INNOVATIONS IN AUDIO • AUDIO ELECTRONICS • THE BEST IN DIY AUDIO

www.audioxpress.com

Market Update
Bone Conduction Transducers
By Mike Klasco and J. Martins

Practical Test & Measurement

The Benefits of Using
Structure-Borne Sound
in Headset Signal Processing
By Hans W. Gierlich

R&D Stories
The Science Behind
the Taction Transporter
By Silmon J. Biggs and John Steinberg

Simulation Techniques
Loudspeaker Driver
Displacement Decomposition
By René Christensen

R&D Stories
The Audio Foundry Automotive
Audio Development Platform
By Adam Levenson

Market Perspective
OTC Hearing Aids: Is Innovation
Leaving the Rule Behind?
By Andrew Bellavia
JANUARY 2023
MANUFACTURING / DESIGNING
TRANSFORMERS & ENCLOSURES
SINCE 1917

Hammond Manufacturing | www.hammondmfg.com

ax
From Ideas to Manufacturing
January 2023 ISSN 1548-0628 If you are holding a printed copy of this issue of audioXpress or reading it on a
screen, you have probably already noticed that this publication is unlike anything else
www.audioxpress.com out there in the audio industry. This publication was originally born from what was
considered once a very niche market segment of people who preferred a hands-on
approach to audio systems and reproduction, and liked to modify and (hopefully) improve existing
audioXpress (US ISSN 1548-0628) is published monthly,
equipment, as well as design and build their own gear from scratch.
at $50 per year for the US, at $65 per year for Canada,
That spirit inspired the late Edward T. Dell, Jr. (1923–2013) to devote decades of his life to the DIY
and at $75 per year Foreign/ROW, by KCK Media Corp.,
audio passion, leading to the creation of hallmark publications including Audio Amateur (1970), Audio
P.O. Box 417, Chase City, VA 23924, US Periodical Postage
Electronics, Glass Audio, Speaker Builder, Multimedia Manufacturer, and finally audioXpress, starting in
paid at Chase City, VA, and additional offices. January 2001. What many people recognized was that their passion for audio electronics and speaker
building, among other things, fostered many ideas that led to the foundation of new companies, business
careers, and audio innovation. Many of the original audioXpress readers were actually active members of
Head Office:
the industry and others got involved because of it. All over the world, the pages of Ed Dell’s publications
KCK Media Corp.
and the valuable article contributions by many hundreds of experts, inspired new ventures globally.
I joined audioXpress as editor-in-chief in 2013, and I was one of those people who these magazines
P.O. Box 417
had touched, so when invited I accepted the challenge of leading this publication with the inspiration
Chase City, VA 23924, US
and conviction that there was a lot more that we could do to serve our audio product development and
Phone: 434-533-0246
design readership.
Fax: 888-980-1303
We are now entering a new expansion phase in the trajectory that was initiated in 2013, when we
embraced our mission to follow audio innovation, true to our motto, Advancing the Evolution of Audio
Technology. And we continue to expand on our mission to offer the leading Audio Development and
Subscription Management: Design Magazine and Website. For that, you will note the important contributions from new authors
audioXpress and audio experts in almost every issue. This January 2023 issue is no different and I am particularly
P.O. Box 417
glad to welcome Andrew Bellavia and René Christensen, who I hope will be regular contributors to this
Chase City, VA 23924, US
Phone: 434-533-0246
publication. When you read their articles, I am certain you will agree.
E-mail: [email protected] A fundamental contributor to this project, as many of our authors recognize, has been the work of Jan
Internet: www.audioxpress.com Didden, who for many years has been listed as our Technical Editor, although his involvement was mostly
voluntary. As we aim to meet the evolving needs of our audience, I reached out to him to become more
closely involved with the publication, essentially expanding on what he always did brilliantly.
Postmaster: Send address changes to: I am extremely pleased to announce that Jan Didden has agreed to take on the formal duties as
audioXpress
Technical Editor of audioXpress (we decided that there wasn’t a better title for what he does), expanding
P.O. Box 417
his involvement and contributions to this publication. And appropriately, I leave you with a word from him.
Chase City, VA 23924, US
J. Martins
Editor-in-Chief
Advertising:
Strategic Media Marketing, Inc. It is a privilege for me to be part of the audioXpress team. I cut my writing teeth many, many
2 Main Street years ago producing articles for The Audio Amateur, the precursor to this title, founded by Ed Dell.
Gloucester, MA 01930, US Then, in 2010 I struck out on my own and started Linear Audio, publishing a 200-page “bookzine”
Phone: 978-281-7708 with some 10 technical audio articles every six months. Linear Audio’s author roster grew to more
Fax: 978-281-7706
than 100; experienced designers, well-known names in audio, as well as accomplished amateurs. You
E-mail: [email protected]
know who you are. In 2017, I decided I wanted to spend more time on research and equipment design
Advertising rates and terms available on request. and stopped publishing new Linear Audio issues. Occasionally I hooked up a prospective author with
audioXpress, and this activity will now expand.
I believe that design and/or construction articles are an indispensable part of a technical audio
Editorial Inquiries: journal. It is important to know what is out there, what’s available, who offers what, how it performs,
Send editorial correspondence, suggestions, and how it tests. But at the start of it all, there is an engineer or designer who creates the equipment
and contributed articles to:
that is proposed, tested, offered, and advertised. I want to give this part of the community the visibility it
audioXpress, Editorial Department
deserves, to inform and educate and hopefully inspire technically inclined persons to take up audio design.
I am looking forward to once again interacting with established and budding authors and designers to
P.O. Box 417
explore this great vocation of ours! So, get in touch with me at [email protected].
Chase City, VA 23924, US
E-mail: [email protected]
Jan Didden
Technical Editor

Legal Notice: The Team

Each design published in audioXpress is the intellectual Publisher: KC Prescott Associate Editor: Shannon Becker
property of its author and is offered to readers for their
Controller: Chuck Fellows Graphics: Grace Chen, Tori Zienka
personal use only.
Any commercial use of such ideas or designs without prior
Editor-in-Chief: João Martins Advertising Coordinator: Heather Childrey
written permission is an infringement of the copyright Technical Editor: Jan Didden
protection of the work of each author.
Regular Contributors: Bruce Brown, Bill Christie, Bob Cordell, Joseph D’Appolito, Dimitri Danyuk,
Vance Dickason, Jan Didden, Scott Dorsey, Gary Galo, Chuck Hansen,
© KCK Media Corp. 2022
Richard Honeycutt, Charlie Hughes, Mike Klasco, David Logvin,
Printed in the US
Ward Maas, Oliver Masciarotte, Nelson Pass, Christopher Paul,
Bill Reeve, Michael Steffes, Nora Wong, Stuart Yaniger, Ethan Winer

4 | January 2023 | audioxpress.com

 OUR NETWORK

SUPPORTING COMPANIES
ACO Pacific, Inc. 39 NTi Audio, Inc. 63
Amadeus 30 Parts Express International, Inc. 68
Audio & Loudspeaker Technologies International (ALTI) 45 Purifi Audio 65
Beyma 61 Redco Audio 25
Cine Gear Expo 33 Resonado 13
COMSOL, Inc. 9 SB Audience 23
DSP Concepts 29 Scan-Speak A/S 31
Glass Acoustic Innovations Technology Co., Ltd. (GAIT) 37 Taction 64
Hammond Manufacturing, Ltd. 3 Triad Magnetics 67
HEAD acoustics 21 Warkwyn 65
KAB Electro-Acoustics 51 Wavecor 17
KLIPPEL GmbH 49 Western Electric 2
Menlo Scientific, Ltd. 59

NOT A SUPPORTING COMPANY YET?

Contact Peter Wostrel ([email protected], Phone: 978-281-7708, Fax: 978-281-7706)
to reserve your own ad space in the next issue of the magazine.

audioxpress.com | January 2023 | 5

 Contents

Features
26 The Benefits of Using 54 Loudspeaker Driver
Structure-Borne Sound in Headset Displacement Decomposition
Signal Processing By René Christensen
By Hans W. Gierlich This valuable tutorial offers a unique overview
This article details the latest testing strategies on the possibilities of three-dimensional analysis
for bone conduction technology used in product of loudspeaker driver displacement, using Phase
development for headsets and in-ear devices. It Decomposition and Axial-Symmetrical Decomposition
showcases new possibilities for manufacturers techniques.
to better evaluate device performance, since no
appropriate test equipment and methods were 58 Power Transformer Parameters,
commercially available until now. Selection, and Testing
Part 3 — Insulation Materials,
40 OTC Hearing Aids Winding Bobbins, and Testing Methods
Is Innovation Leaving the Rule Behind? By Chuck Hansen
By Andrew Bellavia In the previous two articles in this series, Chuck
A wave of new developments promises to solve the Hansen detailed the history of transformer cores
most pressing need for people with milder hearing and their various construction methods. This article
loss, allowing mass-market hearables to supplant explores transformer insulation materials and winding
over-the-counter hearing aids with less stigma. bobbins, and testing methods for the insulation
system.
46 The Audio Foundry Automotive
Audio Development Platform
By Adam Levenson
This article details the possibilities of the Audio
Foundry platform for automotive audio prototyping
and to support the development of audio experiences
COVER ART in vehicles with powerful design tools and services.
Haylou PurFree Bone The result of a collaboration between Tymphany and
Conduction Earphones DSP Concepts.

6 | January 2023 | audioxpress.com

 Volume 54 – No. 1 January 2023

Drivers and Transducers Departments

MARKET UPDATE
8 Bone Conductors Going Mainstream 4 From the Editor’s Desk
By Mike Klasco and J. Martins
Bone conduction technology that sends vibrations directly to the inner ear, bypassing the 5 Supporting Companies
eardrums, is seeing increasing use—evolving from its origins in medical hearing devices to
mainstream consumer and strategic communication applications. The advantages in open-
ear wireless applications have great appeal and the audio industry is responding.

18 Building a Haptic Dimension

The Science Behind the Taction Transporter
By Silmon James Biggs and John Steinberg
Haptic motors are finding applications in gaming, music, VR/AR, and therapeutic/wellness.
Websites
The Transporter is a patented transducer that delivers faster more accurate bass and is www.audioxpress.com
designed to work with the main audio driver in headphones. It subtly vibrates against
the user’s skin to produce an extra dimension that can be described as a high-fidelity www.voicecoilmagazine.com
extension. www.cc-webshop.com
www.loudspeakerindustrysourcebook.com
www.circuitcellar.com
www.linuxgizmos.com

Columns
SOUND CONTROL
34 Acoustical Test Sources
By Richard Honeycutt
Richard Honeycutt identifies different types of acoustical testing and explores @audioxp_editor audioxpresscommunity
the science behind the choice of a sound source for acoustical testing. Some
commercially available examples of sound source examples are presented. linkedin.com/company/audioxpress

audioxpress

audioxpress.com | January 2023 | 7

ax Market Update — Bone Conduction Transducers

Bone Conductors
Going Mainstream

By
Mike Klasco and
J. Martins

Bone conduction technology that sends vibrations directly to the inner ear, bypassing the eardrums,
is seeing increasing use—evolving from its origins in medical hearing devices to mainstream
consumer and strategic communication applications. The advantages in open-ear wireless
applications have great appeal and the audio industry is responding.

Popular dynamic headphone drivers, the most critical by connecting a rod attached to a musical instrument between
component and the voice of the headphone, are mostly commodity someone’s teeth. And Beethoven is said to have used that technique
stuff and all are too close a cousin to the buzzer and microspeakers (putting a rod in his mouth) to hear the piano after losing most
from which they descended. This should not be a surprise as many of his hearing. This was the genesis of bone conduction, which
Asian headphone manufacturers started out producing buzzers eventually would already be perfected in the 20th century for
and injection-molded plastic parts. hearing aids and medical applications.
Conventional earphones do the job nicely for telephony An interesting overview of the many inventions, applications,
communications and music listening for most consumer and patents using some form of bone vibration and actuators
applications. The drivers on these designs produce sound by is available in a continuously updated article written by John
vibrating the air that reaches the external ear, eardrum, middle Mortensen (see Resources), discussing the hearing aid pioneers,
ear, and finally the inner ear. Yet, in-ear models are uncomfortable Bell Labs and Sonotone, and the patent disputes.
for many people for quite a few reasons. We all have experienced During the 1970s, a Swedish medical team perfected the idea
the pressure and lack of ventilation in our ear canals, the occlusion of implanting a vibration plate directly into the mastoid bone
effect of standing waves when our ears are plugged up, along with (the bone behind the ear) to transmit to the inner ear where
the loss of situational awareness. Loose-fit earbuds have changed the vibrations can be processed by the brain as sound, and use
the earphone market forever, as demonstrated by the massive an attachment connector for an external audio processor. This
success of Apple’s original AirPods. patented process is known as a bone-anchored hearing aid (BAHA),
Leaving the ear totally open by design, bone conduction and has been perfected with transcutaneous devices that are fully
“headphones” typically transmit vibrations directly to the human implanted under the skin. The sound processor is then connected
skull and skin, bypassing the ear canal, straight into the inner using magnetic attachments, allowing users to regularly upgrade
ear. The key value of bone conduction is that our ears are left their systems as technology evolves, without the need for any
completely open, allowing us to have a parallel hearing channel more surgeries, and more importantly, avoiding the dangers of
to our surroundings, while still enjoying the music. But to get an open wound.
to designs that are able to rival traditional earphones in music As an example, Cochlear, a leading implantable hearing solutions
fidelity, we still have some way to go. Yet the key proposition in specialist recently received U.S. Food and Drug Administration (FDA)
bone conduction is already appealing to a significant number of approval for the latest Cochlear Nucleus 8 Sound Processor, which
consumers and professionals. is not only smaller and lighter but is also the first behind-the-
ear cochlear implant sound processor using the next-generation
Sound Through Our Bones Bluetooth LE Audio technology—which has clear benefits for these
Long before the electroacoustic transducer was invented, applications.
bone conduction was already understood. In the 15th century Bone conduction technology has proven to be effective for
Italian physician and mathematician Girolamo Cardano, a prolific users with conductive or mixed hearing loss—basically people
inventor, described the fact that deaf people could “hear” music with a functioning cochlea but problems with the outer or middle

8 | January 2023 | audioxpress.com

Take the Lead
in Acoustics
with COMSOL Multiphysics®
Multiphysics simulation drives acoustics innovation by providing
insight into every design aspect that influences product
performance. The ability to account for coupled physics phenomena
lets you predict, optimize, and virtually test a design under real-
world conditions – even before a first prototype is built.

