AES 2022 International Automotive Audio Conference – Program

(Provisional - subject to change)

Day 1 (June 8th):

TimeDearborn Ballroom FoyerGrosse Point RoomFairlane RoomDearborn Ballroom, Salon II-V
8:00 am

Conference Registration Open

Location: Dearborn Ballroom Foyer

Demo & Exhibition Open

Location: Dearborn Ballroom, Salon II-V (All Day)

8:15 am
8:30 am

Opening Comments

Location: Grosse Point Room


Conference Chair: Roger Shively (JJR Acoustics)
8:45 am

Keynote 1:

The Automotive Industry Transformation and the Demand for High In-Car Audio Quality

Location: Grosse Point Room


Keynote speaker: Mathias Johansson, PhD Dirac Research AB
9:00 am
9:15 am
9:30 am
Morning Coffee

Location: Dearborn Ballroom, Foyer

3 x 10 min Company Presentations

Location: Grosse Point Room

APEI Session 1:

Audio Weaver Development Platform / Voice
Hands-on experience with Audio Weaver centered on an automotive voice use-case

Location: Fairlane Room


(DSP Concepts)

Description:

  • 12 Participants receive a free evaluation board from STMicroelectronics

  • Observation seats also available


Agenda:
  • Audio Weaver Overview

  • Audio Weaver Designer GUI

  • Standard Designer Processing Modules

  • Audio Weaver Use Case

    • Automotive voice: external mic use case and the use of the RAPID Kit
9:45 am
10:00 am

Paper Session 1:

System Architecture and Hardware - Microphones, Test/Measurement method and equipment, Loudspeakers

Location: Grosse Point Room


Nonlinear Speaker Control in Automotive Sound Applications
by Joachim Schlechter, Dr. Stefan Irrgang, Stephan Schonfeld - Klippel

Enhanced Perceptual Rub & Buzz Measurement for Testing Automotive Speakers
by Steve Temme - Listen, Inc.

Electromagnetic Interference in Electric Vehicles and its Impact on AM Audio
by Pooja Nair (Xperi), David Layer (NAB), Ashruf El-Dibary (Xperi)

Improved Audio System Playback by Automatic Head Position Detection and System Adjustment
by Jonathan Lane, Hannah Perkins, Daniel Field (Panasonic)


10:15 am
10:30 am
10:45 am
11:00 am
11:15 am
11:30 am
11:45 am
12:00 pm

Tutorial 1:

High-Performance Equalizer Proposal for Personalized Listening in Vehicles

Location: Grosse Point Room


Speaker: Jay Yoo (Radsone Inc.)
12:15 pm
12:30 pm
Lunch

Location: Dearborn Ballroom, Foyer

12:45 pm
1:00 pm
(1:10pm) 2 x 10 min Company Presentations

Location: Grosse Point Room

1:15 pm
1:30 pm

Paper Session 2:

Sound Reproduction in Cars - Computer-Aided Tuning, Sound Field Control

Location: Grosse Point Room


Automated Control of Reverberation Level Using a Perceptional Model
by Philipp Weber, Christian Uhle, Sarah Hölscher, Matthias Lang and Thomas Bachmann - Fraunhofer IIS

Object-Based Audio as Platform Technology in Vehicles
by Christoph Sladeczek, Mario Seideneck, Wolfgang Lorenz, Katrin Pursche - Fraunhofer IDMT, Benjamin Schneider - DSP Concepts

Stereo upmix design for shaping sound experiences
by Søren Christensen, Pedro Højen-Sørensen, Morten Hansen, Lars-Johan Brännmark - Dirac

Experiential Benefits of Sound Zones for Automotive Applications
by Henrik Oppermann, Sofia Checa - Sennheiser (Schweiz) AG

APEI Session 2:


Voice & AI
Introduction to Machine Learning for Automotive Voice: Keyword Spotting using TFLM and Audio Weaver

Location: Fairlane Room


(DSP Concepts and Quilt Data)

Keyword spotting is a huge part of modern voice user interface (VUI) systems. Learn about the basics of how these systems work and how to train your own using Tensorflow. We will cover everything from feature generation and model training to deploying to production.

Agenda

  1. Introduction to ML for Audio,

  2. ML support in Audio Weaver,

  3. Keyword spotting using Tensorflow,

  4. Keyword spotting feature generation, model training, deployment
1:45 pm
2:00 pm
2:15 pm
2:30 pm
2:45 pm
3:00 pm
3:15 pm
3:30 pm
Afternoon Coffee

Location: Dearborn Ballroom, Foyer

3 x 10 min Company Presentations

Location: Grosse Point Room

3:45 pm
4:00 pm

Panel Discussion 1:

TC-AA White Paper: In-car Acoustic Measurements - Frequency Response

Location: Grosse Point Room



Moderator: Jayant Datta (White Paper Chair)
Panelists: Steve Hutt (Equity Sound Investments), Laurent Gagliardini (Stellantis)
4:15 pm
4:30 pm
4:45 pm
5:00 pm

Tutorial 2:

Position Dependent Amplitude Response in Automotive Loudspeakers

Location: Grosse Point Room


Keynote speaker: Mark Ziemba (Panasonic)
5:15 pm
5:30 pm
5:45 pm
6:00 pm
Break
6:15 pm
6:30 pm
Evening Reception

Location: Dearborn Ballroom, Salon II-V

6:45 pm
7:00 pm
7:15 pm
Day 1's Session ends at 6:00pm | Join us for a Casual Reception in the Dearborn Ballroom demo and exhibition space

Day 2 (June 9th):

