Stefania Serafin, Aalborg University - Copenhagen, Denmark
Vesa Välimäki, Aalto University - Espoo, Finland
Sound synthesis covers not just analog and MIDI synthesizers but the full range of algorithmic approaches to sound generation. Similarly, procedural audio is characterized by the philosophy of "sound as process, not samples," Procedural audio has been enthusiastically embraced by the games industry, and sound synthesis in all its forms may revolutionize sound design for animation, games, VR, augmented reality, and across the creative industries. This Workshop gives an overview of recent developments in the field. It brings together leading researchers to explain the key concepts and discuss new approaches and technologies. It will be relevant to practitioners, enthusiasts, and researchers wishing to gain a deeper understanding of this rapidly changing field.
![]() | This session is presented in association with the AES Technical Committee on Audio for Games and AES Technical Committee on Audio for Cinema |
Thursday, May 24, 10:30 — 12:00 (Arena 3 & 4)
Tutorial: T11 - Total Timbre: Tools and Techniques for Tweaking and Transforming Tone
Presenter:Alex Case, University of Massachusetts Lowell - Lowell, MA, USA
Recordists shape timbre through the coordinated use of several different signal processors. While equalization is a great starting point, the greatest tonal flexibility comes from strategic use of additional timbre-modifying signal processors: compression, delay, reverb, distortion, and pitch shift. This tutorial defines the timbre-driving possibilities of the full set of studio effects, connecting key FX parameters to their relevant timbral properties, with audio examples that reveal the results. This multi-effect approach to timbre enables you to extract more from the effects you already use, and empowers you to get the exact tones you want.
Thursday, May 24, 10:45 — 12:15 (Scala 3)
Tutorial: T13 - Perceptual and Physical Evaluation of Guitar Loudspeakers
Presenter:Wolfgang Klippel, Klippel GmbH - Dresden, Germany
Loudspeaker and headphones generate distortion in the reproduced sound that can be assessed by objective measurements based on a physical or perceptual model or just by a subjective evaluation performed in systematic listening tests. This tutorial gives an overview on the various techniques and discusses the way how measurement and listening can be combined by auralization techniques to give a more reliable and comprehensive picture on the quality of the sound reproduction. The separation of signal distortion in speech signals allows to assess signal distortion which are for most stimuli below the audibility threshold. Further analysis of the separated distortion signals gives clues to identify the physical root cause of loudspeaker defects which is very crucial for fixing design problems.
Thursday, May 24, 12:15 — 13:45 (Scala 3)
Workshop: W11 - Deep Learning for Audio Applications
Chair:Konstantinos Drossos, Tampere University of Technology - Tampere, Finland
Panelists:
Qiuqiang Kong, University of Surrey - Guildford, Surrey, UK
Stylianos Ioannis Mimilakis, Fraunhofer Institute for Digital Media Technology (IDMT) - Ilmenau, Germany
Christian Uhle, Fraunhofer Institute for Integrated Circuits IIS - Erlangen, Germany; International Audio Laboratories Erlangen - Erlangen, Germany
Deep learning is currently the most active research field. It has led to impressive improvements over former state-of-the-art methods in, for example, image classification, speech recognition, and text translation.
This workshop introduces the concepts of deep learning and shows how it is applied to various problems of the audio engineering community, e.g., acoustic event detection, signal enhancement, source separation, and classification.
![]() | This session is presented in association with the AES Technical Committee on Semantic Audio Analysis |
Thursday, May 24, 17:00 — 18:00 (Scala 3)
Tutorial: T17 - Use of Delay-Free IIR Filters in Musical Sound Synthesis and Audio Effects Processing
Presenter:Federico Fontana, University of Udine - Udine, Italy
The delay-free loop problem appears when an audio electronic system, typically an analog processor, is transformed in the digital domain by means of a filter network preserving the structural connections among nonlinear components. If such connections include delay-free loopbacks then there is no explicit procedure allowing for the computation of the corresponding digital filter output.