» comsol.com/feature/acoustics-innovation
ax Market Update — Bone Conduction Transducers

ear. It’s highly beneficial for users with single-sided deafness, combined with conventional electrodynamic and balanced armature
where a bone conduction device picks up sounds on the non- transducers.
functioning side and sends them as vibrations through the bone Evolution in microphones is a key area for hearing aids, since
to the functioning cochlea on the other side. that’s how environmental sounds are picked up to be transmitted
Due to the extensive use in hearing aid applications pioneered to the transducer. The processor is also a vital component,
by the Scandinavian leading brands in this field, bone conduction cleaning the signal picked up by the microphones, filtering the
transducers (the vibration actuator) have been constantly perfected key frequencies (e.g., for speech intelligibility), and optimizing
in combination with external processors, microphones, and the transmission through the skull bone, or through the skin,
depending on the design type.
These transducers are now relatively common in consumer
applications, evolving from the concept that was perfected in
Sweden by Bo Håkansson, professor of electrical engineering at
the Chalmers University of Technology, in Gothenburg. Håkansson
devised the balanced electromagnetic separation transducer (BEST)
a new type of bone conduction driver, which is now used by all the
leading hearing aid companies such as Oticon Medical (Demant
Group), and offered by Danish company Ortofon, as the holder of
the BEST license for conventional bone conductors.
The BEST transducers have managed to expand the frequency
range transmitted through bone conduction up to 10kHz, combining
a new type of electrodynamic transducer that uses a balanced
suspension principle to counterbalance the forces between the
vibrating parts and reduce distortion. The design also allows for
much smaller transducers that are still able to generate vibrations
down to 250Hz with higher sensitivity.
As stated, Oticon is the company that took this technology to
the next level in the treatment of hearing loss, but in the last 10
years, the technology was embraced for consumer applications.
Since a bone conduction system leaves the ear canal open, this is
a preferable experience by many users who cannot stand in-ear
contact, or simply are more interested in remaining aware of
their surrounding sounds, and are willing to accept the trade-
Cochlear, a leading company in implantable hearing solutions, recently
received FDA approval for the Cochlear Nucleus 8 Sound Processor, which off in fidelity.
is now the smallest and lightest behind-the-ear cochlear implant sound Also, as the effects of prolonged headphone use at unhealthy
processor available in the industry, already using the next-generation volumes is being documented, bone conduction is being promoted
Bluetooth LE Audio technology. as a much healthier alternative, particularly for children.
In these consumer applications, the product design sometimes
combines traditional bone conduction transducers with other
external transducers that act outside of the ear canal. Those can
combine haptic actuators optimized for extended low frequencies
(bass) and also conventional dynamic transducers that are beaming
to the area just outside the ear—such as the recent Bose Sport
Open Earbuds, a truly wireless design optimized for sports, fitness,
and wellness.
With the recent FDA approval of the long-expected rule
establishing Over-the-Counter (OTC) hearing aids, the audio
industry will continue to explore these combinations and open ear
designs. Manufacturers such as Sonion (another stalwart in hearing
solutions from Denmark) offers both bone conduction sensors and
vibration sensors, microphones, and balanced armature receivers.
In 2018, Sonion started a partnership with Valencell to expand the
The RadioEar B-81 high output bone transducer is based on the Balanced possibilities of using biometric sensors in the ear and optimize
Electromagnetic Separation Transducer (BEST) principle and widely used designs for in-ear and on-ear applications. Bone conduction pickup
in audiometry applications because of its higher output level at low sensors (an area we will explore in a separate article) will be key
frequencies with lower distortion. in the future of true wireless earbud and hearable designs.

10 | January 2023 | audioxpress.com

The Consumer Perspective augmented reality (AR) content, navigation, and work-from-home
Bone conduction technology was experimented with in early applications. The design needs interchangeable ear wings (silicone
headphones using vibration actuators. In 1994, a patent for loop fins) to provide a secure fit.
“Bone-conductive stereo headphones,” was granted to H. Werner The search for open-ear awareness is precisely where Auricle,
Bottesch. That patent described acoustical transducers “devised a young startup from Denmark, has focused. The Auricle open-
for the delivery of stereophonic soundwaves through the mastoid ear design uses two separate bone conduction units connected by
bone structure of the human skull.” This basically inspired the an active wire, and focuses on situational awareness. The design
concept for using bone conduction for leisure listening and music
reproduction, an alternative to conventional dynamic transducers.
Ten years later, the market had become somewhat familiar
with the concept of bone conduction headphones. But it was
the transition from conventional wired headphones to wireless
that generated attention from consumer electronics companies.
Products such as the Audio Bone established the headband form-
factor for use in fitness activities in 2008, but remained connected
by a wire to the sound source. With Bluetooth becoming more
reliable five years later, wireless bone-conduction headphones
hit the market, creating a category that has grown ever since.
The concept was instantly embraced by consumers who enjoy
fitness, running, and cycling, and obviously commuters who
appreciate the resulting safety of staying aware of environmental Typical bone conduction drivers combine a coil where the number of
windings affects the electrical impedance, a recoil device, or a moving
sounds. Some countries (e.g., the United Kingdom) even passed
part, two blade springs and magnets, all contained inside a housing, the
legislation determining that only open-ear bone-conduction
design of which influences the response due to resonances.
headphones can be used in competitive athletics and any sports
that take place in public environments (e.g., marathons).
For those activities, a key requirement is the need for the
bone conduction designs to remain in place even during intense
activity. And the fact that bone conduction needs to wrap around
the users’ head and create a tight grip is naturally an advantage
in those use cases. Of course, other companies have realized that
not all users appreciate the force exerted on their heads, and new
designs are now more focused on comfort for extended daily use,
such as in the office or for home work. The flexible neckbands
in comfortable rubber or silicone are now available for heads of
all shapes and sizes, allowing users to still appreciate the bone-
conduction headphones “hug” around the head, while keeping the
transducers firmly in place.
In any case, bone conduction products remain restricted to Panasonic had some success with its water- and sweat-resistant RP-
the headband or neckband form factor precisely because of the BTGS10-k wireless bone conduction headphones, that paved the way for
need to exert pressure. True wireless bone conduction designs the consumer category.
are a challenge, even when using some kind of clamp on the ear.
That’s why some companies have tried combining bone conduction
with other traditional transducers.
Bose’s solution to keeping the ear canal open has not taken
the bone conduction path. The Bose Sport Open Earbuds features
an unoccluded design that leaves the ear canals completely open.
They’re designed to rest securely on the natural curve of the ear,
using acoustic ports to channel sound directly into the ear. Sound
waves are sent through the air for a more conventional listening
experience by two 16mm drivers—one in each earbud—and a
dipole design to minimize the spread of sound, while funneling
Sonion, the manufacturer of components and solutions for hearing aids
it down the ear canals. and in-ear headphones, offers voice pick up bone sensors and also offers
The Sony LinkBuds are another interesting example, introducing its miniature magnetic vibration transducers for use in bone vibration
an open diaphragm ring design that delivers improved audio applications. The company also offers hybrid solutions with bone
transparency at the expense of bass response, yet ideal for conduction.

audioxpress.com | January 2023 | 11

ax Market Update — Bone Conduction Transducers

holds to the tragus, a small and unique cartilaginous structure to 50Hz, in order to achieve better fidelity for music reproduction.
of the ear, located on the anterior margin of the auditory canal The design has all the potential to eventually become completely
that offers an effective placement for bone conduction without true wireless in the next generation.
obstructing the ear canal. After extensive testing and research, the As Pedro Costa, the company’s founder explains, “We have
company perfected a very light enclosure that is able to remain found the placement on the tragus to be the most efficient location
firmly attached. The tragus grants a backward slope, guarding relative to the amount of power required to drive the transducer—
the opening of the auditory canal, which also proved ideal to irrespective of stimulated frequency. In other words, by attaching
hold a clip-on bone conduction transducer. The company is now the auricles to the tragus, we get the most bang for the buck in
optimizing the design for a new-generation bone/haptic conduction terms of force (aka how much sound we can transmit via the skin
transducer, able to deliver an extended frequency response down and bone relative to the output wattage/voltage) and frequency
range/response ‘balance’.”
For cyclists and motorcyclists, bone conduction transducers
are also now frequently incorporated in all types of helmets,
drawing in great measure from the experience in military and
aviation applications. In those applications, pressure is naturally
exerted by the helmet, even though not always in the right place.
Aside from the ergonomics, bone conduction transducers excel
in the human voice frequency range and can easily be optimized
for voice reproduction. That helps explain why many of the bone
conduction products on the market are popular among consumers
who use them for calls and listening to audiobooks and podcasts.

A Haptic Perspective
Ortofon Microtech offers a range of bone conduction transducers, Bone conduction, with its roots in military communications
including specialized versions for use in hearing clinics in combination and the hearing aid industry, has included vibration pickup as
with headphones for audiometry checks, models for hearing aids, and an alternative or to augment airborne sound pickup. Haptics
conventional models for speech or music signals. create an experience of touch by applying forces, vibrations, or
motions to the user. These actuator feedback devices are also
needed to create the illusion of substance and force within the
virtual world—applications in user and human-machine interfaces,
gaming, virtual reality, simulations, and the so-called metaverse.
Their functions and locations are truly all over the place with end-
users coming into contact with haptics in smartphones/tablets and
other touchscreens, wearables in gaming, sport fitness, medical,
wellness, and more.
In previous audioXpress articles we have explored bass shakers
for home theater, music performance, flight simulators, amusement
rides, and pilot training, as well as haptic actuators generating
reassuring vibrations for wellness or a simple “handshake” in a
button to acknowledge our order selection. Yet, by far, the most
popular use for haptics are the miniature devices in smartphones
that alert us when in our pockets.
Those haptic motors that most of us initially encountered in
flip phones used a Eccentric Rotating Mass (ERM) micromotor
with an off-center weight on the shaft. The ERM vibration motor
has faded, except at the bottom end of the market, as the more
sophisticated Linear Resonant Actuators (LRA) vibration motor is
now common for higher-grade applications.
Devices such as the TDK Mini PowerHap actuators have been
introduced for these more simple user interaction applications. But
EnListen launched specialized headphones for sound listening training
these traditional haptic devices only deliver simple vibration buzzes
based on a popular model produced by AKG, with an added bone
conductor inserted in the right ear cuff to send sound signals through the
due to the narrow frequency range and non-agile response. In
right mastoid bone. A dial could control how much conduction sound goes contrast, larger panel actuators (under the screen applications) and
to the left ear relative to the sound delivered to the right ear. Set at 10, today’s bone conduction devices are transducers that reproduce
both ears received the same contributed volume. a vibrational response analogous to the audio signal.

12 | January 2023 | audioxpress.com

 Beyond
conventional.

We invented the high performance racetrack driver.

Designer drivers with proprietary Res-Core™ motors

deliver unrivaled linearity and pistonic behavior.

Select drivers available now through licensed manufacturers.

Reach out to [email protected] to learn more.
ax Market Update — Bone Conduction Transducers

BeStar Technologies is a recognized haptics and piezo

actuator vendor. Aside from a number of haptics devices, BeStar
has introduced the ArFi Haptic Analyzer, a complete and unique
measurement system to be able to quantify haptics. ArFi is a
modular system that provides options to measure vibro-tactile
feedback as well as acoustic feedback, separately and in parallel.
Its strong advantage is clearly reflected during the design phase
of haptics, as it allows developers to quantify their haptic design
and tune it to the given application. ArFi displays the measured
data and calculates the intensity value that reflects the human
perception of the intensity.
In general, when we search for bone conduction products
online, we will find numerous “best of” lists that include products
Shokz, the company recognized for its patented bone conduction technology
and open-ear headphones, recently launched OpenRun Pro Mini, a smaller that do not use bone conduction. There are several interesting
version of Shokz’s most premium bone conduction sports headphones, the products that may not strictly be classified as bone conduction
OpenRun Pro. The new design is powered by patented Shokz TurboPitch 9th and have found applications in earphone or headphone drivers.
generation bone conduction technology, which adds two bass enhancers Some examples would be the aforementioned Bose OpenAudio
located inside the transducers for improved sound quality and enriched earphones, which use a dipole configuration to project the audio
bass, while keeping both ears open to surrounding ambient sounds. In in an open-ear design. Bose OpenAudio technology is also used
addition, spoken words are softer and more natural-sounding, making in the Bose Frames range of sunglasses, allowing users to hear
OpenRun Pro Mini a great choice for audiobooks and podcasts.
music and even take calls, and they don’t need the tight fit that
bone conduction would require.
Tectonic Audio has a transducer for the same purpose for off-
the-ear headphones (some readers might remember the legendary
AKG-K1000 off-the-ear design from a few decades ago). Tectonic’s
ancestry goes back to the NXT speaker, a flat panel driven by an
“exciter.” Its present product lines include exciters, and many
can also function for haptic applications and the smallest models
for earphones.
There have also always been special cases for first responders,
military, and other applications that require situational awareness
In 2020, Shokz launched the Opencomm headset, in a brilliant move that where an unobstructed ear canal is beneficial. Bone conduction
combined its previous generation model (Aeropex) with a microphone
devices can be designed to be difficult to recognize by others while
boom. Bone conduction is proven to work very well with voice frequencies
keeping the wearer informed. This is ideal for overt operations
(the reason why it’s used for military and fighter pilot helmet
applications), and since workplace situations and working from home or just a more discrete look (smart glasses avoid the cyborg look
(usually) offers less noisier environments, the combination results in a of the early Bluetooth devices).
great alternative compared with convention communication headsets that An example of an innovative specialist developer of bone
isolate the user. conduction devices for first responders, military, and the hearing
impaired is TEMCO from Japan. A pioneer of bone conduction
established in 1985, TEMCO introduced the world’s first bone-
conduction earphone microphones, communication systems for
helmets, bone-conduction sunglasses, and many other form-
factors. The company’s high-performance devices are small in
size, convenient to wear, resistant to shock and sweat, rain-proof,
and suitable for soldiers and firemen.
Shenzhen Voxtech is sort of the Chinese version of TEMCO,
with its products also focused on military personnel, security
forces, and sports enthusiasts. The British have BAE Systems,
another company using bone conduction for military helmets
and sailing teams.
Most of these bone conduction transducers are voice coil/
magnetic motor structures (electro-dynamic) along with a few
Staying aware of the surroundings and being able to enjoy music, take
calls, receive instructions, or listen to a reference track while playing on piezo types. Perhaps very soon microelectromechanical systems
stage. There are many advantages for these open-ear headbands using (MEMS) capacitive or piezo MEMS devices will be joining the club
bone conduction, and larger brands are paying attention. down the road.

14 | January 2023 | audioxpress.com

 Back a decade ago, AMI, a Bayer Material Science company,
attempted to break into the audiophile end of the headphone
market with Vivitouch, its electroactive polymer (or EAP in short).
A thin sheet that consisted of two electrode layers sandwiching a
dielectric elastomer film, and when a voltage is applied, the two
attracting electrodes compress the entire sheet.
This actuator provided the physical “oomph” to the conventional
headphone speakers, with both a powerful yet subtle contribution.
A few years back, AMI was acquired from Bayer by Parker Hamilton,
which is focused on industrial applications and Vivitouch is no
longer available.
For now, earphone designers can make hybrid combo designs
integrating both earphone drivers with bone conduction. Gaming
technology has always strived to deepen player engagement and
immersion. Auricle, a Danish startup, designed an open-ear bone conduction design
Haptics further narrow the gap between gaming and real life that holds to the tragus, a small and unique cartilaginous structure of the
ear that offers an effective placement and doesn’t require the headband
and can be used to extend the frequency response perception to
from other typical bone conduction designs. The patented design required
the lower frequencies. Mad Kat and Able Planet headphones used
extensive research and optimization of the transducer type/design and
Vivitouch, Razer has worked with Woojer and Lofelt, Corsair works placement.
with Taction, and Skullcandy offers its own Sensory Haptic Bass
solutions. Other gaming headphones have long offered tactile
enhancement.