TimeDearborn Ballroom FoyerGrosse Point RoomFairlane RoomDearborn Ballroom, Salon II-V
8:00 am

Conference Registration Open

Location: Dearborn Ballroom Foyer

Demo & Exhibition Open

Location: Dearborn Ballroom, Salon II-V (All Day)

8:15 am
8:30 am

Opening Comments

Location: Grosse Point Room


Conference Vice-Chairman: Dan Field (Panasonic)
8:45 am

Keynote 2:

Immersive Audio Experience in Automobiles

Location: Grosse Point Room


Keynote speaker: Andreas Ehret (Dolby)
9:00 am
9:15 am
9:30 am
Morning Coffee

Location: Dearborn Ballroom, Foyer

3 x 10 min Company Presentations

Location: Grosse Point Room

APEI Session 3:

Modeling & Measurement

Location: Fairlane Room



(COMSOL, Harman)

Mark Cops (COMSOL), Martin Olsen (Harman)

"Simulation-Driven Audio Design and Optimization"
Engineers and scientists use the COMSOL Multiphysics® software to simulate designs, devices, and processes in all fields of engineering, manufacturing, and scientific research. This leads to fewer physical prototypes, lower costs, better products, and faster time to market. The goal of this presentation is to introduce COMSOL Multiphysics® and highlight four important features for acoustic simulation.

"Integrated Virtual Engineering for (Pre-)development of Automotive Audio Solutions"
Computer aided engineering has become an essential part of most development processes in the automotive industry, and the value proposition is recognized by still increasing proportion of corporate organizational layers. This presentation provides a complete overview of the virtual engineering (pre-)development stages associated specifically with designing future car-audio systems. Three primary categories of simulation models are presented and illustrated with examples of relevant case studies, along with a walk-through of technical considerations related to estimation of acoustical boundary conditions and car interior sound field characterization.

9:45 am
10:00 am

Panel Discussion 2:

Making Immersive Audio in Cars a Reality

Location: Grosse Point Room



Moderator: Rüdiger Fleischer (Dirac)
Panelists: Andreas Ehret (Dolby), Stefan Meltzer (Fraunhofer), Wolfram Jähn (Audi), Mark Ziemba (Panasonic), Herbert Waltl (mediaHYPERIUM)
10:15 am
10:30 am
10:45 am
11:00 am
11:15 am
11:30 am

Paper Session 3:

Evaluation of Sound Quality

Location: Grosse Point Room


Automotive Audio System Evaluation over Headphones Based on the BRIRs of Different Listening Positions
by Yukun Pei, You Li, Pablo Ripollés - New York University

Sound Quality Assessment of Car Audio Systems
by Magnus Schaefer - Head Acoustics

11:45 am
12:00 pm
12:15 pm
12:30 pm
Lunch

Location: Dearborn Ballroom, Foyer

12:45 pm
1:00 pm
(1:10pm) 2 x 10 min Company Presentations

Location: Grosse Point Room

1:15 pm
1:30 pm

Paper Session 4:

Virtual Acoustics - Simulations
Machine & Deep Learning - Automatic Speech Recognition, Acoustic Event Detection

Location: Grosse Point Room


Achieving Maximum Audio Processing Throughout on the Latest Automotive Chipsets
by P. Beckmann, J. Whitecar, J. Peil, and K. Hegde - DSP Concepts

A Study on Speech Intelligibility Performance of Automotive Voice Microphones
by Yu Du, Balazs Varga, Viktor Dobos - Harman

System based approach for testing the power handling capabilities of automotive loudspeakers
by Joerg Lichtenstein, Marquis Fields, Dr. Martin Kreißig - Burmester Automotive Audio, GmbHc

An Acoustic Front-End to Speech Recognition in a Vehicle
by Mark Every, Xueman Li - BlackBerry QNX

Seeing with Sound: Detection and localization of moving road participants with AI-based audio processing
by Dr. Jeff Sieracki, Matthias Boehm, Prachi Patki, Matthew Caggiano, Markus Noll - Reality.ai and Infineon

APEI Session 4:

Computer Model Verification - Tips & Tricks to Ensure Trustworthy Audio Measurements

Location: Fairlane Room


Dan Foley (AxioMetrix)

Verifying the accuracy of a computer model is highly contingent on getting accurate measurement results. In many instances, measurement results are often blindly accepted when in actuality the measurements are highly suspect. This is especially the case when the test equipment can cost several thousand or dollars or more. This workshop will present a variety of tips one can use to ensure that measurement results do indeed reflect actual audio performance.



1:45 pm
2:00 pm
2:15 pm
2:30 pm

APEI Session 4:

Careers in Automotive Audio

Location: Fairlane Room



(Free to Students; Pre-registration required)
2:45 pm
3:00 pm
3:15 pm
3:30 pm
3:45 pm
4:00 pm
Afternoon Coffee

Location: Dearborn Ballroom, Foyer

3 x 10 min Company Presentations

Location: Grosse Point Room

4:15 pm
4:30 pm

Panel Discussion 3:

TC-AA White Paper: In-car Acoustic Measurements - Max SPL

Location: Grosse Point Room



Moderator: Jayant Datta (White Paper Chair)
Panelists: Steve Temme (Listen, Inc), Hans Lahti (Harman), Pat Dennis (Nissan)
4:45 pm
5:00 pm
5:15 pm
5:30 pm

Tutorial 3:

Virtual Speaker Integration Through Simulation

Location: Grosse Point Room


Keynote speakers:Jeremy Charbonneu, Thibaud Costes, and Alexis Talbot (Hexagon)
5:45 pm
6:00 pm
6:15 pm
6:30 pm
Break
6:45 pm
7:00 pm
Evening Reception