The tutorial will show how a delay-free IIR filter network is designed, realized, and finally computed; and why they have led to successful real-time digital versions of the Dolby B, the Moog and EMS VCS3 voltage-controlled filter, as well as nonlinear oscillators and RLC networks, magnitude-complementary parametric equalizers, and finite-difference time-domain scheme-based models of membranes characterized by low wave dispersion.
![]() | This session is presented in association with the AES Technical Committee on Signal Processing |
Friday, May 25, 14:30 — 16:00 (Scala 3)
Tutorial: T23 - Optimizing Transducer Design for Systems with Adaptive Nonlinear Control
Presenters:Gregor Höhne, Klippel GmbH - Dresden, Germany
Marco Raimondi, STMicroelectronics SRL - Cornaredo (MI), Italy
Modern loudspeakers increasingly incorporate digital signal processing, amplification and the transducer itself in one unit. The utilized algorithms not only comprise linear filters but adaptively control the system, deploying measured states and complex models. Systems with adaptive nonlinear control can be used to equalize, stabilize, linearize, and actively protect the transducer. Thus, more and more demands can be taken care of by digital signal processing than by pure transducer design, opening new degrees of freedom for the latter. The tutorial focuses on how this freedom can be utilized to design smaller and more efficient loudspeaker systems. Examples are given for how existing transducer designs can be altered to increase their efficiency and which new challenges arise when driving speaker design to its limits.
Friday, May 25, 15:00 — 16:30 (Arena 3 & 4)
Workshop: W27 - Artificial Intelligence in Your Audio
Chair:Jonathan Wyner, M Works Studios/iZotope/Berklee College of Music - Boston, MA, USA; M Works Mastering
Panelists:
Jonathan Bailey, iZotope
Joshua D. Reiss, Queen Mary University of London - London, UK
AI has been part of the listener's experience for many years . . . . now it is influencing the development of tools made for music production and music creation. We will look at how it is being used, the promise it holds in developing new tools and the challenges it presents for music engineers and producers.
![]() | This session is presented in association with the AES Technical Committee on Recording Technology and Practices |
Saturday, May 26, 09:00 — 10:30 (Arena 3 & 4)
Student / Career: SC11 - Creating Audio Plugins with MATLAB
Presenter:Gabriele Bunkheila, MathWorks - Cambridge, UK
The first AES MATLAB Plugin Student Competition is now open for submissions – The entry deadline is August 15, 2018, with showcase and awards scheduled for the upcoming 145th Convention of the AES in New York City.
This optional tutorial covers the technical foundations needed to enter the competition. After attending, you will be able to build a simple VST plugin using MATLAB. You will learn about structuring code for real-time efficiency, defining clear interfaces to tune parameters via interactive controls, testing generated plugins against original designs, and much more.
The session will make use of practical coding examples – prior programming experience will be beneficial but is not required. A video recording will be available online after the Convention.
For more information about the competition, please visit "http://www.aes.org/students/awards/mpsc/">competition link
Saturday, May 26, 10:30 — 12:00 (Scala 2)
Tutorial: T28 - Intelligent Acoustic Interfaces for High-Definition 3D Audio
Presenter:Danilo Comminiello, Sapienza University of Rome - Rome, Italy
This tutorial aims at introducing a new paradigm for interpreting 3D audio in acoustic environments. Intelligent acoustic interfaces involve both sensors for data acquisition and signal processing methods with the aim of providing a high-quality 3D audio experience to a user. The tutorial explores the motivations for 3D intelligent interfaces. In that sense, the recent standardization of the MPEG-H has provided an incredible boost. Then, how to design 3D acoustic interfaces is analyzed, involving ambisonics and small arrays. Moreover, the main methodologies are introduced for the processing of recorded 3D audio signals in order to "provide intelligence" to interfaces. Here, we leverage the properties of signal processing in the quaternion domain. Finally, some examples of 3D audio applications are shown involving intelligent acoustic interfaces.