Bone Conduction Brands

Many brands are introducing more refined extended frequency
range bone conduction earphone models. From the early days
of the Panasonic RP-BTGS10 bone conduction in 2013, all the
way to the market leader in the segment that is Shokz (formerly
Aftershokz) things have progressed greatly. There are quite a
few bone conduction earphones and we can’t list them all, but
for a sense of the market we will briefly mention some notable
big brands and specialist vendors.
Raw bone conduction transducers available in the market for
consumer electronics originate mainly with a handful of Asian myFirst, a Singapore-based tech brand, offers bone conduction
vendors, where there are also multiple OEM/ODM bone conduction headphones for children, which provide a safe way to listen to music
and watch movies or videos while still protecting hearing health. The
earphone suppliers. We mentioned Danish companies, Ortofon and
vibrations from the bone conduction also allows children’s ears to be
Sonion, which typically offer solutions oriented for hearing aids
open to external noise and are designed to sustain rough use.
but can also provide custom designs for consumer applications.
Sonion’s bone conducting actuators and sensors can be found in
hearing aids and communication, fitness, and lifestyle devices.
In the US, Dayton Audio offers the Bone Conduction Exciter
(BCE) range that can be used both to drive a surface or in bone
conduction earphones and glasses.
Lofelt from Germany has its patented L5 actuator that is used
in gaming applications, such as headphones, with past design
wins with Razer and others. The company is now off-the-market,
acquired this summer by Meta (Facebook)
Murata Manufacturing from Japan offers miniature vibration
devices utilizing the characteristics of piezoelectric ceramic that are
low power, have electromagnetic resistance, and support surface-
mount device (SMD) mounting, among other performance features
not found in conventional electromagnetic vibration devices. Piezo,
a division of Midé Technology in the US, is a ceramic actuator Haylou PurFree BC01 Bone Conduction Bluetooth sport headphones are a
specialist. The company offers piezoelectric actuators and motors Bluetooth 5.2, IP67 waterproof design with dual-device connectivity for
that can be used in a wide range of haptic applications. athletic activities.

audioxpress.com | January 2023 | 15

ax Market Update — Bone Conduction Transducers

In general, the leading manufacturers in this space have

perfected their home bone-conduction drivers in-house in an
effort to optimize the frequency response. Panasonic has been
producing bone conduction earphones for more than a decade
precisely because it develops its own transducers.
Shokz has also developed its own patented transducers. Shokz
has a large market share of the bone conduction headphones,
regularly updating its range to incorporate the latest Bluetooth
wireless technology and features.
The Shokz OpenRun model is highly respected and has received
The Shokz Opencomm UC bone conduction headset features an adjustable
good reviews, which is not so common with bone conduction
DSP noise-canceling boom microphone that also eliminates background
noise in loud environments to ensure clear voice transmission in various headphones.
work settings. A quick overview of the existing offerings in the bone conduction
space reveals an endless number of Shokz emulators from more
Sound Solutions in Austria—aka Sound Solutions International recognizable brands such as Philips and Orange (yes, the guitar
(SSI) owned consecutively by Philips and Knowles and now part amp company), to a significant number of smaller brands, including
of Foxconn’s FIT group, with production facilities in China, R&D Damson, WGP, Vidonn, Tayogo, Sandoo, AKASO, Sentien Audio,
in Austria, and offices in the US—is another potential supplier for Mojawa, and others—all offering very similar designs.
highly customized applications. Another emerging bone conduction application is swimming
In Asia, Os Electronic is a manufacturer best known for balanced earphones. Shokz has developed the OpenSwim waterproof design
armature drivers but also offers bone conduction transducers for for that purpose. Bone conduction is a viable solution for swimming
earphones. Reui Men (aka Billion Sound Electronic Technology) and many existing models are waterproof.
from Taiwan offers many off-the-shelf exciters with the smallest Bone conduction is also becoming common in smart glasses and
applicable to bone conduction earphones. AAC Technologies virtual reality (VR) devices. While effective for this application, the
provides advanced high-performance compact, low power, linear designs require a tight fit between the bone conduction transducers
haptic actuators with strong output. Other suppliers with haptic and the wearer, which glass frames usually do not have—or it
transducers include DLTech Audio, DMEGC Magnetics, and KOTL. wouldn’t be a comfortable experience. ax

Resources Sources
H. Werner Bottesch, “Bone-conductive stereo headphones,” AAC Technologies | www.aactechnologies.com
US Patent 5323468, https://patents.justia.com/patent/5323468
BeStar Electronics Co., Ltd. | www.bestartech.com
B. Håkansson, Chalmers,
Cochlear | www.cochlear.com
www.chalmers.se/en/staff/Pages/bo-hakansson.aspx
Dayton Audio | www.daytonaudio.com
B. Håkansson, “The balanced electromagnetic separation transducer
a new bone conduction transducer,” National Library of Medicine, DMEGC | www.chinadmegc.com
February 2003, https://pubmed.ncbi.nlm.nih.gov/12597176 GameChanger Products, LLC
M. Klasco, “Bass Shakers (Part 1): Enhancing the Deep Bass www.audioboneheadphones.com/product/audio-bone-1-0
Experience with Tactile Energy,” King Long Electromechanical Co., Ltd. | www.kotl.com.cn
https://audioxpress.com/article/bass-shakers-part-1-enhancing-the-
deep-bass-experience-with-tactile-energy Murata Manufacturing Co., Ltd. | www.murata.com

M. Klasco, “Bass Shakers (Part 2): Motion Platforms Are Nidec Corp. | www.nidec.com
Bass Shakers on Steroids,” https://audioxpress.com/article/ Ortofon | www.ortofon.com
bass-shakers-part-2-motion-platforms-are-bass-shakers-on-steroids
Piezo.com | www.piezo.com
M. Klasco, “Speakers and Transducers - Haptic Devices for
Augmentation, Entertainment, and Wellness,” RadioEar | www.radioear.us/products
https://audioxpress.com/article/speakers-and-transducers-haptic- Reui Men Co., Ltd. | www.billionsound.com
devices-for-augmentation-entertainment-and-wellness
Shokz | www.shokz.com
J. Mortensen, “Bone Conduction Technology Leading to Headphones
Sonion | www.sonion.com
In the Early 1800s Through the 20th Century,”
https://techevaluate.com/bone-conduction-technology-leading-to- Tectonic Audio Labs | www.tectonicaudiolabs.com
headphones-in-the-early-1800s-through-the-20th-century
Temco Japan | www.temco-j.co.jp
“What are bone-conduction headphones?” Philips,
www.philips.co.uk/c-e/so/sound-hub/what-are-bone-conduction-
headphones.html

16 | January 2023 | audioxpress.com

ax R&D Stories

Building a Haptic
Dimension
The Science Behind the
Taction Transporter

By
Silmon James Biggs and John Steinberg

Haptic motors are finding applications in gaming, music, VR/AR, and therapeutic/wellness.
The Transporter is a patented transducer that delivers faster more accurate bass and is
designed to work with the main audio driver in headphones. It subtly vibrates against the
user’s skin to produce an extra dimension that can be described as a high-fidelity extension.

Engineers and audiophiles talk about the So how can you get deep bass? Producing low-
spectrum of sound as spanning 20Hz to 20,000Hz. frequency acoustic sound at high volumes means
That’s only partially correct. Most adults can’t moving a lot of air. A good home theater subwoofer
hear above 14kHz. More importantly, humans in can have drivers that are 15” or more in diameter. A
general don’t actually hear 20Hz very well either. PA system for a stadium concert can have hundreds
But everyone—the deaf and hearing alike—can feel of them.
120Hz and 20Hz, and even lower. In fact, “hearing” When those giant speakers crank up, you don’t
deep bass is mostly feeling it. just hear them—you feel them. Your skin vibrates.
Bass perception happens along an audio-tactile Your organs vibrate. And (for most people) your
spectrum. The lowest notes are sensed primarily brain interprets that vibration as sound, even though
with the skin, and the highest primarily with the it mostly isn’t being processed by what you normally
ear, but all of them are naturally perceived as a mix think of as hearing.
of both senses. That’s a big part of the reason why Perceiving music is like perceiving food. We talk
it is so hard to get that “live sound” feeling with about tasting food, just as we talk about hearing
headphones—normal headphones can’t produce a music. But your sense of smell is an essential
critical aspect of the live experience. component of the enjoyment of a good meal. If
Does frequency response down to 30Hz, 20Hz, you have a cold or your sense of smell is damaged,
and below matter? The answer is “yes.” Rock, hip- food becomes bland and uninteresting. Just as the
hop, and EDM have huge amounts of energy at input from your nose makes a steak or a fine wine
those levels. more of a sensory delight, the experience of feeling
Jazz and orchestral music, plus movies and the bass is what makes going to a concert or dance
games do, too. If you’ve never heard your favorite club so much more powerful than listening to the
music on a system that can reproduce ALL the same music with only your hearing.
frequencies that are present in the recording, you Conventional headphones can’t deliver that
will probably be shocked to learn how much you experience—they’re like drinking a Cabernet while
have been missing. holding your nostrils closed. (The vast majority of

18 | January 2023 | audioxpress.com

home and car stereo speakers can’t do it either, Why Flat Frequency Response Matters
for that matter.) How loud should that note be? Musicians and
Audio drivers in most headphones are (at most) sound engineers try hard to get every sound
only 1” to 2” in diameter. Some headphones claim recorded at just the right level. If your speakers
to be able to produce sound down to 20Hz. They or headphones aren’t flat, you hear something
might even measure acoustically “flat.” But your
ears are very insensitive to deep bass. For a 15Hz
sound to be perceived as equally loud as a 75Hz
sound, it has to be 30 times louder—loud enough
to quickly cause serious hearing loss, in fact. So
headphones that measure “flat” aren’t perceived
as flat (Figure 1).
On the other hand, the human sensory system is
extremely sensitive to touch. And it is very sensitive
to vibration in exactly the frequency range where
hearing lets us down. So a 15Hz signal applied to
the skin (depending on the part of the body) only
needs velocity 2x higher than one at 75Hz to seem Figure 1: Sensitivity of hearing. At low level our hearing is most sensitive at frequencies
between 1kHz to 4kHz approximately, which is 1,000,000 (60dB) more sensitive than at
equally strong. The Figure 2 graph shows touch
bass frequencies close to 40Hz. Phons represent our hearing experience at different sound
versus hearing sensitivity and looks a lot like a
levels, described as equal loudness curves. At high sound levels we tend to experience all
crossover between a subwoofer and a satellite frequencies at a similar level. The sensitivity of our hearing reduces with increased level
speaker, doesn’t it? and so does the apparent frequency response.

Doing Haptics with Sound

When integrated into an over-ear headphone, the
Taction Transporter complements the main audio
driver by vibrating the skin to produce deep bass.
It does this by subtly moving the cushions of the
headphones against our head.
It seems simple, but if done wrong, that
movement creates a bunch of problems. Other
headphones have tried to do this before, with
disappointing results. One problem is that their Figure 2: Touch versus Hearing Sensitivity. One tact is the sensation of 1dB of normal skin
transducers weren’t as good as Transporters. (More vibration at 250Hz on the base of the thumb with a contact area of 2.9cm2, where 0dB is
about that later in the article.) 0.0002 m/s velocity. (Source Image: Curve derived from data in Verillo, IEEE Transactions
How you move the cushions matters too. Most Man Machine Systems, 1970)
tactile bass transducers to-date have tried to
move the headphone cup on the same axis as the
audio transducer works, which means moving the Figure 3: Taction
headphone cup toward and away from your skull. Transporters move the
headphone cups front-
Moving the earcups this way usually causes serious
to-back around the ear,
issues with the frequency response of the acoustic
reducing distortion and
driver, because it dynamically changes the volume transmitting vibration
of air inside the earcup. more efficiently. This was
Taction Transporters work differently. The the subject of Patent No.
transducers move the headphone cups in the US10,390,139, describing
sagittal plane–that is front-to-back along the hardware that can
surface of the skin (Figure 3). With this approach, accurately produce a wide
distortion is dramatically reduced. It is also a more range of tactile frequencies
at a perceptually constant
efficient way to transmit vibration to your skin. In
intensity, moving the
a properly tuned headphone, Transporters give you
skin in multiple axes, and
epic, subterranean bass without screwing up the using signal processing
acoustic response. But none of that would matter to enhance the user’s
if our transducer wasn’t the most powerful, flattest experience of audio
tactile driver ever produced. spatialization.

audioxpress.com | January 2023 | 19

ax R&D Stories

different from what they intended. (Note: There is ERMs have been used for years to make cellphones
some debate about what the ideal perceptually flat vibrate in silent mode.
response is for headphones but the disagreements A few companies have used them in things like
are pretty minor.) Flat frequency response matters game controllers and headphones (Figure 4). While
for tactile drivers, too. they can produce a range of frequencies, they have
It’s pretty easy to make a low-fidelity tactile bass a major limitation: There is a fixed relationship
driver. So some folks have offered headphones that between frequency and given power output. If you
use eccentric rotating mass (ERM) motors (Photo 1). want it to go louder, it has to rotate at a higher
frequency. If you want it to play a lower frequency,
it has to get quieter.
ERMs can only produce output along a single
line—there is no “area under the curve.” So the
range of possible outputs from an ERM is quite
small, and the effect won’t sound much like music
for most program material (Figure 5).
Another type of device that’s been tried in
headphones and other devices, including seat-
mounted rumblers, is called a Linear Resonant
Actuator (LRA). Like ERMs, these can be small and
inexpensive (Photo 2). But also like ERMs, they are
incapable of delivering high-fidelity bass response.
Photo 1: This is an example of an Eccentric Rotating Mass (ERM) motor.
The biggest problem with LRAs is in the name:
resonance.
Speaker and headphone designers work hard to
0.1 eliminate (or at least compensate for) resonances
Can only produce so they don’t color the sound. LRAs, on the other
Results on the line hand are all about resonance. That makes them very
velocity [m/s]

0.01 efficient at playing a single note. If an LRA resonates

at middle C on a piano, it can play middle C very
loud with very little power. But C sharp will be barely
0.001 audible, as will B natural. And that’s all folks—all the
other notes on the piano will be inaudible (Figure 6).
Some LRAs used in headphones are so resonant
0.0001 that they ring like crazy if you just tap the cup of
1 10 100 1,000 10,000 the headphone with your finger—even when the
frequency [Hz] headphones aren’t plugged in.
So how do Transporters compare to LRAs and
ERMs? Let’s start by comparing them in terms of
Figure 4: ERM motors can only produce a single volume level for each possible frequency.
frequency response (Figure 7). Taction Transporters
are flat ±4dB from 15Hz to 85Hz. They can
produce output at any level from the threshold of
0.1 perceptibility up to actually blurring your vision at
Can’t produce here
any of those frequencies. Wide frequency response
Can’t produce here is critical to good tactile performance. But that’s
velocity [m/s]

0.01 only one of several important parameters.