Location: Dearborn Ballroom

7:15 pm
7:30 pm
Banquet

Location: Dearborn Ballroom, Salon II-V

7:45 pm
8:00 pm
8:15 pm
8:30 pm
Social Event

Location: Dearborn Ballroom

Music by

E3 Detroit ( Jazz Trio)

8:45 pm
9:00 pm
9:30 pm
Day 2's Session ends at 6:30pm | Join us for a Reception, Banquet, and Social Event in the Dearborn Ballroom demo and exhibition space

Day 3 (June 10th):

TimeDearborn Ballroom FoyerGrosse Point RoomFairlane RoomDearborn Ballroom, Salon II-V
8:00 am

Conference Registration Open

Location: Dearborn Ballroom Foyer

Demo & Exhibition Open

Location: Dearborn Ballroom, Salon II-V (All Day)

8:15 am
8:30 am

Opening Comments

Location: Grosse Point Room


Conference Papers Chairman:Yu “Dewey” Du (Harman)
8:45 am

Invited Paper:

AI in Automotive Audio: Approaching Dynamic Driving Sound Design

Location: Grosse Point Room


Keynote speaker: Rodrigo Schramm and Robin Doerfler(Impulse Audio Lab GmbH)
9:00 am
9:15 am
9:30 am
Morning Coffee

Location: Dearborn Ballroom, Foyer

3 x 10 min Company Presentations

Location: Grosse Point Room

APEI Session 5:

Dirac: Intelligent Audio Tools

Location: Fairlane Room


9:45 am
10:00 am

Panel Discussion 4:

TC-AA White Paper: In-car Acoustic Measurements - Impulsive Distortion

Location: Grosse Point Room



Moderator: Jayant Datta (White Paper Chair)
Panelists: Jonaton Ewald (Volvo), Doug Holmi (Bose) , Stefan Irrgang (Klippel)
10:15 am
10:30 am
10:45 am

APEI Session 5 (continued):

Materials Science

Location: Fairlane Room


11:00 am
11:15 am
11:30 am

Paper Session 5:

Active Sound Management and Solutions - Active noise control system architectures

Location: Grosse Point Room


NVH meets Audio: Active Sound Design
by Jeroen Lanslots, Scott MacDonald - Siemens and Markus Bodden, Torsten Belschne - Neosonic

ANC Platform Simulation with Consideration of the Coupling Effect of Secondary Sources in Automotive Application
by Akshay Jagadish Khatokar, Samira Mohamady, Allahyar Montazeri - IAV GmbH

11:45 am
12:00 pm
12:15 pm
12:30 pm
Lunch

Location: Dearborn Ballroom, Salon I

12:45 pm
1:00 pm
(1:10pm) 2 x 10 min Company Presentations

Location: Grosse Point Room

1:15 pm
1:30 pm

Panel Discussion 5:

Noise and Sound

Location: Dearborn Ballroom, Salon I



Moderator: D Scott McDonald (Siemens)
Panelists: Gareth Williams (Mentor), David Trumpy (Harman), and Tzvika Fridman (Silentium)
1:45 pm
2:00 pm
2:15 pm
2:30 pm
2:45 pm
3:00 pm

Paper Session 6:

Active Sound Management and Solutions -
Active noise control system architectures,
Vehicle sound design (internal and external) - Simulation

Location: Grosse Point Room


Virtual pass-by noise sound synthesis from transfer path analysis data
by Mansour Alkmim, et al. - Siemens, KU Leuven, DMMS core Lab and Wim Desmet - KU Leuven

Pop and Burble Triggered Sounds Development Process
by Frank Valeri - General Motors, and Maxime Fouillet, Simona De Cesaris, Yvone Buhl - Mueller-BBM Active Sound Technology

Virtual Speaker Integration Through Simulation
Jeremy Charbonneu, Thibaud Costes, and Alexis Talbot (Hexagon)

3:15 pm
3:30 pm
3:45 pm
4:00 pm
4:15 pm
4:30 pm

Farewell

Location: Gross Point Room


Conference Chair:Roger Shively (JJR Acoustics)
4:45 pm
Day 3's Session ends at 4:45pm | Thank you, and Safe Travels!

Keynote Abstracts

Keynotes:

“The Automotive Industry Transformation and the Demand for High In-Car Audio Quality” (45 mins)

Mathias Johansson, PhD
Co-founder & Chief Product Officer – Dirac Research AB

Many automotive industry leaders talk of an on-going paradigm shift that resembles the disruptive change that the mobile phone industry underwent a decade ago. In this talk, we discuss how this industry transformation has already increased the demand for high audio quality of in-car entertainment systems. We discuss the primary drivers, and what the demand and opportunities for higher audio quality means in practice.

“Immersive Audio Experience in Automobiles” (45 mins)

Andreas Ehret
Director Automotive – Dolby

Content creation, playback technology, and consumer preferences all point toward the growing momentum of immersive audio in the consumer entertainment industry. Is the automotive audio industry ready to embrace this revolution? We will face new opportunities and challenges for how to make immersive in-car entertainment technically possible across a growing number of use cases. There exists tremendous future potential for enabling these types of experiences, and by analyzing this topic, we are helping unlock what this revolution will mean for the industry and consumers.