Power That Doesn’t Corrupt

0.001 As previously discussed, your skin includes
sensitive receptors for tactile signals. But there
is a threshold below which those signals will go
0.0001 unnoticed. If a headphone manufacturer goes
1 10 100 1,000 10,000 to all the trouble and expense of adding tactile
frequency [Hz] transducers to a pair of headphones, the system
had better be capable of producing enough tactile
Figure 5: ERM motors cannot play louder or quieter without also changing frequency. output to be felt.

20 | January 2023 | audioxpress.com

 on e
In - e th e b
ar dev o
ic e t e s t i n g – f a i t h f u l t

Bone conduction simulation – the game changer

REAL-LIFE-PROOF.
www.head-acoustics.de  �  
ax R&D Stories

Photo 2: Here is an example There are several different ways of measuring

of a Linear Resonant
power as delivered by a tactile transducer.
Actuator (LRA) driver.
(Headphone designers, take note: Your head-and-
torso simulator will not properly measure tactile
output. Tactile transducers don’t (intentionally)
produce changes in air pressure.) Taction offers
dedicated test equipment based on three-axis
accelerometers to measure tactile output.
You can measure tactile output as acceleration,
as velocity, or as displacement. The measurement
that best correlates with user experience, in our
opinion, is velocity. A transducer that delivers
constant velocity over a given frequency range will
be perceived by a user as providing flat frequency
response.
So how much velocity is required? The threshold
for perception of a tactile signal varies from person
to person, and is different for different parts of
your body. But a good baseline is 5 mm/second.
The ERMs and LRAs used in other headphones
that claim tactile output are severely challenged
on this score. The LRA used in one well-known
headphone produces a tactile signal above the
perceptual threshold only around the resonant
frequency. Unfortunately, when it’s doing this,
acoustically it loudly rings like a bell at about 75Hz.
An ERM could theoretically deliver strong output
at the top end of its frequency range, but (a) as
previously discussed, the only way to increase the
output of an ERM is to increase its frequency, which
obviously doesn’t deliver flat output; and (b) the
ERMs commonly used just aren’t that powerful. In
Figure 6: The frequency response of a LRA is shown here. These are so resonant that a typical headphone, the Transporter can deliver
power can drop 40% just 2.5Hz away from the resonant peak. at least 5mm/s of earcup velocity from 13Hz to
above 100Hz.

The Need for Speed

Audio drivers with quick rise and fall times
make music sound more dynamic, but rise time
is especially important for gaming. Saving a
few milliseconds of reaction time can mean the
difference between killing and being killed. There’s
research that shows that your body actually
processes tactile inputs faster than audio inputs.
But a faster path from skin to brain to action
does no good if it takes forever for the signal
to become strong enough to reach your skin’s
perceptual threshold.
Fast rise time is key. Fast fall times matter, too.
Three aspects of transducer design are critical to
achieving quick rise and fall times: power, mass,
and damping.
Figure 7: This graph shows a comparison of the low-frequency output among Taction ERMs have long rise times—on the order of
Transporters (blue area), Eccentric Rotating Mass motors (line), and Linear Resonant 150 milliseconds or more. LRAs rise faster than
Actuators. ERMs, but only at their resonant frequency. At other

22 | January 2023 | audioxpress.com

 Figure 8: A Taction
Transporter reaches 90% of
frequencies the rise time is effectively “never.” Even its potential output in just
so, it takes many cycles for an LRA to generate 8 milliseconds. For music,
enough oomph to deliver its strongest output. That’s that means the kick of a kick
because they tend to have a lot of mass relative to drum feels tight and natural.
the power of their motors. Transporters are a very
different story (Figure 8).
The LRA takes about 160 milliseconds to reach
90% of its potential output. That kind of delay is
easily noticeable. (Audio-tactile mismatch gets
noticeable when its around 40 milliseconds or larger.)
That’s bad for music. It’s even worse for gaming.
By comparison, a headphone, driven by a Taction
Transporter can reach 90% of its potential tactile
output in just 8 milliseconds. That means the audio
and tactile perceptions feel properly synchronized.
So how did we create a tactile transducer with
strong output and fast rise times? As Newton figured
out, force requires acceleration and mass. Taction
brings plenty of both.
To be effective, tactile transducers need to have Resources
mass—significantly more mass than other tactile Corsair HS60 Haptic, Stereo Gaming Headset with Haptic Bass, www.corsair.com
transducers have used. The Transporter’s moving
dB Enterprises, www.db-ent.com (Sales Contact for Taction: Dave Lindberg
mass weighs 18 grams. That’s 3 to 10 times more ([email protected])
mass than you would find in previous comparable
Taction Technology, www.tactiontechnology.com
transducers.

audioxpress.com | January 2023 | 23

ax R&D Stories

Figure 9: This graph To make that mass move quickly, we also need
shows a power handling
some serious acceleration and that requires a
comparison.
powerful motor. The motor in the Transporter is
more powerful than the motor used in any other
current tactile headphones we know of and 10 times
stronger than the most powerful potentially usable
ERM we know of (Figure 9).
More mass, plus more acceleration, equals
more force. Enough force (at the highest setting) to
literally blur your vision. (While most of us might not
want that, some folks do, and it provides abundant,
low-distortion headroom for the rest of us.) It
also makes Transporter transducers viable for a
number of other applications outside the context
of headphones.
The third critical factor in building a responsive
transducer is damping. Damping resists movement.
There are many ways to do this, and the properties
of damped systems can be very complex.
For purposes of our tactile transducer, you can
think of damping as a method for (a) reducing
resonances and (b) helping the transducer stop
moving, which improves fall times. Transporters
use patented damping techniques that give
unprecedented control, and excellent rise and fall
times (Figure 10).
That kind of fall time is critical to reproducing
Figure 10: Here’s a
music, and to making sure that sounds like
comparison between the explosions get out of the way of other important
fall time of the Taction cues in gaming. The result, enables the Transporter
Transporter and the LRA transducers to offer dramatically improved transient
used on a competitor’s response compared to all previous attempts at using
headphone. tactile transducers in headphones.

The Most Advanced Transducer

From the extensive research conducted by
Taction, we can finally describe the characteristics
of the ideal tactile transducer (Photo 3):

• Equally strong at all relevant frequencies—from

as low as 10Hz to as high as 100Hz or so
• Able to produce any frequency in that range

About the Authors

Silmon James Biggs, Ph.D, is the founder and CTO John Steinberg is the CEO of Taction Technology,
of Taction Technology, and the inventor of the Taction founding CEO of EcoFactor, Inc., a pioneer in the
Transporter haptic transducer. A full-stack tech guru, Internet of Things, Big Data, and the Smart
from mechanical design and digital circuits to software Home. John is a Silicon Valley, CA, veteran and
and everything in between, James invested years in a former musician (Berklee College of Music,
finding better ways to interface people with technology. Oberlin Conservatory). He worked in high-end
Fascinated by the tactile aspect of sound since he was stereo stores as a teenager and as an adult,
a kid, it drove him to study human biology at Stanford, treasures his 700+ LPs and vintage audio gear,
and do postdoctoral research at the MIT Touch Lab, some of which date back to the 1950s. John has
including work for DARPA on man-machine interfaces. He also helped also been issued more than 50 patents.
develop electroactive polymer transducers for Artificial Muscle, Inc.

24 | January 2023 | audioxpress.com

 at any output level—loud, soft, or anywhere in The simple electronics enable systems using
between standard audio components. More importantly, the
• Starts and stops on a dime Transporters are quiet, and extend the high-fidelity
frequency response all the way to 250Hz. ax
Already implemented in commercially available
products, such as the Corsair HS60 Haptic Stereo
Gaming Headset, Taction now sells off-the-shelf
haptic motors or offers a manufacturing license
agreement. The published specifications for the
Transporter transducer detail the ability to reach
the lowest tactile bass note of 12Hz with a velocity
of 5mm/s, and reach the highest tactile bass note
of 120Hz at that same velocity. Rise time is only
0.008 seconds, and fall time 0.090 seconds. The
transducer is able to move the headphones’ earcups
with a top speed of 50mm/s with minimal impact
on headphone acoustics. The intensity is adjustable,
maintaining a flat frequency response between 15Hz
to 85Hz (±4dB).

Overall Conclusion
We believe the Transporter haptic transducers
are able to offer an unmatched deep bass sensation.
They are fast, with an instant start and no lag, and Photo 3: The Taction Transporter transducer weighs 24g and the device package measures
they stop instantly, with no ringing. 47.5mm × 40.5mm × 8.0 mm.

203-502-7600

audioxpress.com | January 2023 | 25

ax Practical Test & Measurement

The Benefits of Using

Structure-Borne Sound in
Headset Signal Processing

By
Hans W. Gierlich

This article details the latest testing strategies for bone conduction technology used in
product development for headsets and in-ear devices. It showcases new possibilities
for manufacturers to better evaluate device performance, since no appropriate test
equipment and methods were commercially available until now.

The use of headsets in communication is headset microphone. The latest technology helping
increasing, and the demand for improving voice to improve this situation is bone conduction sensors,
quality in all remote communication continues especially in combination with in-ear headsets.
unabated. While in the past, mostly packaged and However, testing and optimizing such devices is
corded headsets were used with mobile phones, difficult since no appropriate test equipment is
today’s trend is toward advanced wireless Bluetooth commercially available.
headsets in a completely different price and quality To improve the situation, we complement
range. the HEAD acoustics Head-and-Torso Simulator
Nevertheless, a huge problem is still using such (HATS) HMS II.3 LN-HEC (with type 4.4 artificial
headsets in noisy situations while simultaneously ear according to ITU-T P.57 [1]) with an actuator
moving the headset microphones farther away from capable of generating structure-borne sound in
the talker’s mouth. This comfort for the user results the artificial ear. This way, a realistic simulation of
in a substantial decrease in signal-to-noise ratio the human structure-borne voice signal for in-ear
(SNR) for the talker’s voice signal captured at the headsets is possible.

Figure 1: Individual
structure-borne signals are
compared to the individual
air-borne signals at the
mouth reference point (MRP)
of each test person’s mouth.

26 | January 2023 | audioxpress.com

 This article provides background on human
structure-borne sound measurement and simulation;
on potential benefits of using structure-borne sound
as an additional input signal for headset signal
processing; and testing strategies and test results.

Human Structure-Borne Sound—

Measurement and Simulation
The first step in headset testing with structure-
borne sound is understanding the human structure-
borne sound transmission of the user’s own voice to
the user’s ear, where the bone conduction sensors of
headsets are most likely placed. For this purpose, we
implemented a structure-borne sensor in an in-ear Figure 2: An average structure-borne signal is compared to the average air-borne signal
headset mockup and performed measurements at the MRP.
on a variety of human individuals. The goal is to
measure the individual spectra of the air-borne
and structure-borne sound. These measurements
provide an overview of the spread of individual
differences and are the basis for deriving average
transfer functions.
Furthermore, we received information on
whether there might be significant differences
between males and females. As part of the testing,
we conducted measurements with 11 female and
24 male subjects. Figure 1 shows the individual
structure-borne signal compared to the air-borne
signal at the mouth reference point (MRP). Figure 2
shows the average of these signals for female and
male talkers. The difference shows mainly in the
frequency range below 300Hz, where female voices Figure 3: This is a comparison of simulated vs. human average structure-borne sound.
generally do not provide any signal energy due to
their higher fundamental frequency. Consequently,
we chose the average male spectrum as a target
spectrum for structure-borne sound simulation.
The next step is simulating human structure-
borne sound. We extended the artificial head
HMS II.3 LN-HEC with an actuator, reproducing
the average human structure-borne sound in the
artificial ear precisely at the location where we
conducted the human measurements. Figure 3
shows the spectrum of the simulated structure- Figure 4: These are signals at a near-end noise canceller.
borne sound of the new LN-HEC ear compared to
the average human structure-borne sound. Please
note that, due to the insufficient SNR of the sensors
available today, structure-borne sound simulation
currently can only be measured reliably up to 3kHz.
As seen in Figure 3, the simulation is correct within
a tolerance of ±3dB.

Headset Signal Processing and

Potential Improvements with
Structure-Borne Sound
The primary purpose of using structure-borne
sound in headset signal processing is the improved Figure 5: Test sequence A and the corresponding analysis is shown here.

audioxpress.com | January 2023 | 27

ax Practical Test & Measurement

separation between a near-end talker and near-end

background noise, a near-end talker and a near-
end impairing talker, and a near-end talker and a
far-end talker (double talk detection).
A second benefit of using structure-borne sound
could be transmitting the structure-borne signal
instead of the near-end microphone signal, at
least in the low-frequency domain. The structure-
borne signal is almost free of background noise
and, therefore, ideally “noise-canceled.”
Figure 4 illustrates the signals received at a
noise canceller. The separation of near-end speech
and impairing signals can significantly improve with
Figure 6: Noise spectra is shown in an uplink with and without bone conduction simulation a structure-borne signal. What can be expected in
Device 1, background noise: public noise from ETSI TS 103 224 simulated with 3PASS [2], [4]. terms of enhanced signal processing? The noise
canceller is an adaptive filter that estimates the
background noise and tries various strategies to
reduce background noise by assessing its signal
power and trying to minimize the noise level. The
better the noise estimation, the better cancellation
works.
Separating noise and speech, however, is
complex. The noise canceller may diverge, and
the speech signal may get degraded. When able
to distinguish better between speech and noise
based on structure-borne sound, the adaptation
control can be much improved. Adaptation can be
frozen or adaptation speed reduced, and the noise
canceller’s divergence may be avoided.
Figure 7: Noise spectra is shown in an uplink with and without bone conduction simulation
Consequently, a higher-quality speech signal can
Device 2, background noise: public noise from ETSI TS 103 224 simulated with 3PASS [2], [4].
be provided in conjunction with low background
noise. Similar behavior can be expected with
concurrent talkers. Assuming only the voice of the
talker wearing the headsets should be transmitted,
the concurrent talker can be treated as background
noise.

Figure 8: Test sequence A is shown with a concurrent talker and the corresponding analysis.

About the Author

Hans W. Gierlich started his professional career in 1983 at the
Institute for Communication Engineering at RWTH, Aachen. In
February 1988, he received a Ph.D. in electrical engineering. In 1989,
Hans joined HEAD acoustics GmbH in Aachen as vice-president. Since
1999, he has been head of the HEAD acoustics Telecom Division and
in 2014, he was appointed to the board of directors. Hans is mainly
involved in acoustics, speech signal processing and its perceptual
effects, QOS and QOE topics, measurement technology, and speech transmission and
audio quality. He is active in various standardization bodies such as ITU-T, 3GPP, GCF, IEEE,
TIA, CTIA, DKE, and VDA. He is vice chair of the European Telecommunications Standards
Institute (ETSI) Technical Committee for “Speech and Multimedia Transmission Quality”
and was chair from 2016–2020. In 2021 he received an ETSI fellowship award. Figure 9: Two positions of a concurrent talker are shown
here.

28 | January 2023 | audioxpress.com

AUDIO FOUNDRY A
DSP CONCEPTS | TYMPHANY
COLLABORATION

Featuring the Rig™ powered by Audio Weaver®

purpose-built labs for automotive audio prototyping

• Early concept validation

• Accelerated development
• Experiential design flexibility
• In-vehicle concept presentation
• Fully equipped workspaces
• Dedicated prototype vehicle
• Onsite audio experts

Learn more at audiofoundry.dspconcepts.com

ax Practical Test & Measurement

Measurement Results of Actual is placed in the test room, and the headsets are
In-Ear Headsets carefully fitted to the HATS. The speech signals used
We performed measurements of various in all experiments are based on Recommendation
headsets based on this fundamental research. ITU-T.P501 [3]. Since measurements in steady-
The test setup is a typical laboratory-based setup state conditions likely won’t show any differences,
using a realistic sound field simulation with eight we created specific test sequences targeting
loudspeakers and pre-recorded background noises, performance differences with bone conduction
as described in ETSI TS 103 224 [2]. The HATS simulation.
Test sequence A, shown in Figure 5, focuses
on the potential divergence of the noise canceller,
which may happen if suddenly a background noise
situation occurs. In this situation, speech and noise
may not be separatable anymore, and speech might
get degraded while the noise is not sufficiently
canceled.
Measuring the remaining background noise
directly after the speech signal has stopped provides
a good idea of the noise cancellation performance.
The amount of noise reduction indicates the noise
canceller’s efficiency in converging reliably even
while speech is present. Two example results are
given in Figure 6 and Figure 7.
The differences in the remaining noise spectra
Figure 10: Speech spectra of the concurrent talker is shown in an uplink with and without with and without structure-borne sound are
bone conduction simulation, concurrent talker in Position 1. apparent, as well as the difference between the

30 | January 2023 | audioxpress.com

headset performances. While Device 1 provides an
improvement of up to 30dB just in the frequency
domain up to 500Hz, Device 2 reduces the
background noise by up to 30dB over the complete
speech spectrum transmitted.

Experiment B
A similar test sequence can be used to evaluate
the performance with a concurrent talker present.
For this test, no background noise is used. Instead, a
second HMS simulates the concurrent talker. The test Figure 11: Speech spectra of the concurrent talker is shown in an uplink with and without
sequence for this experiment is shown in Figure 8. bone conduction simulation, concurrent talker in Position 2.
You can use various talker positions in this
experiment. In this article, we discuss one
specifically interesting combination. Figure 9
shows two concurrent talker positions. In Position
1, the artificial head faces the talker wearing the
headsets directly. In Position 2, the concurrent talker
is talking in the same direction.
The test results for these two conditions are
pretty interesting. While for Device 1, neither
a difference between the two positions of the
concurrent talker nor a difference between bone
conduction simulation nor the lack thereof could
be observed (Figure 10). Figure 12: Test sequence B and the corresponding analysis are shown here.

audioxpress.com | January 2023 | 31

ax Practical Test & Measurement

The result for Device 2 is entirely different. any attenuation. When turning 90 degrees, Device
Talkers, facing the wearer of the headset and 2 differentiates that the second talker may not be
talking in their direction, are transmitted without talking to the headset wearer anymore and seems to
be treated like background noise if bone conduction
Part A Part B is present (Figure 11). The attenuation is about
Bone Conduction 10dB.
Off On Off On
Simulation
G-MOS 2,4 2,4 2,1 2,5 Experiment C
N-MOS 3,6 3,6 3,7 3,6 A third test used the following test sequence
S-MOS 2,8 2,9 2,5 2,8 shown in Figure 12. This test focuses on the
noise canceller’s potential divergence if speech
Table 1: S-, N-, and G-MOS results for Device 1 with background noise is presented after background noise. Here we
focus on the speech signal (S-MOS) deterioration
Part A Part B and a change in noise reduction performance.
Bone Conduction
Off On Off On
A perceptually motivated procedure for such
Simulation evaluations is 3QUEST [5], as per ETSI TS 103 281
G-MOS 2,2 2,6 2,1 2,5 [6], which allows for measuring these differences.
N-MOS 3,8 3,8 3,9 3,6 The test results (Table 1 and Table 2) for both
S-MOS 2,6 3,1 2,4 3,0 devices show a clear improvement in speech quality
(S-MOS) with bone conduction sensoring. In both
Table 2: S-, N-, and G-MOS results for Device 2 in the presence of background noise devices, the improved speech detection allows for
better control of the noise canceller. The speech
signal is less deteriorated, leading to higher speech
quality in the presence of background noise. This
way, devices with bone conduction sensors can be
tested and optimized reliably in the lab to tune the
noise canceling algorithms best possible.

Experiment D: Double Talk

A n ot h e r im p r ove m e nt wit h s t r u c t ur e -
borne sensing can be expected in double-talk
situations. In almost any communication device,
echo cancellation prevents the far-end listener
from hearing the echo produced at the near-
end. In general, the echo canceller (Figure 13)
is an adaptive filter modeling the echo path and
Figure 13: The fundamental functionality of an echo canceller and the associated signals
are diagrammed here. trying to minimize the echo signal by subtracting
the inverse echo signal. The echo canceller may
diverge if the impairing near-end speech signal
References is not reliably detected. The adaptation control
[1] Recommendation International Telecommunication Union (ITU) ITU-T P.57 quality is vital, and the better the echo can be
(06/21): Artificial Ears separated from the near-end talker signal, the
[2] European Telecommunications Standards Institute (ETSI) ETSI TS 103 224: better the echo cancellation will work in case of
Speech and multimedia Transmission Quality (STQ); A sound field reproduction double talk. Using the bone conduction signal may
method for terminal testing including a background noise database improve this separation since it represents solely
[3] Recommendation ITU-T P.501: Test signals for use in telephony and other
the near-end speech signal.
speech-based applications
Conclusions
[4] HEAD acoustics datasheet: 3PASS Structure-borne sound simulation is crucial
[5] HEAD acoustics datasheet: 3QUEST when testing modern headsets. Only with the
complementing simulation of air-borne and
[6] European Telecommunications Standards Institute (ETSI) ETSI TS 103 281: structure-borne sound generated by the human
Speech and multimedia Transmission Quality (STQ); Speech quality in the voice can it be possible to evaluate and optimize
presence of background noise: Objective test methods for super-wideband and
fullband terminals
headsets in different conversational situations
comprehensively. ax

32 | January 2023 | audioxpress.com

 ax Sound Control

Acoustical Test Sources

By
Richard Honeycutt

The choice of a sound source for acoustical testing and reporting is decisive and should be
determined according to the applications. Let’s explore the characteristics and benefits of
available sources for acoustical measurements.

Acoustical testing can be classified according to the test sound field as a balloon whose shape and coloration represent the
setup: Response testing involves a sound-level meter (SLM) and/ sound pressure level (SPL) radiated in each direction.
or a real-time analyzer (RTA) to measure the sound level and/ The third way is a section drawing with a vertical or horizontal
or spectrum produced by a source. Stimulus-response testing cut through the balloon. Such sections often show multiple lines,
incorporates an acoustical source whose sound output is known, illustrating the variation in directivity with frequency.
and measurements of the resulting sound field are made in order
to determine how a room or other enclosure affects the sound. Source Directivity
Sound-system performance, reverberation, and echo testing Figure 1 shows the circular plot characterizing horizontal and
exemplify this type of test. vertical directivity of an omnidirectional source. Of course, the
The second type of testing requires a source with a certain balloon plot would be a perfect sphere. Figures 2–5 show the
directivity (polar response) and frequency response.
The directivity of an acoustical source depends upon
its physical characteristics and size.
A source that is small compared to the
wavelength of the sound being investigated is called
a point source and is omnidirectional. Directivity can
be represented in any of several ways.
One way is a simple graph of the beamwidth
plotted versus frequency. The directivity graph of
an omnidirectional source would be very boring: a
horizontal line showing a beamwidth of 360° at all
frequencies. Another way is to represent the radiated

0º

–5dB

–10dB

–15dB

–20dB

–25dB

270º 90º

180º
Figure 2: The horizontal (upper) and the vertical (lower) directivity of the human voice
vary with frequency; the standard deviation is indicated by the colored shading. (Image
.Figure 1: The horizontal and the vertical directivity of an Source: Christoph Pörschmann, Johannes M. Arend: “Analyzing the Directivity Patterns of
omnidirectional source are perfect circles. Human Speakers,” www.researchgate.net/publication/343344151)
.
.
34 | January 2023 | audioxpress.com
directivities of other common sources. For most of For line arrays, the horizontal directivity is
these sources, at lower frequencies, the directivity typically more-or-less constant at 120° to 150° with
would be less (the “watermelon” shape of the balloon some narrowing at the highest frequencies, and the
smaller in diameter); at higher frequencies, greater. vertical directivity varies from omnidirectional at the
lowest frequencies to some essentially constant value
(often 10° to 30°) from the lower critical directivity
frequency—at which the array height is about ¼
wavelength—to the upper frequency response limit.
For stimulus-response testing of an installed
sound system used for reinforcement of speech
or singing, a speaker having a reasonably flat
frequency response and a polar response similar to
that of the human voice can be used. The advantage
of feeding a vocal mic using a speaker, rather than
simply feeding pink noise to the system is twofold:

• The acoustical characteristics—frequency

response and directivity— of the vocal mic,
which affect the gain-before-feedback, are
included in the test.
Figure 3: the directivity balloon of an enclosed two-way
8” cone speaker at 3.15kHz shows strong beaming.
• The acoustical environment—stage-monitor
leakage and reflections from a lectern table or
music stand—are included.

Figure 5: The directivity

model of a line array
speaker at 3.15kHz is shown
by the balloon (upper)
Figure 4: The directivity of a 50-degree horn speaker at and the vertical and the
3.15kHz is shown by the balloon (upper) and the vertical horizontal balloon sections
and the horizontal balloon sections (below). (below).

audioxpress.com | January 2023 | 35

ax Sound Control

Photo 3: The NTi Talkbox is a high-

SPL test source with flat frequency
Photo 1: The B7K mouth simulator can be used to response that simulates the human
simulate a human-mouth acoustical source. voice directivity.

Photo 2: Another source that simulates a human

mouth is the B&K Head-and-Torso Simulator (HATS).

Two excellent examples of a test source that can be

used in this fashion are the Brüel & Kjaer (B&K) mouth
simulator (Photo 1) and the Head-and-Torso Simulator
(HATS) (Photo 2). The HATS provides more accurate
directivity compared to a human head. The NTi TalkBox
is a less expensive option (Photo 3). The polar directivity
plots are shown in Figure 6.
For measuring reverberation time (RT ), an
omnidirectional source should be used because the concept
of RT as developed by Wallace Sabine assumes perfect
diffusion of direct and reflected sound in the room being
tested. No real point-source speaker small enough to be
omnidirectional in all octaves would be able to produce
a high enough SPL to overcome the background noise.
The most common RT measurement is RT30, which is
twice the time required for sound to decay by 30dB. For
accurate measurements, the background noise must be
at least 30dB—preferably 40dB or more—below the test
signal, if pink noise is used, as is usually done. (Swept-
sine measurements can take advantage of synchronous
averaging to cancel out some of the background noise,
but not many swept-sine RT analyzers are available.)
This means that test signal levels ranging from 60dBZ
to 90dBZ (flat weighting) may be required, depending
upon the room noise level. In order to achieve these signal
levels, test speakers using multiple drivers are employed.
To provide uniform radiation in all directions, 12 identical
speakers are often used, mounted in a dodecahedral
Figure 6: The directivity of the Talkbox in the critical 1kHz to 4kHz octave bands is configuration. The drivers are fed an equalized pink noise
close to that of the human voice. signal. The resulting frequency response is not truly flat

36 | January 2023 | audioxpress.com

in terms of energy per Hertz, but it can be made flat in terms of be very important. Such a room would have equally absorptive
energy per octave. As sound guru Pat Brown has said, “the object finishes on all surfaces, and no shapes that could cause focusing or
here is to excite the room, not to sound good.” block sound from certain areas. Most real rooms have one or more
characteristics that prevent the dispersion from being uniform:
Onmi Sources
Several well-known manufacturers of acoustical test equipment • Ceilings that are more absorptive than the other room
make excellent dodecahedral test speaker (“dodecs”). Examples boundaries (e.g., acoustical tile ceilings in classrooms and
are the NTi DS3 and the B&K Model 4292L, both of which provide conference rooms
omnidirectional sound at levels somewhat above 120dB. The
Italian company Outline has recently introduced the Global Source
Radiator (GSR) with a companion radio-remote-controlled white/
pink noise generator, 1kW digital amplifier, and other measurement
accessories (Photo 4).
Commercially available dodecs typically range in diameter
from 31cm to 39cm, and weigh from 8kg to 11kg. Most of them
are truly dodecahedral in form. The Outline GSR is unique in
that it is spherical, reducing diffraction and edge reflections that
could cause response irregularities. Good commercial dodecs are
expensive, so many audio amateurs and a few audio pros have
decided to build their own. A DIY dodec article written by Pat Brown
is available online (see Resources).

The Importance of Omnidirectionality

For RT measurements in a room with perfectly uniform Photo 4: The Outline HSR dodec and its companion amplifier provide
dispersion of sound, using an omnidirectional source would not omnidirectional radiation at over 130dBSPL.

audioxpress.com | January 2023 | 37

ax Sound Control

• Most of the total absorption provided by an

audience located near the floor (auditoria)
• Highly asymmetrical shapes (long, narrow
rectangular rooms or rooms whose plan-view
form is irregular)
• At least some glass walls (many modern
conference rooms)
• Opposing acoustically reflective walls of glass,
gypsum, or masonry
• Wall or ceiling areas that focus sound (concave
wall or ceiling sections)

In such rooms, an omnidirectional source will

provide more reliable RT measurements in some
or all octave bands than will a directional source.
As an example, consider the cruciform church
sanctuary model shown in Photo 5. The tan-colored
floor areas represent pews, two-thirds of which
are occupied by people, meaning that most of
Photo 5: This computer model of a cruciform church sanctuary has several features that the acoustical absorption is near the floor. The
call for the use of an omnidirectional source for measuring RT.
longitudinal and transverse arms of the sanctuary
are long and narrow. The peaked ceiling reflects
sound toward the center.
Figure 7 shows the RT versus frequency using an
omnidirectional source. At 500Hz, the RT at various
points in the room ranges from a minimum of 7.78
seconds to 7.87 seconds. Figure 8 shows the RT of
the cruciform sanctuary measured using a full-range
horn as a test speaker. This speaker’s significant
directivity begins at about 200Hz, reaching 50° ×
50°at 1kHz and above.
In the computer model, the speaker is aimed at
the rear center seating section. At 500Hz, the RT at
Figure 7: The RT versus frequency (octave bands) measured using an omni source is various points in the room ranges from a minimum
shown above. The 500Hz RT is 7.72 seconds. of 7.84 seconds to 8.12 seconds. In this room, using
a directional speaker rather than an omni as the
test source results in negligible difference in the
measured RT.
Resources Modifying the room model by adding 2-1/8”
P. Brown, “Build Your Own Dodecahedron Loudspeaker,” SynAudCon, March 2010, cloth-wrapped fiberglass panels on the rear walls
www.prosoundtraining.com/2010/03/17/build-your-own-dodecahedron-loudspeaker and the end walls of the “crossbar” introduces more
inhomogeneity in the boundary absorption of the
“Technical Data TalkBox,” NTi Audio,
room, and changes the RT range to 4.29-5.03 seconds
www.nti-audio.com/Portals/0/data/en/TalkBox-Specifications.pdf
for the spherical source, and 4.17-5.71 seconds for
Sources the horn source, a difference of about 9% in average
RT. So the difference in RT measurement results is
High-Frequency Head and Torso Simulator | Type 5128 greater when the room is less diffuse. In such a
Brüel & Kjær | www.bksv.com room, measurement errors resulting from the use
of a directional source can be significant.
Mouth Simulator | B&K 4227
Normally, the sound source for RT measurements
Brüel & Kjær | www.bksv.com
is placed at the location from which the sound
NTi TalkBox normally emerges when the room is in use. For
NTi Audio | www.nti-audio.com example, in an auditorium, the source might be
placed about head height on the stage. In a simple

38 | January 2023 | audioxpress.com

rectangular room, placement of the source at a
corner can excite up to eight times as many room
modes as center placement.
In any case, the dist ance bet ween the
microphone and the sound source must be
significantly larger than the critical distance of the
room—the distance at which direct and reverberant
sound pressures are equal—given approximately
by: DC≈0.1(V/RT)1/2 where V = room volume in m3.