“The Future of Audio and How the Consumer is Changing what is being Development in Automotive Audio” (45 mins)

TBD – Harman

Paper Session 1 (Wed 8 June 10:00am) System Architecture and Hardware - Microphones, Test/Measurement method and equipment, Loudspeakers

“Nonlinear Speaker Control in Automotive Sound Applications” (30 mins)
by Joachim Schlechter, Dr. Stefan Irrgang, Stephan Schonfeld – Klippel

Sound design in cars, especially sound control, such as echo cancellation, noise control, sound zones require highly linear and in most cases time invariant behavior of the audio playback system. The weakest part in such a complex system is the electro-acoustic transducer (loudspeaker). Using a new control technology those problems inherent in traditional transducers can be considerably reduced. In this paper the consequences of those improvements for audio design in cars are elaborated.

“Enhanced Perceptual Rub & Buzz Measurement for Testing Automotive Speakers” (30 mins)
by Steve Temme – Listen, Inc.

Loudspeaker Rub & Buzz faults are a problem for automotive manufacturers as they sound harsh and immediately give the perception of poor quality. There are two places such faults can occur – during speaker manufacturing and installation of the speaker in the car. Perceptual distortion measurements are often considered the holy grail of end-of-line testing because rejecting speakers with only audible faults increases yield. Although such measurements have been around since 2011, production line adoption has been slow because until now, sensitivity to background noise has made limit-setting challenging. In this paper, we introduce a new algorithm that uses proprietary technology to reduce the impact of background noise on the measurement and offer more repeatable results. This facilitates limit setting on the production line, and makes it a truly viable production line metric for increasing yield. This metric also has applications for end-of-line automotive quality control tests. A buzzing loudspeaker in a car disappoints the customer and is costly to replace. It is also challenging for a service center to determine exactly where the buzzing is coming from and whether it is caused by a faulty loudspeaker or bad installation. Results from various algorithms will be shown, and their correlation to subjective and other non-perceptual distortion metrics explained.

“Electromagnetic Interference in Electric Vehicles and its Impact on AM Audio” (30 mins)
by Pooja Nair, Ashruf El-Dibary – Xperi, and David Layer – National Association of Broadcasters

Electromagnetic interference (EMI) generated by electric vehicles (EVs) can compromise the reception quality of AM signals. This paper explains some effects of EMI on AM reception and audio quality in EVs, and the importance of EMI mitigation, with the help of field test data using EVs. The reception performance of analog AM and all-digital AM HD Radio™ signals is then briefly compared based on results from analysis and field measurements. Finally, field test results for internal combustion engine (ICE) vehicles and EVs demonstrate the robustness of all-digital AM signals and the superior AM audio experience provided by AM HD Radio technology.

“Improved Audio System Playback by Automatic Head Position Detection and System Adjustment” (30 mins)
by Jonathan Lane, Hannah Perkins, Daniel Field – Panasonic

The current time-honored automotive sound system adjustment approach is to set the tuning equalization (EQ) of the audio system through objective and subjective evaluation means, to achieve a single average setting for the range of listener positions. This is shown to result in degraded audio playback for those whose seating position puts their ears on the fringes of the normally measured space (5th through 95th percentile, mixed gender adults). Deviation over the space is especially notable in the case of nearfield speakers, like headrest and overhead speakers. Methods to remedy this sub-optimal outcome are investigated and an automatic means to make suitable adjustments resulting from this study is experimentally evaluated.

Paper Session 2 (Wed 8 June 1:30pm) System Architecture and Hardware - Microphones, Test/Measurement method and equipment, Loudspeakers

“Automated Control of Reverberation Level Using a Perceptional Model” (30 mins)
by Philipp Weber, Christian Uhle, Sarah Hölscher, Matthias Lang and Thomas Bachmann – Fraunhofer IIS

This paper describes a method for estimating the perceived intensity of reverberation in audio signals and controlling the level of an artificial reverberation signal such that an artificially reverberated output signal has similar reverberation properties as the corresponding input signal. The estimation uses a linear regression model with sub-band Inter-channel Coherence and Spectral Flatness Measure as input features which is trained with listening test data. For the adaptation to the artificial reverberation control signals are computed based on temporal modulation properties and correlation between the input channel signals and applied to compute an equivalent reverberation level. The resulting quantity is post-processed using signal-adaptive integration. For the evaluation the proposed method has been applied to control the reverb send gain of artificial reverberation in a car.

“Object-Based Audio as Platform Technology in Vehicles” (30 mins)
by Christoph Sladeczek, Mario Seideneck, Wolfgang Lorenz, Benjamin Schneider – Fraunhofer IIS

Current studies show that vehicle interiors will change more than they have in decades. This is due to simultaneously occurring megatrends. Autonomous driving in particular allows the driver to focus less on what is happening on the road. This leads to new usage concepts that shift attention to the interior experience. This brings with it completely new demands on sound systems. The availability of immersive entertainment technologies used for new comfort functions and mobile working will be essential. Due to the changed focus of attention, not only the personalized sound staging itself will take on a new importance, but also its correct spatial mapping. This applies to a wide variety of functions such as interior staging, driving safety, well-being or communication. In this context, the effort required to create the audio content will take on a significant role, requiring a new unified interface for spatial presentation in order to limit production efforts. This paper describes how object-based audio as a platform technology can be used to meet these requirements. Based on specific use cases, a new workflow is presented that has been implemented for use in series production. The concepts for rendering technology, audio tuning process and implementation on resource-limited hardware are described.