Overall Impressions
So you need to pay close attention to the
performance of an acoustic test source any time Figure 8: The RT versus frequency (octave bands) measured using a 50°×50° horn source
is shown above. The 500Hz RT is 7.72 seconds.
you make a stimulus-response measurement.
Because having a flat frequency response matters.
About the Author
If you are measuring RT for design purposes, the
Dr. Richard Honeycutt fell in love with acoustics after his father
people, clothing, and number of occupants introduce
brought home a copy of Leo Beranek’s landmark text on the subject
so much variability in a venue’s actual use that when Richard was in the ninth grade. Richard is a member of the
precise measurements of the empty venue are North Carolina chapter of the Acoustical Society of America. Richard
often not important enough that a omnidirectional has his own business involving musical instruments and sound systems.
source must be used. However, if you are making He has been an active acoustics consultant since he received his PhD
in electroacoustics from the Union Institute in 2004. Richard’s work
measurements to verify compliance with a standard,
includes architectural acoustics, sound system design, and community
it is best to use an omnidirectional source. ax noise analysis.

Celebrating 45 Years!
ACO USTICS BEGINS WITH ACO™
MEASUREMENT
Polarized
& MICROPHONES
Electret <1Hz to >120kHz
Windscreens
3 to18 inch &
Custom
Weather/UV
Resistant
– that is the questionTM

Community/Industrial
Noise Monitors/Alarms
Indoors and Out

Microphone Systems
Standalone
IEPE Powered
Phantom Power
Custom
Tel: +1-650-595-8588

Serving End-Users and OEMs Worldwide

audioxpress.com | January 2023 | 39

ax Market Perspective

OTC Hearing Aids

Is Innovation
Leaving the Rule Behind?
By
Andrew Bellavia

A wave of new developments promises to solve the most pressing need for people with milder hearing
loss, allowing mass-market hearables to supplant over-the-counter hearing aids with less stigma.

To much fanfare in the US, the Federal Drug Administration’s An unstated premise is that a device’s core functionality is
(FDA) over-the-counter (OTC) hearing aid rule took effect in October determined by design and therefore readily identified as a hearing
2022. Devices are now available both in stores and online. This is aid out of the box. The FDA barely recognized the possibility that
the culmination of a process that properly began five years earlier additional software might change the nature of a device, considering
with the passage of the law requiring the FDA to create such a mainly that a traditional hearing aid might be given self-fitting
category for self-perceived mild to moderate hearing loss. Five capability later on by unlocking or downloading a program.
years being an eternity in the wearables world, in a March 2022 Therefore, two basic principles undergird the OTC hearing aid rule:
article [1], I suggested that the assumptions used to write the rule
were becoming obsolete even as the process was concluding. A • Selective amplification is needed to ameliorate mild to
fresh look at developments illustrates how quickly this may come moderate hearing loss.
about. CES will be a showcase for the initial suite of OTC hearing • Such amplification must be provided for at the time of
aids, while at the same time providing clues to the coming post- manufacture, even if a self-fitting component is activated by
OTC revolution. software afterward.

Ground Shifting Under Our Feet Both of these will prove untrue, bypassing the OTC rule in
What we typically think of as a “hearing aid” is a device meant different ways and opening the door to increased innovation in
to compensate for damage to, or deterioration of, the cells of hearing devices.
the inner ear that convert sound into impulses sent to the brain.
Damage is usually selective, causing the ear to be more or less Amplification and the Cocktail Party Problem
sensitive to sound at different frequencies. To compensate, a People with more severe hearing loss require selective
hearing aid can be programmed by a hearing care professional amplification to understand other people even in quiet settings.
to provide varying levels of amplification at different frequencies Others whose hearing loss is milder, or even nonexistent according
according to an individual’s hearing loss profile. A modern hearing to a standard hearing test, often hear fine in quiet settings but have
aid is a sophisticated device, providing more control than just the difficulty in loud restaurants or pubs. This “cocktail party problem”
amount of amplification, but selective amplification remains the is one of the most difficult because the noise is created by other
core feature (Figure 1). voices. The ability to discriminate between the voice one wants to
The new FDA rule defines an OTC hearing aid in the same hear and all the others is one of the most sophisticated functions
way, specifically as a device that “must allow the user to cause of the entire auditory system from ear to brain. It doesn’t take
frequency-dependent changes based on the user’s preference” [3]. much to impair this ability, which is why a significant number of
An OTC hearing aid also provides control over the amplification people without measurable hearing loss report difficulty hearing
profile but with settings determined by the end user rather than in noise (Figure 2) [4],[5].
a hearing care professional, either through a hearing self-test or That leaves open the question if amplification is even necessary
choice of profiles. for people who are challenged only in loud situations. All that

40 | January 2023 | audioxpress.com

is lacking is an effective way to separate nearby
speech from the din. The company that solves
this problem will bypass hearing aid regulations
entirely, opening up the possibility of offering
“restaurant mode” in mass-market true wireless
stereo (TWS) earphones. Even Apple has moved in
this direction, adding by software update what it
calls “conversation boost” (directional microphones)
and ambient noise reduction to AirPods Pro. With up
to 7dB of signal-to-noise ratio (SNR) improvement in
hearing [6], these modes offer meaningful benefit
in a popular consumer device (Figure 3). And that
is just the beginning.
Traditional hearing aids improve speech in
noise performance in similar ways to AirPods Pro,
using a combination of directional mics and noise
filtering. Referring to Figure 1, hearing aids are more Figure 1: Block diagram of a typical hearing aid processor [2]
sophisticated in that they perform amplification, Figure 2: The estimated US
compression, and noise reduction on per-channel adult population with self-
basis to squeeze out higher SNRs than is possible reported hearing difficulty
operating on the entire spectrum as a whole. In and no audiometric hearing
addition, sound scene analysis dynamically controls loss (box C) is shown here.
the filter bank as the auditory situation changes.
Because the noise profile in a restaurant is mainly the
babble of other voices, there are limits to the efficacy
of acoustic filtering, even under dynamic control.

A Solution from the Broadcast World

Those in the broadcast world are familiar with
machine learning-based noise reduction through
programs, such as Krisp. These work on a completely
different principle. Rather than performing acoustic
filtering, machine learning systems are trained to
separate speech from noise and pass on only the
speech. Often a deep neural network is used. This
is a method of analyzing an input similar to the
way the human brain works. Just as a person can
learn to recognize the difference between a cat
and a dog from life experiences, a deep neural
network can learn the difference between speech
and noise by being trained with speech recordings
in different scenarios. The result is an inline system
for separating out the speech and sending it
through without the noise (Figure 4). Krisp wasn’t
even launched until 2018 and still requires a PC’s
processor and resources to create the deep neural
network in software for running its noise reduction
program. That’s a far cry from implementing such
a system in-ear. The situation is changing rapidly,
however, and the effects will be revolutionary.

ML Noise Reduction Goes Portable

A company called Whisper was the first to apply
machine learning (ML) deep neural networks to the Figure 3: Hearing SNR improvement is illustrated with the AirPods Pro hearing features
hearing device case, launching its first product turned on.

audioxpress.com | January 2023 | 41

ax Market Perspective

in 2020. It was still not possible to run the neural

network in-ear, so the Whisper system uses a separate
“Brain,” about the size of a mobile phone and using
a similar processor. The hearing aids themselves
are of a classic design and work without the Brain.
When the Brain is turned on, received audio is passed
from the hearing aids to the Brain, de-noised, and
returned at very low latency. Not only does the use of
an external device allow enough processing power to
run the neural network locally, it is also updatable. As
of this writing, Whisper has delivered its fifth major
software update to its customers (Figure 5). It is not
likely that such a system would gain mass-market
acceptance, however. Consumers will balk unless
the entire system is in-ear and provides noticeable
Figure 4: This shows a deep neural network with N hidden layers [7]. benefit. This is where it gets interesting.

A Neural Network in Your Ear

With today’s technology, typical processors that
Figure 5: A Whisper hearing can be incorporated into a hearable do not have the
aid uses a separate “Brain,” resources to run a deep neural network, especially
about the size of a mobile considering limitations on allowed power consumption.
phone and using a similar That is poised to change as companies develop
processor. (Image Source: processors specifically designed to address this
Whisper)
need. Greenwaves Technologies is sampling its GAP9
processor with a nine-core RISC-V compute cluster
that they claim “is perfectly adapted to handling
combinations of neural network and digital signal
processing tasks delivering programmable compute
power at extreme energy efficiency” [8]. Greenwaves
has been demonstrating noise reduction and other
applications on the GAP9, and has reserved a suite
at CES.
Femtosense has taken a different approach by
designing a neural network in hardware. It uses what
they call a “sparse” architecture, activating only those
nodes in the network that provide useful results at
any given moment. They claim significantly reduced
power consumption as a result, since only a fraction of
the notes are turned on at any given time (Figure 6).
Although the Femtosense chip could be implemented
as a stand-alone co-processor, it is more likely to find
a home as an IP block within a multifunction SoC for
the sake of efficiency. Having already demonstrated
its de-noising algorithm running on a development
board at CES 2022, it will be most interesting to see its
progress at the 2023 event. Solutions such as these
will bring hearables one step closer to providing
broadcast quality denoising in-ear, potentially
revolutionizing how consumer devices address the
cocktail party problem without amplification.

Evolving Toward Smart Hearables

Figure 6: This is an illustration of a Femtosense processor showing active nodes at a given Hearable devices have been following the same
moment. (Image Source: Femtosense) trajectory as mobile phones, in three stages. The

42 | January 2023 | audioxpress.com

first stage was the introduction of mobile phones end solution fuses our advancements in attention
whose sole purpose was to originate and receive and audio software, sensors, hardware, and design”
calls. Next came feature phones with more advanced in a modular platform (Figure 10). It is not hard
functions, such as email and web browsing to imagine how beamforming mics and ML-based
preprogrammed. Finally, with the release of the denoising algorithms could be enhanced by knowing
original iPhone, came the app store ecosystem. By which person the user is attempting to hear.
democratizing access to what amounted to a pocket- Future hearable apps could detect the microphone
sized computer, innovation exploded. Today, we and sensor capability of a particular device and
are at the “feature hearables” stage, but the same configure itself for optimum performance. Much
progression is about to happen, with important like picking out a smartphone based on a mix of
consequences in the hearing space. hardware features and performance (think screen
One company working to bring a true operating resolution, camera performance, and sound quality),
system to hearables is Bragi, which pioneered true
Figure 7: This illustration
wireless devices in the previous decade and more shows a Bragi in-ear app
recently pivoted solely toward developing its OS. ecosystem. (Image Source:
Today, Bragi offers a range of apps that can be Bragi)
pre-configured at the factory or sold afterward
through an app store. Though the complexity of
the apps is limited by the processing power of the
supported SoCs from Airoha and WuQi, a glimpse of
the future is seen by the inclusion of Mimi Hearing
Technologies’ app for personalizing streaming
audio playback to one’s individual hearing profile
(Figure 7).
While Bragi is working with existing SoC suppliers,
Sonical is taking a different approach by developing
its own processor (Figure 8). Optimized to run its
OS, Sonical claims it will “provide the world’s first ear
computer that has the capabilities and performance
for next-generation ear worn products.” In a webinar
hosted by Danish Sound cluster, Gary Spittle,
founder and CEO, promised that the “world’s first
ear computer platform” will launch at CES, and that
“many plugins are already available.”