“Stereo upmix design for shaping sound experiences” (30 mins)
by Søren Christensen, Pedro Højen-Sørensen, Morten Hansen, Lars-Johan Brännmark – Dirac

In the currently evolving spatial audio paradigm, an ever-present question is how to meaningfully reproduce two-channel stereo recordings on multichannel systems designed for 3D or `immersive’ formats. The key challenge is to find a way of reproduction that, on the one hand, satisfies listeners’ expectations on immersion and, on the other hand, remains faithful to fundamental characteristics of the original recording, such as, e.g., tonal balance, image stability etc. To pragmatically enable a flexible shaping of these characteristics, an attractive signal chain is to first process the stereo signal into a number of intermediate, spatially decoded, channels that may subsequently be filtered and combined into output speaker channels. This paper presents some insights and results from explorations in this domain, with a focus on design principles towards low-artifact nonlinear stereo upmixing.

Experiential Benefits of Sound Zones for Automotive Applications” (30 mins)
by Henrik Oppermann, Sofia Checa – Sennheiser (Schweiz) AG

New technologies and software developments are pushing the boundaries of what is possible with spatial audio. The automotive industry has already started to pick up on the benefits spatial audio can bring to the in-car experience. In this paper, we take a specific focus on spatial audio for automotive as it relates to personalised sound zones. Personalised sound zones open up brand new sonic opportunities for in-cabin entertainment, communication, safety, and more. With the combination of algorithms and hardware we can re-think the interactions with audio inside the car and solve many accepted pain points of the in-car experience.

Paper Session 3 (Thurs 9 June 11:30am) Evaluation of Sound Quality

“Automotive Audio System Evaluation over Headphones Based on the BRIRs of Different Listening Positions” (30 mins)
by Yukun Pei, You Li, Pablo Ripollés – New York University

In recent years, car manufacturers have consistently upgraded the audio systems of their vehicles, with audio aficionados adding further modifications to them. The acoustics of the vehicle cabin and the sound effects of the audio systems have become one of the most important topics of the Research and Development Departments of manufacturers. For example, Tesla has implemented different spatial audio algorithms in its models to achieve a better listening experience. By capitalizing on impulse responses, binaural audio technology provides the opportunity and the flexibility to virtually generate the sound effects of a particular space without the requirement of physically being in that space. We used binaural technology and a Tesla Model 3 vehicle to develop a standardized procedure for the evaluation of car audio systems. A perceptual listening test was integrated into this study to verify the procedure and to further evaluate the audio system of the Tesla Model 3.

“Sound Quality Assessment of Car Audio Systems” (30 mins)
by Magnus Schaefer – Head Acoustics

The sound quality of a car audio system is an integral part of the perceived quality of a car for many customers. Accordingly, car manufacturers are always looking for ways to improve and optimize their audio systems. This contribution presents different approaches for quantifying the perceived audio quality. As an example, several real car audio systems are evaluated both by a specific listening test and a novel instrumental assessment method that is based on a recently introduced set of analyses for audio quality. A comparison of the results of both approaches shows that the instrumental method provides the same information as the listening test in a fraction of the time.

Paper Session 4 (Thurs 9 June 1:30pm) Virtual Acoustics - Simulations | Machine & Deep Learning - Automatic Speech Recognition, Acoustic Event Detection

“Achieving Maximum Audio Processing Throughput on the Latest Automotive Chipsets” (30 mins)
by Paul Beckmann, Adam Levenson – DSP Concepts

 Automotive audio systems continue to grow in complexity with an ever growing set of use cases. System architects must now consider not only playback processing, but also telephony, voice control, safety sounds, individual sound zones, and even road noise cancellation. These use cases require huge compute resources that must be distributed across multiple cores. This paper explores the audio processing throughput of the latest standalone automotive DSPs and multicore SOCs with integrated DSPs. The effects of caching, memory architecture, threading, and multicore are presented and show you how to configure your system for maximum audio processing throughput.

“A Study on Speech Intelligibility Performance of Automotive Voice Microphones” (30 mins)
by Yu Du, Balazs Varga, Viktor Dobos – Harman

Automotive voice microphones are commonly used in vehicle audio systems for hands-free communication. One important performance measure of such a voice microphone is the speech intelligibility (SI). Unfortunately, this measure cannot be directly derived simply using the microphone datasheet. Based on both a subjective method (ANSI S3.2-2009) and an objective method (ANSI S3.5-1997), this paper investigates the SI score judged by human listeners versus the speech intelligibility index (SII) score calculated from an established algorithm. Three typical automotive voice microphone designs differing in directivity and frequency response (FR) characteristics are tested. It is shown that no intelligibility differences are noticeable among three microphones at low background noise levels. With medium to high level and non-wind induced noises, the unidirectional or omnidirectional microphone with a rising FR exhibits certain advantage over the omnidirectional microphone with a flat FR. With wind turbulence inducted noises which are typically at high levels, the unidirectional microphone sees the greatest performance degradation. In Addition, an approximate linear correlation is demonstrated between the subjective SI and the objective SII scores for automotive applications, which enables the use of the simple SII method to predict the SI score without physically conducting the laborious SI test.

“System based approach for testing the power handling capabilities of automotive loudspeakers” (30 mins)
by Joerg Lichtenstein, Marquis Fields, Dr. Martin Kreißig – Burmester Automotive Audio, GmbH

In this paper we investigate stresses on loudspeakers in an automotive audio system using real world scenarios and thereafter develop a test procedure for evaluating their power handling capabilities. At first some general considerations about a given automotive audio system and the measurement setup, including measurement signals, are made. Then measurements of several state variables are made using automotive speakers, automotive amplifiers, and a selection of worst-case music content. After analyzing the state variables under load, we generate an measurement signal. To validate the method the same measurements with the same loudspeakers are made in a “lab-setup”, using the derived measurement signal. Those measurements are compared to derive a test procedure.

“An Acoustic Front-End to Speech Recognition in a Vehicle” (30 mins)
by Mark Every, Xueman Li – BlackBerry QNX

The acceptance of speech as a primary user-interface in vehicles depends on how well speech recognition can overcome challenging conditions including high levels of noise, echo and competing speech, in which accuracy is known to degrade. To mitigate this, an acoustic front-end using the QNX Acoustics for Voice software library preprocesses multichannel microphone data from the vehicle and provides a cleaned signal to the recognizer. We demonstrate how three components of the front-end: beamforming, acoustic echo cancellation and zone interference cancellation, lead to significant improvements in word error rates.

“Seeing with Sound: Detection and localization of moving road participants with AI-based audio processing” (30 mins)
by Dr. Jeff Sieracki, Prachi Patki, Matthew Caggiano – Reality.ai, Matthias Boehm, Markus Noll – Infineon

Existing ADAS solutions for car environmental awareness (cameras, LiDAR, ultrasonic, etc.) typically require targets to be in a clear line of sight from the sensor. The target must be illuminated by some source of energy, so systems are affected by dust, weather, lighting, and obstacles. We address those limitations using a passive acoustic solution that “listens” to the environment. It can hear potential targets around corners or out of sight over a distance, providing early warning that supplements and improves other ADAS systems. We aim to detect a variety of road participant including sirens, as well as approaching vehicles, bicycles and even pedestrians. We discuss use cases and challenges, present an inexpensive reference architecture based on automotive grade components, and report on the state of development with initial validation results.

Paper Session 5 (Fri 10 June 11:30am) Active Sound Management and Solutions - Active noise control system architectures

“NVH meets Audio: Active Sound Design” (30 mins)
by Jeroen Lanslots – Siemens and Scott MacDonald, Markus Bodden, Torsten Belschne – Neosonic

NVH automotive engineers used to spend their whole career on reducing noise and vibration issues. The job is to come up with counter measures to remove or reduce these issues, leading to better performance, durability, and/or comfort. Digital twins have become very instrumental to validate design improvements without requiring physical prototypes. With vehicle electrification we enter the era where the powertrain is no longer the biggest contributor to interior and exterior noise. This makes that NVH engineers now must start adding sound to their vehicle designs again. That domain is called Active Sound Design and resides in between several domains: NVH and audio. This paper explores how Active Sound Design (ASD) can benefit from the Digital Twin approach.

“ANC Platform Simulation with Consideration of the Coupling Effect of Secondary Sources in Automotive Application”(30 mins)
by Akshay Jagadish Khatokar, Samira Mohamady, Allahyar Montazeri – IAV GmbH

The passenger comfort has always been of pivotal importance in the interior design of an automobile. One important aspect in reaching this goal in automotive industry is design and implementation of an effective active sound management system with the ability of personalizing the acoustic environment inside the car. Due to the complexity of the sound field inside the car cabin, having a high-fidelity model that reflects all details is a challenging task. Therefore, in this paper we
develop a simulation platform to be able to evaluate the performance of the ANC system and the distribution of the sound field as a result of this mechanism. This helps to get a better in-sight on the behaviors of the sound field inside the cabin before its actual implementation. One important feature of this model which may also have a large effect on the performance of the ANC system is the inclusion of a full-scale numerical model of the loudspeaker. The realistic model of the loudspeaker developed in this way allows modeling the effect of loudspeaker coupling amongst in an enclosed space and investigate its effect on the ANC system. The model is compared against the mathematical simplified model of the enclosure developed in the previous work by the authors to see how the approximate geometry and simplified model of the loudspeaker would degrade the performance of the ANC system and measure the changes in the acoustic radiation impedance of the loudspeaker.

Paper Session 6 (Fri 10 June 3:00pm) Active Sound Management and Solutions - Active noise control system architectures, Vehicle sound design (internal and external)

“Virtual pass-by noise sound synthesis from transfer path analysis data”(30 mins)
by Mansour Alkmim, et al. –  Siemens, KU Leuven

This work describes the sound synthesis of moving sources for the evaluation of pass-by noise scenarios. The proposed approach is based on transfer path analysis (TPA) from vehicle components to a semi-circular microphone array, enabling a decomposition in terms of spherical harmonics. The source signals are estimated by setting microphone indicators around source locations (engine compartment, tires) and by solving TPA inverse problems. A key feature lies in the simultaneous measurement of the pressure data at indicator and semi-circular array microphones. The spherical harmonic domain allows the sources to be moved smoothly in space, enabling the simulation of any trajectory including Doppler effect. Finally, the propagated source signal is decoded using an inverse spherical transform to synthesize binaural output. This approach renders a dynamic spatial audio environment that can be synchronized with a visualization tool. To illustrate the technique, preliminary results of an electric vehicle measured indoors are shown.

“Pop and Burble Triggered Sounds Development Process”(30 mins)
by Frank Valeri – General Motors, and Maxime Fouillet, Simona De Cesaris, Yvone Buhl – Mueller-BBM Active Sound Technology

Automotive customers have become accustomed to triggered sounds in vehicles for alerts and communications. The field of triggered sounds is expanding to include personalization and customer exciting triggered sounds [1]. These triggered sounds are broadcasts from the infotainment system and are both triggered and controlled via the vehicle CAN messages. This level of complex triggering and control allows OEMs to craft excitement generating triggered sounds while providing audible feedback which corresponds to vehicle operation. Pop and Burble (P&B) Enhancement is one example of a triggered sound that can generate excitement and provide a perception of powerfulness for the vehicle [2]. This paper will discuss the benefits and potential application of P&B enhancement, the process to develop the sounds, control logic and a software solution set to create, implement, and tune P&B enhancement.