A New Wave of Innovation

By providing a platform for software developers
to innovate without having to create corresponding
hardware, smartphones have opened up whole new
worlds. Think about all the smartphone apps you use
and imagine a similar ecosystem running natively
in your true wireless earphones.
Coming back to the cocktail party problem, there
are companies working to mimic the brain’s ability Figure 8: Sonical’s hearable processor and app ecosystem wants to provide ’the world’s
to concentrate on and filter the desired speaker first ear computer.” (Image Source: Sonical)
even more closely. The smart hearable ecosystem
will provide the means to bring these innovations About the Author
to practical use. Prior to founding AuraFuturity, a marketing consulting company
In the same webinar at which Sonical teased focusing on in-ear and hearing, Andrew Bellavia had experience in
its CES presence, Neurotech software company international sales, marketing, product management, and general
management. Audio has been both an abiding interest and a
Segotia discussed research performed at KU Leuven market he served professionally in these roles. Andrew has been
(Figure 9) as basis of its own work to detect the deeply embedded in the hearables space since the beginning and is
specific voice being attended to by measuring EEG recognized as a thought leader in the convergence of hearables and
signals [9]. EEG sensor developer AAVAA also sees hearing health. He has also been a strong advocate for hearing care
a future in hearables, highlighting that its “end-to- innovation and accessibility. Andrew’s home base at CES will be the ALTI suite, Lido 3101B.

audioxpress.com | January 2023 | 43

ax Market Perspective

a consumer will be able to choose a hearable with Regardless of Approach, Consumers Win
optimum performance for their primary use cases, Even with techniques such as de-noising speech
knowing they will still work with a wide range of and beamforming mics steered by detecting which
other apps. It all amounts to a playground for sound one is attending to, selective amplification will
innovation that will yield exciting results for hearing not be totally banished from hearing enhancement
enhancement. devices. But it looks increasingly likely that even
amplification will be available from third-party app
developers.
Already Mimi Hearing Technologies’ app for
personalizing streaming audio, an unregulated
hearing feature, is becoming available on both the
Bragi and the Sonical platforms. It is well within
the capability of app developers to offer a similar
program for the microphones, essentially converting
a smart wireless earphone into an OTC hearing aid
after the fact.
It is difficult to imagine how the FDA will regulate
Figure 9: This Segotia slide describes the KU Leuven study cited in text. (Image Source: earphones, which do not themselves provide
Segotia) amplification even if a percentage of consumers
will buy an app from a third party afterward, and
non-amplifying hearing features are not regulated
at all. The newly released OTC hearing aid rule
will be left by the wayside should these promising
developments become reality.
O n e t hin g i s c l e a r. W hil e t h e p a c e o f
developments may in time give regulators and
some other stakeholders fits, the consumer
stands to benefit from all the innovation coming
to hearables. With the ability to “try before you
buy” a variety of hearing apps and load them into
mass-market consumer earphones, more people
will find an effective solution for their hearing
and lifestyle, with less stigma. That can only be
Figure 10: AAVAA’s EEG sensor implementation is shown. (Image Source: AAVAA) a good thing. ax

References
[1] A. Bellavia, “Is the FDA’s OTC Hearing Aid Rule Already [6] N. Chong-White, PhD, J. Mejia, PhD, J. Valderrama-Valenzuela,
Obsolete?” LinkedIn, March 3, 2022, www.linkedin.com/pulse/ PhD, and B. Edwards, PhD, “Evaluation of Apple AirPods Pro with
fdas-otc-hearing-aid-rule-already-obsolete-andrew-bellavia Conversation Boost and Ambient Noise Reduction for People with
Hearing Loss in Noisy Environments,” Hearing Review, March 22, 2022,
[2] L. Gerlach, G. Payá-Vayá, and H. Blume, “A Survey on Application
https://hearingreview.com/hearing-products/hearing-aids/psap/
Specific Processor Architectures for Digital Hearing Aids,” J Sign
apple-airpods-pro-for-people-with-hearing-loss-in-noisy-environments
Process System, March 20, 2021,
https://doi.org/10.1007/s11265-021-01648-0 [7] J. Moolayil, “A Layman’s Guide to Deep Neural Networks,”
Towards Data Science, July 24, 2019,
[3] “Medical Devices; Ear, Nose, and Throat Devices; Establishing
https://towardsdatascience.com/a-laymans-guide-to-
Over-the-Counter Hearing Aids,” Federal Register, Volume 87,
deep-neural-networks-ddcea24847fb
No. 158, p. 50703 (page 6 of rule).
[8] Greenwaves Technologies, “GAP9 Product Brief,”
[4] D. R. Moore, M. Edmondson-Jones, P. Dawes, H. Fortnum,
https://greenwaves-technologies.com/wp-content/uploads/2022/06/
A. McCormack, R. H. Pierzycki, and K. J. Munro, “Relation between
Product-Brief-GAP9-Sensors-General-V1_14.pdf
speech-in-noise threshold, hearing loss and cognition from 40-69
years of age” PloS one, 9(9), e107720, September 17, 2014, [9] R. Zink, S. Proesmans, A. Bertrand, S. Huffel, M. de Vos,
https://doi.org/10.1371/journal.pone.0107720 “Online detection of auditory attention with mobile EEG: closing
the loop with neurofeedback,” 10.1101/218727, November 2017,
[5] B. Edwards, “Emerging Technologies, Market Segments, and
www.researchgate.net/publication/326717755_Online_detection_
MarkeTrak 10 Insights in Hearing Health Technology,” Seminars
of_auditory_attention_with_mobile_EEG_closing_the_loop_with_
in hearing, 41(1), 37–54. https://doi.org/10.1055/s-0040-1701244,
neurofeedback
February 10, 2020, www.ncbi.nlm.nih.gov/pmc/articles/PMC7010484

44 | January 2023 | audioxpress.com

 ALTI 2023

Hospitality Suite
Boost Your Location, Boost Your Business
Meet | Organize | Regroup | Relax

Where Audio Business Gets Done @CES

ALTI
A S S O C I AT I O N
Audio & Loudpeaker
Technologies International

The Venetian Expo and Convention Center

(former Sands Expo) Level 3

Contact Barry Vogel: Official Media Sponsored

+1 602-388-8669 Partner by
or [email protected]
ax R&D Stories

The Audio Foundry Automotive

Audio Development Platform
By
Adam Levenson

This article details the possibilities of the Audio Foundry platform for automotive audio prototyping
and to support the development of audio experiences in vehicles with powerful design tools and
services. The result of a collaboration between Tymphany and DSP Concepts.

It’s rare to find genuine innovation in the audio Steve Ernst agrees with this fundamental definition
tech field, and even rarer to get an opportunity to but expands on it calling it “a place, specialized
contribute to a groundbreaking project. Back in equipment, and services for OEMs to come to and
June 2022, at the Audio Engineering Society (AES) accelerate their audio development.”
Automotive Audio Conference in Dearborn, MI, Matt Marchese provided the more practical vision.
surrounded by gleaming luxury demo cars, DSP “The Audio Foundry is also synonymous with the
Concepts and Tymphany set up a live demonstration partnership and collaboration between DSP Concepts
in the corner of the room. Automotive OEMs and and Tymphany. It’s also synonymous with the space,
Tier 1s began lining up to step inside the brushed and soon to be spaces, available at a global level for
aluminum acoustic hulk known as the Rig, powered Tier 1s, OEMs, and technology providers.”
by Audio Weaver and displayed on a big screen
perched in front of the Rig’s windshield. A Changing Market
As Tymphany’s Matt Marchese put it, “You Naturally, we asked the team to explain the
could see it, sitting in the Rig with these OEMs market conditions that inspired the creation of the
and them just having this OMG moment. It was an Audio Foundry and what called for the solution. “I
eye opening moment, even for us.” Soon after, the always like to say that this is an unprecedented
Audio Foundry was born. time in the industry where the technology in the
To understand the vision and the scope of this vehicles is changing, as are the vehicles themselves,”
totally new offering for the automotive industry, I sat Marchese responds.
down with the Audio Foundry team, which includes Whitecar says, “I think there are two key aspects:
Matt Marchese (Automotive Business Unit Leader, First, the complexities of the audio experiences in
Tymphany), Evan Bunner (Engineering Services a vehicle are changing rapidly. For 20 years, audio
Team Leader, Tymphany), John Whitecar (VP of has been pretty staid in cars but the explosion in
Product Management, DSP Concepts), and Steve electric vehicles (EVs) has really opened up a number
Ernst (Head of Automotive Business Development, of new use cases, a number of audio experiences.
DSP Concepts). Plus concurrent use cases drive all of these audio
Naturally, each member of this team has a functions at the same time: playback experiences,
particular vision of the project, as with different voice experiences, emergency vehicle detection,
pieces of the puzzle. John Whitecar calls Audio AVAS... All of that is coming together all at one
Foundry “a way to accelerate audio development.” time. The other aspect of this is the fact that the

46 | January 2023 | audioxpress.com

compute power available is so vast now, that it’s And Marchese confirms, “We’ve heard from
very cost effective to implement these features. almost every OEM that they’re looking to create
People have known about these features for years, unique experiences in the vehicle. They’re looking
but they were too costly. The capability of silicon, for more and more ways to differentiate and this
when you’ve got multiple DSP cores integrated into idea of experiences in the vehicle has become a
a part, suddenly opened up a whole new world of main driver and audio helps support that.”
what you can get done.
These developments are also fairly recent. I The Automotive Audio Challenges
would say probably the last three years. But when This in summary explains the requirements
you look at the compute power perspective, the and the market conditions. But this project directly
Qualcomm Snapdragons with embedded Hexagons, reflects the need to find new solutions for the
that’s kind of the first platform that really enabled challenges the OEMs and Tier 1s are facing when
all this. It allows moving audio processing from an building audio systems. The experienced team
external DSP into an embedded processor. Obviously behind Audio Foundry is in a unique position to
Tesla came out with EVs over 10 years ago, but it’s understand these new audio experiences in vehicles,
only recently that the market has really started to and consequently illustrate what problems and
pick up. You look at what’s going on in China. You solutions they proposed to solve.
look at every major OEM that has now pivoted and According to Ernst, “The OEMs and Tier 1s
shifted to EV manufacturing. The whole industry have not changed the way they’ve done business
is moving right now.” for the last 25 years, 50 years. That is clearly the
Steve Ernst adds, “What’s also driving this are challenge for them because we promote this concept
requirements. There’s a use case expansion, there’s of the modern sourcing model, which is completely
this bandwidth opportunity or processing power juxtaposed against this idea of the black box. There’s
capability, but there are also now requirements that so much at stake now, and OEMs want more control
are driving it. Many of these are not elective. When and ownership, but it changes how they do business
you look at these new groups that are leveraging and where the skill sets are, and that’s the biggest
audio, it is out of a pure legality perspective: they hurdle for them.”
have to do it. There are international standards “When we talk to the OEMs, even when they
related to EVs. Because now, for example, you need emotionally make this determination to use Audio
to create external sound in order to alert pedestrians. Weaver—as an example—it comes back to, ‘How do
Autonomous would be the other driving force: when I do that?’ And, ‘I know I need it, I know I want it,
autonomous is engaged, there are safety alerts and and I gotta have it, but I don’t know how to do it,
things that have to take place in order to make sure and I don’t have the facility, talent, or resources to
that person stays engaged and understands the state do it.’ Part of the Audio Foundry’s value proposition
of the vehicle when it’s driving in autonomous mode. is our ability to support the OEMs no matter where
So it’s beyond, call it fun and entertainment, it’s now they are in terms of readiness to achieve their audio
moved into a whole different realm of requirements goals. This is a discovery platform for them now to
in order for the vehicle to be drivable.” figure out how they’re going to manage this new

The Audio Foundry

platform offers purpose-
built labs for automotive
audio prototyping,
offering listening rooms,
measurement rooms, studio
rooms, and anything needed
to support engineers with
powerful design tools and
services.

audioxpress.com | January 2023 | 47

ax R&D Stories

development cycle and purchasing strategy.”

Whitecar added that, “The automotive industry
has been very hardware focused. Traditionally, they
are hardware mechanical people putting stuff into
cars. So that’s the other big paradigm shift: The
industry is moving very quickly to a software model.
And so, back to Steve’s point, they don’t know how
to do it because it’s a completely different skill set.
It’s going from: ‘I can build you a box that does
these two features,’ to ‘I’ve got to get this feature
into my software environment onto a main SoC, or
our main module that’s controlling everything.’ So
you’re changing the skill sets, you’re also changing
how they organize teams.”
Whitecar continues, “It used to be, ‘I have a
team that builds this box, and I’ve been building
this box for every car for the last 20 years, and my
A fundamental part of the Audio Foundry platform, the Rig is an SUV-sized audio
team knows everything there is to know about this
development tool used for creating new acoustic solutions, layouts, and proof of concepts.
box.’ And now that box is gone. So now it’s like ‘How
It offers an easily adjustable, modifiable acoustic environment with 24+ locations for
speaker placement with replaceable panels, delivered with services and maintenance
do I get this done?’ ‘My normal team is broken,
packages. my approach is broken, my processes are broken.’
So they have to figure out how to pivot and move
in to this heavy software domain that leverages
major hardware components. Now they have to get
teams to work together. The NVH (Noise, vibration,
and harshness) team has to talk to the audio team,
Steve Ernst, Head of has to talk to the infotainment team, and the voice
Automotive Business team. They all now have to work together to build
Development, joined DSP these products.”
Concepts in January 2021
in Santa Clara, CA. He has a
Who Is It For?
30-year sales and marketing
So, who is asking these questions and seeking
career in automotive
audio, having worked as
the new solutions? What is the role of these people
business unit leader in at the OEMs and Tier 1s?
companies such as Harman According to Ernst, “It’s not the VP, it’s that layer
International, stationed in below. The person who owns the department that
the US, Japan, and South makes the actual deliverable. They’re getting the
Korea. assignment from the VP to say, ‘Hey, by the way,
we’re a software company now. Deliver everything
to me as software, not hardware.’ So it’s that layer
below management that has to figure out what that
means and has to define how it’s going to happen.”
Matt Marchese, Automotive
Marchese s ays, “I also think about the
Business Unit Leader at re-empowerment of the automotive audio engineer.
Tymphany, is based in San There’s been this heavy reliance on the Tier 1 to
Francisco, CA. Matt began bring forward the solutions, bring forward the
his career in automotive innovation, and I think that’s where you can now
with in-vehicle electronics see the shift. Audio engineers are now wanting
at Ford/Visteon, worked to take control of their own destiny and do this
at CARLAB, an automotive
work themselves. I think the Audio Foundry really
product planning consulting
helps support and gives them a venue to explore,
firm for nearly every global
OEM, and previously worked
experiment, and develop.”
for THX, supporting audio/ “And, especially for some of the startup
video certification and companies, where there’s this idea of creating more
licensing. consumer centric approaches like having removable

48 | January 2023 | audioxpress.com

 Evan Bunner, Engineering
Services Leader, Tymphany,
based in San Rafael, CA, is
a proven Automotive Design
Engineer, experienced
speakers or having headphone sets. They’re in Product Development
basically creating a whole different experience in and Vehicle Engineering.
the audio system, which requires hardware aspects, Evan has been developing
transducer aspects, system aspects, and then there automotive components
are software aspects that all have to be proven since the beginning of
his career in 2013, when
out and put it into program. So Audio Foundry
he started by developing
gives managers a way to work on these concepts
composite structures
and present them to the decision makers. I think for racing vehicles and
Steve’s right on: I think it’s that level at which supported NVH research
somebody owns the problem of having to deliver the activities for Honda.
features, and they may not have the headcount or
the facilities to actually pull it off,” explains Whitecar.