“Virtual Speaker Integration Through Simulation “(30 mins)
Jeremy Charbonneu, Thibaud Costes, and Alexis Talbot – Hexagon

With the shift of vehicle manufacturing to rely heavily on a simulated design process, multiple iterations of prototype vehicles are no longer available. Finessing the fine details, such as the tuning of performance audio systems must therefore be completed on fewer vehicle builds as the components, mountings and materials are varied. Adding to this complexity is the reduced duration of the design window as the development time for new vehicles is getting shorter. Design engineers must be as efficient as possible in their efforts and use the best-in-class tools at their disposal; gaining insights earlier into their design process to make informed decisions.

Tutorials

“High-Performance Equalizer Proposal for Personalized Listening in Vehicles”
(30 mins)

Jay Chul-Jae Yoo – Radsone Inc.

Inside vehicles is one of the best listening environments. Music streaming apps, YouTube, podcasts from smartphones become major audio sources in cars while HD radio or satellite radio broadcasting are still useful in them. People would like to listen personalized playlists also in vehicles through smartphones connected with by USB or Bluetooth. So, audio sources become more and more personalized but car audio reproduction systems are mostly fixed to pre-tuned settings except their tone controls. Nowadays, many users are familiar to sound effects because they can use and test various effect apps in smartphones. Resultantly, their knowledge of sound was enriched and their own sound preferences become clear. Such users would like to realize personalized listening settings in vehicles too. The most used sound effect shall be an equalizer. Depending on audio sources (e.g. radio, music streaming, USB mp3), music genres, ages, where they are from, the preferred EQ target curves are not same, so the  predetermined tunes by automakers cannot satisfy such diversities. When new potential buyers test-drive, they check also the sound of the car audio systems. If they feel the sound thin or too boosted in their senses, they would not like to buy the cars. Most vehicles only provide tone balance controls, which are the minimal features of EQ. Definitely the tone control is a kind of level that is too insufficient in the age of smartphones. Automakers should consider giving users more control over the sound. In terms of devices, a car has a good system that can provide a high-performance EQ. In the case of smartphones, EQ, which requires processing power, is not provided properly due to limitations in screen size and battery life of them. A recent vehicle audio system has a horizontal screen display of sufficient size and an amplifier with powerful DSPs built-in. It is possible to provide a good enough EQ for consumer control on the display. Various audio sources used in the vehicle are ultimately played back through the vehicle’s amplifier, so applying an EQ to the vehicle’s amplifier enables an integrated control for all sources. In addition, if an EQ is applied at the end of the audio path, the amplifier, it is advantageous in terms of sound quality. Currently, the tone control provided by most vehicles has the advantage of being easy to operate, but the operation range is too limited and the same settings applies to all audio sources. If you increase the bass of the tone control while listening to the radio, you have to adjust it again when listening to speech such as a podcast to enhance the
speech intelligibility. On the other hand, direct controlling various parameters of a Parametric EQ used in pro audio seems too over-spec for the vehicle EQ. In EQ apps provided by smartphones, graphic EQs are mostly used, and users have experience with them.

Considering this situation, the proposed vehicle EQ is a 10-band Graphical Equalizer with ISO center frequencies. After measuring the existing EQ apps, we found that the center frequencies are quite different according to apps, the gain responses are not accurate, even worse in the low bands, and various problems are in clipping processing. The proposed EQ has the entire structure consisting of a soft switch for preprocessing, an EQ main block, and a finalizer for post-processing which includes a DRC and a soft clipper along with methods for realizing accurate frequency responses according to EQ adjustments, a practical clipping algorithm that can be activated for saturated signals. In EQ user interfaces, storage buttons to save EQ values according to each audio source were provided for one click loadings and the whole frequency response curve display according to specific band gain adjustments with GEQ band sliders were provided for more accurate tuning.

“Position Dependent Amplitude Response in Automotive Loudspeakers” (60 mins)
Mark Ziemba – Panasonic

This Tutorial will review previous research in this area and how it audibly impacts the accuracy of music reproduction in automotive audio systems.

  • New measurements on current loudspeakers that show difficulties in this area, and correspondingly some current loudspeakers that use specific design elements to overcome these anomalies.
  • Live audible demonstrations of different loudspeakers that exhibit this AM distortion compared to others that have better performance. This performance will be monitored through the sound reinforcement system with video so every conference attendee can see the loudspeaker under test and experience the sound of the distortion with music examples.
  • Some old and some new techniques will be used to make the distortion more audible to the attendees.
  • Klippel linearity measurements on several these loudspeakers will be shown so we can isolate or eliminate the causes of these audible anomalies.
  • Some conclusions will be offered and discussion entertained on loudspeaker design trade offs. How to optimize different measurable loudspeaker parameters for specific applications is always a valuable discussion topic and should be of interest to all audio engineers.

“Virtual Speaker Integration Through Simulation ” (60 mins)
Jeremy Charbonneu, Thibaud Costes, and Alexis Talbot – Hexagon

With the shift of vehicle manufacturing to rely heavily on a simulated design process, multiple iterations of prototype vehicles are no longer available. Finessing the fine details, such as the tuning of performance audio systems must therefore be completed on fewer vehicle builds as the components, mountings and materials are varied. Adding to this complexity is the reduced duration of the design window as the development time for new vehicles is getting shorter. Design engineers must be as efficient as possible in their efforts and use the best-in-class tools at their disposal; gaining insights earlier into their design process to make informed decisions.

This tutorial will aim to demonstrate an effective workflow for assessing sound system performance in an automotive compartment. This multi-stage novel approach utilizes a finite element method for developing the excitation boundary condition and the discontinuous Galerkin method (DGM) for propagating results within the vehicle cavity. The excitation pattern to be injected into the models will be based on a 1-Dimensional loudspeaker model which directly employs the Thiele and Small parameters of the loudspeaker supplier to transform an input voltage signal into a vibration of the speaker membrane. The final model compares the traditional absorption approach using an impedance boundary condition, to a workflow including 3-Dimentional DGM porous components within the treated cabin. To illustrate the compatibility between DGM techniques and GPU acceleration, a comparison has been performed using a single node with a single GPU that match runtime observed when using several high-performance computing nodes and hundreds of cores.

Panel Discussions

TC-AA White Paper: In-car Acoustic Measurements – Frequency Response (90 mins)
Moderator: Jyant Datta (White Paper Chair)
Steve Hutt (TBD), Laurent Gagliardini (Stellantis)

In general

Currently, there is no consensus in automotive on how to measure essential attributes. There are multiple methods of measuring system performance – all different. There is a high need for alignment between OEMs and Tier1 suppliers, and standard measurement procedures are needed:

  • Measurement conditions will be well known – would allow comparisons (and repetitions)
  • Make system specs/requirements meaningful, faster to compare, analyze and write requirements
  • Easy to interpret results (measurements conditions are known)
    This could lead to standardized measurements

Frequency Response

Frequency Response of an audio system is a quantitative measure of amplitude variation as a function of frequency in response to a constant input signal. The test signal is chosen with some quantity such as voltage (for analog input) or digital level (for digital input) with amplitude held constant. The frequency response range is typically defined over the audible frequency range 20 Hz to 20 kHz or a more limited range.

The Frequency Response may be described with variance limits, for example a +/- 3dB over the frequency range of interest either in “absolute” or “relative” terms. It should be noted that average smoothing of the Frequency Response modifies the amplitude of variance limits.

Making Immersive Audio in Cars a Reality (90 mins)
Moderator: Rüdiger Fleischer (Dirac)
Andreas Ehret (Dolby), Stefan Meltzer (Fraunhofer), Wolfram Jähn (Audi), Mark Ziemba (Panasonic), Herbert Waltl (mediaHYPERIUM)

Automotive might be the most ideal environment to enjoy all aspects of immersive music on a day-to-day basis. All players across the ecosystem must collaborate to implement high-quality immersive music in cars to create the experience artists originally intended. The panelists will dive into greater detail about bringing immersive audio into cars.

Dive into greater detail regarding immersive audio in cars. Highlight industry trends and discuss how high-quality immersive entertainment experiences are implemented in cars. Discuss the different players and how an in-car immersive audio experience is made possible through collaboration with different partners across the ecosystem. Mention experiential aspects of what it’s like to hear immersive audio in the car, highlighting artistic intent and the value it brings to consumers.

TC-AA White Paper: In-car Acoustic Measurements – Max SPL (90 mins)
Moderator: Jyant Datta (White Paper Chair)
Steve Temme (Listen, Inc), Hans Lahti (Harman), Pat Dennis (Nissan)

Consumers want to compare how loud different infotainment systems can play inside a car. Maximum SPL (Max SPL) is the maximum sound pressure level that a car’s infotainment system can reproduce inside its cabin with all its windows, sunroof, convertible top, etc. closed. There are many ways this can be measured but to keep it simple, two methods are recommended: overall Max SPL and Max SPL Spectrum, regardless of distortion level. Distortion is very difficult to characterize in a modern day infotainment system because of all the signal processing which was not designed to play back sine waves.

  • SPL – important in system-level specs, puts demands on loudspeakers, amplifier, and integration
  • SPL – easy to compare between systems
  • THD – can be used to calibrate max. SPL (e.g. limit to 10% THD)
  • THD – can be linked to the perceptual domain

TC-AA White Paper: In-car Acoustic Measurements – Impulsive Distortion Measurements – Max SPL (90 mins)
Moderator: Jyant Datta (White Paper Chair)
Jonaton Ewald (Volvo), Doug Holmi (Bose) , Stefan Irrgang (Klippel)

Impulsive distortion is never an intended design choice and is not related to the design goals. It is related to unpredictable nonlinear dynamics caused by overload and defects. While loudspeaker drivers may exhibit impulsive distortion (rub and buzz), the primary cause is most often vehicle integrity, i.e. when the vehicle’s structure manifests vibrations as buzz, squeak and rattle.

They have a higher crest factor than the regular nonlinear distortion, which is a unique characteristic.

The impulsive distortion contains high frequency harmonic and non-harmonic components in the reproduced audio signal, which are easily audible for a low frequency stimulus and degrade the sound quality significantly.

The commonly used phrase “Squeak and Rattle” describes specific (mostly mechano-acoustic nature) but no general symptoms that are related to impulsive distortion.

Low frequency excitation producing high mechanical load is a critical test signal.

Impulsive distortion symptoms are much easier to detect by “normal consumers” than harmonic distortion or a change of the frequency response by several dB. When comparing audio performance in cars, usually the annoyance of present impulsive distortion is dominant and needs to be addressed.

Noise and Sound (90 mins)
Moderator: Scott McDonald (Siemens)

Panelists:
Gareth Williams (Mentor), David Trumpy (Harman), and Tzvika Fridman (Silentium)

ANC/RNC
AVAS, Exterior Sound Design
A2B Analyzer Products
Testing Systems and Subsystems

Choose your country of residence from this list:










Skip to content