The Audio Foundry’s Backstory “I’ve known John for some time, since I was
So, how does a partnership like this get started? working in the consumer side of the business. In
According to Whitecar, Marchese had the vision. getting feedback from Tymphany’s own individual
Tymphany was working on a proof of concept and pursuits and noting how often Audio Weaver would
had already developed an SUV-sized automotive become part of those discussions, it just became
audio development environment for its own internal apparent that, within the development realm, we
uses. But it became clear that there was some didn’t have the software side to complete a platform,”
synergy between the hardware and the software, Marchese explains “The funny thing is that we
and the possibility of building out experiences and spent all of the summer of 2020 talking to software
bringing a platform and a partnership together that companies about how hardware and software could
OEMs and Tier 1s could use. So Marchese called DSP go together. And honestly, we had a very difficult
Concepts and that was the genesis of it. time finding a partnership that really stuck or had

KLIPPEL
Analyzer System

Output-based Testing for Any Loudspeaker

 Meaningful SPLmax with multi-tone (MTON)
 Room correction with in-situ compensation (ISC)
 3D sound radiation in near and far eld
 Equivalent input harmonic distortion (EID)
 Impulsive distortion (Rub & Buzz)

🌐🌐 www.klippel.de ✉ [email protected]

audioxpress.com | January 2023 | 49

ax R&D Stories

John Whitecar, VP of It was like, you know, these two things coming
Product Management, together to really accelerate the development. You
Automotive at DSP Concepts,
could see it, sitting in the Rig with these OEMs and
from Stanford, CA, is a
them starting to discuss different scenarios in which
world-class automotive
audio technology expert
they could see how this could help improve their
with c-suite and VP level timing, improve their costs, help them to improve the
engineering experience at way they innovate. It was an eye-opening moment,
Tesla Motors, Ford, Visteon, even for us,” recalls Marchese.
and Texas Instruments. He
holds 35 patents covering The Audio Foundry Advantage
audio, acoustic, and As Evan Bunner explains it, the Audio Foundry
receiver signal processing
platform is intended to be a one-stop-shop for
technologies.
proving new ideas, trying something new, and
finding the next game-changing technology in the
automotive industry. “The idea is to have all of the
ingredients available to make that possible for a wide
range of OEMs. At the core, you’ll have a workshop
stickiness to it. This model, this early engagement to do hardware stuff, and a platform to implement
services model is not new to Tymphany. We’ve been software innovations.”
doing this on the consumer and pro audio side, and “There’s this really classic process for designing
we’ve been trying to find a unique approach into products, any product. It goes back to this classic V
automotive. So with our pre-existing partnership diagram where you have requirements on one side of
with DSP Concepts, and knowing what DSP Concepts the V and system validation at the side of the V. You
means in the industry, it was a no-brainer to reach start with the big objectives, work your way down to
out and see if there was an opportunity to expand the little features and pieces of code and individual
the partnership into automotive.” components, and then build the system back up on
Marchese also points to the AES automotive the validation side. So you start big with concepts and
conference in Michigan as the pivotal point, when ideas, and really, the upper left hand corner is fuzzy
the two companies demonstrated the vision of things such as a new experience that an OEM wants
hardware and software working together. The to provide to the customer. It starts very vague. It
positive reactions from the attendees showed that has high-level requirements associated with it. So the
the companies had something there (the full report Audio Foundry workflow is to get those things done
is available in audioXpress, September 2022). quickly,” explains Whitecar
“We had the exact right demographic including “To have a hardware platform and a software
Tier 1s and OEMs sitting in the Rig with Audio Weaver. platform that lets you really spin through ideas quickly.
Then you can start refining those ideas, start pushing
into hardware, pushing software to specific targets.
So you can take this concept that’s fuzzy, done at
the high level, now you can bring it back into the Rig
and test it with closer to production hardware, with a
software environment, and push your way down the
V towards productization. And then you can actually
provide mechanisms for validating what you set out
to do as you get closer to the real products. The whole
idea is that each step is accelerated because you’ve
abstracted a lot of the low-level details out in the
process, which is usually what slows you down. If you
think about a typical approach, if an OEM wants to put
an audio concept into a car today, they have to tear
up a car, they have to cut holes in the car, they have
to redo trim panels, they have to find hardware to run
things on. It’s a very time consuming and arduous
process,” Whitecar continues.
The Audio Foundry Makerspace. Fully equipped, state-of-the-art facilities located in San Another key aspect, as Ernst reminds us, is the
Rafael, CA, and Wales, UK, which can be leased and staffed. difficulty in getting access to the actual car that

50 | January 2023 | audioxpress.com

an audio project is intended to fit. “It’s not easy to
have access to the car. That’s no problem for audio
engineers working for an OEM, but it’s not easy for
all others to get access to a vehicle, especially for
the period of time that they’re going to need to fulfill
all of the requirements and experiments. Especially
if they want to cut and drill into a prototype car.”
It is an experience that Whitecar could confirm
happened many times over his career. “Early
prototypes are rare resources, and so you have less
flexibility. So typically what OEMs will do is they’ll
take an existing car, a production car that’s already
on the market, and maybe it’s been used, and then Hands-On Platform The first demonstration
of Tymphany’s and DSP
you start cutting into that. In addition to offering a hardware and software
Concepts’ environment
“It’s very rare to have a hot off-the-line, early platform and a makerspace, the Audio Foundry for automotive audio
prototype, and then start to hack at it. Because that’s provides support from the DSP Concepts and development during
a very expensive car and there’s a lot of people who Tymphany teams. I asked them to describe some the Audio Engineering
want to get time with that vehicle. So you have to of the ways to work with the Audio Foundry. Society’s 2022 International
work offline. And Audio Foundry is even more of Ernst points out that OEMs and Tier 1s can start Automotive Audio
an ideal offline solution because it’s a clean sheet by just commissioning a design-of-experiment for Conference in Dearborn, MI.
of paper. It’s a 3D sandbox. You can put anything the Foundry team to put together, measure it, and
anywhere, and really think through experiences, and report. “It can start at that level, where it’s called a
not focus so much on how do to get it done. It’s more professional service, and then from that point, it can
about making it possible to achieve what you want expand if they want to be involved. So it could be that
to get done,” adds Whitecar. they want to conduct a DOE, make changes, have this

Interested in learning how

guitar amps work, as well as
how to fix and service an amp?
Then this book is for you.
Jack Darr’s Electric Guitar
Amplifier Handbook
details the following:

• How guitar amplifiers work

• How to make amp repairs
properly and safely
• How to troubleshoot tube
and transistor amps of
all sizes
• Details on “typical”
amp circuits
And much more!

Whether you’re an audio engineer

or a musician with a thirst for knowledge, you’ll
will find the explanations in this book easy to comprehend
as long as you have a general interest in electronics and
a passion for DIY.

Get your copy today! cc-webshop.com

audioxpress.com | January 2023 | 51
ax R&D Stories

Audio Weaver offers AWE

Tune for system tuning of
the entire audio system in
vehicles, including modules
to measure the acoustical
fingerprint of a vehicle
for all seats, a complete
system for filter design and
optimization, and real-time
sound tuning and listening
test tool.
setup, create the hardware, and the Foundry team space itself, this will be the first of its kind, this does
will work with the client as a part of the project, and not exist in the industry today. It’s a space that’s
then collectively create that report. Another option 100% dedicated to automotive audio development.
is that the customer can lease the Audio Foundry An unique space for OEMs and Tier 1s to go, to work
makerspace that’s fully equipped with the Rig, out problems and explore solutions.”
listening/measurement/studio rooms, workbenches, “The space is going to mean different things
3D printers, and more. That enables the customer to to different car companies. I would love for car
do the work themselves with expert Foundry team companies to look at this, even the most developed
support. A third option is that the Foundry can build ones, as an extension of their current capabilities.
a Rig and deliver it to the customer’s facility with Many car companies, older car companies, have very
training and a services package.” developed audio engineering labs, but to the points
“The Audio Foundry team can help support that we’ve made throughout this discussion, those
the collaboration or development of a new system don’t include what the Audio Foundry is providing.
or architecture or proof of concept,” Marchese This may not be their 24/7, 52-week place, but Audio
confirms. “And as Steve also said, the client can do Foundry can be viewed as an extension of their
that independently in the space and utilize the tools workplace, to get benefits that they’re not going to
and the environment to do that. The magic is in the get within their own four walls,” Ernst reinforces.
Bunner also believes that the Audio Foundry
About the Author platform will have different resonances with different
Adam Levenson’s 30-year audio career has evolved companies. “It will be a place to experiment for the
from sound effects design, to business development, to bigger companies, but for the startups, it could be
marketing with roles as Director of Audio at Activision, their day to day workshop.”
GM at CRI Middleware, VP of Product Marketing at Waves, “Some of the smaller car companies don’t even
and now Head of Marketing at DSP Concepts. have a place to tune their car. We know that it’s a
different story for the big car companies: They have
plenty of places to tune and they have their supply
Tymphany is a leading company in the base, but it’s not that way with the small guys,”
design and manufacturing of audio Ernst says.
systems, headphones, acoustic modules, Having extensive experience in supporting
and drivers. With more than 90 years of audio design, development and engineering teams and Tier 1s in audio systems
manufacturing expertise, Tymphany is widely regarded as having the strongest for the automotive industry, Tymphany and DSP
portfolio of technology and design experience in the loudspeaker industry. Concepts understand how such purpose-built labs for
Tymphany manufactures OEM/ODM loudspeakers for many of the world’s leading automotive audio prototyping can become a valuable
audio brands and also develops and sells transducers under the Peerless brand. proposition for early concept validation, accelerate
development, increase design flexibility and facilitate
DSP Concepts is a leading embedded audio in-vehicle concept presentations.
technology supplier and creator of Audio Weaver, The Audio Foundry platform offers fully equipped
the audio development platform that makes workspaces, the Rig development sandbox, onsite audio
audio innovation easy. DSP Concepts equips experts, and the power of Audio Weaver for audio
engineers with real-time workflows to quickly stand up prototypes, collaborate development, with integrated debugging for hardware
and modify designs across teams, and deploy to the most popular chipsets. from and software, and a multifaceted tuning toolkit.
leading semiconductor companies including NXP, Qualcomm, STMicroelectronics, As Whitecar summarizes, for big and small
Cadence, Arm, and others. DSP Concepts technologies are embedded in millions companies, Audio Foundry becomes a perfect facility
of automotive and consumer products, including for automotive brands including to do all kinds of experiments. For more information,
BMW, Mercedes Benz, Porsche, Tesla, and many more. visit https://audiofoundry.dspconcepts.com. ax

52 | January 2023 | audioxpress.com

The “Must Have” reference for
loudspeaker engineering professionals.

Home, Car, or Home Theater!

Back and better than ever, this 7th edition
provides everything you need to become a
better speaker designer. If you still have a
3rd, 4th, 5th or even the 6th edition of the
Loudspeaker Design Cookbook, you are
missing out on a tremendous amount of
new and important information!
Now including: Klippel analysis of drivers,
a chapter on loudspeaker voicing, advice
on testing and crossover changes, and so
much more! Ships complete with bonus CD
containing over 100 additional figures and
a full set of loudspeaker design tools.
A $99 value!
Yours today for just $39.95.

Shop for this book, and many

other audio products, at
www.cc-webshop.com.
ax Simulation Techniques

Loudspeaker Driver
Displacement Decomposition
This valuable tutorial offers a unique overview on the possibilities of
three-dimensional analysis of loudspeaker driver displacement, using
Phase Decomposition and Axial-Symmetrical Decomposition techniques.

By
René Christensen
When looking at a frequency response from a Phase Decomposition
loudspeaker driver (magnitude and phase of the The first technique relies on knowing the complex
pressure in a microphone point), a lot of information pressure in a microphone point and relating the
can be extracted regarding its behavior. The complex displacement across the loudspeaker surface
magnitude response is typically the primary focus, to the pressure, or more specifically to the phase of
and a flat response here is the desired outcome, said pressure. Knowing these quantities, the following
whereas the (oftentimes ignored) phase response displacement components can be established:
holds information about the temporal aspects of the
loudspeaker and generally of the setup regarding
number of drivers and how they combine to the full (
win ( Q, p, R ) = Re + w ( Q ) e )e
− iφref iφref

response. However, the sound pressure is simply a

zero-dimensional signal, which does not necessarily wanti ( Q, p, R ) = Re ( w ( Q ) e ) e
− − iφref iφref

tell us the whole story about the three-dimensional

 π
aspects of the surface vibration and how it generates
( )e
i  φref + 
win ( Q, p, R ) = Im w ( Q ) e
− iφref  2

the sound. To really dissect where, for example, dips

and peaks in the pressure response come from
some other tools can help, and two such tools are with a reference phase of:
the Phase Decomposition and the Axial-Symmetrical
Decomposition techniques, respectively. φref = arg ( p ( P ) ) + π + kR
The Phase Decomposition technique [1] as
well as the Axial Symmetrical Decomposition [2] I outlined the details in two separate articles
have been promoted by Klippel, where they are that I wrote on the subject [3] and [4], so let’s dive
built into measurement systems and so can be into these equations, starting with the last one first.
applied to measurements only. I was involved in Looking at the reference phase, the first part is
the development of an add-on package that could simply the phase of the pressure in the microphone
take in simulated surface (cone, dust cap, surround) point. This is important, as this is basically the phase
displacement data, and analyze it as it had been that we want to compare with the displacement
measured. But shortly thereafter I developed a phase. The middle part indicates a sign shift; this
simulation setup in COMSOL Multiphysics where the relates to the fact that the pressure in general
phase decomposition could be carried out directly correlates with the acceleration of the cone and
based on simulation data, and this approach has the displacement is proportional to the negative
been used in the analysis and synthesis work for acceleration. The final part is basically a subtraction
several clients since then. of the phase related to the path from the integration

54 | January 2023 | audioxpress.com

point Q to the microphone point P; a subtraction of For the COMSOL Multiphysics implementation, the
the irrelevant phase coming from the time delay/ displacements are found via a laser scanner, which
transport delay. The total displacement in a point only considers the axial displacement, whereas in a
Q on the surface is denoted by w(Q), the respective simulation software setup, the user also has access
in-phase, anti-phase, and quadrature components to the radial and normal displacements, which can
as previously listed. The in-phase displacement is lead to more accurate results. (In the present work,
the displacement that is in-phase with the reference the simulation software COMSOL Multiphysics has
phase, which I already explained is the pressure been used throughout.)
phase taking into account the sign and the distance Figure 3 shows how a simple total displacement
aspects. This part of the displacement adds to the can be decomposed into in-phase, anti-phase, and
total pressure. The operator Re+ returns the positive quadrature displacements.
part of the real part. Basically, the equations indicate
Figure 1: The displacement
rotations, projections, and back-rotations according
and pressure phasors are
to Figure 1, which shows the complex displacement
shown here.
phasor w(Q) and the complex pressure phasor p(Q)
along with the component axes.
Similarl y, t he anti-p ha s e dis p lace m e nt
components subtract from the total displacement.
And finally, the quadrature displacement component
neither adds to nor subtracts from the total
displacement, with Im being the imaginary part
operator. Both the in-phase displacement and
the anti-phase ditto have associated in-phase
and anti-phase pressures, respectively, whereas
the quadrature displacement does not affect the
pressure and so has zero associated pressure.
For the Klippel implementation, the pressure
is not measured, but instead calculated via the
Figure 2: A flat vibrating
so-called Rayleigh integral:
P surface S is placed in an
infinite baffle. At point Q
ω 2 ρ0 e − ikR
p ( P) = − ∫w (Q ) dS on the surface, the total
2π S R R w(Q)
displacement is w(Q). The
∞ distance from point Q to a
with the angular frequency ω, the density of the microphone an observation
medium ρ0, the distance from the microphone point ∞ Q ∞ point P is denoted R.
to the points Q on the surface under integration, and S
S being the entire surface to be integrated. Using
the Rayleigh integral, there is an assumption that
the surface is flat and placed in an infinite baffle ∞
(Figure 2).
The latter requirement fits well with how the
typical simulation setup is anyway, but the former
is an issue, since most drivers are not flat. Still, the
overall analysis and conclusion are typically of value,
and a correction can be established before running
the actual phase decomposition. These issues were
explored in an Audio Engineering Society paper that
I co-wrote with Urik Skov [5], but for each case one
must evaluate the effect of not fulfilling the “flat”
requirement. (One thing to note is that while the
varying distance coming from varying depth on
the cone is in fact included in the calculations, the
dipole aspect of the full Kirchhoff-Helmholtz equation
that the Rayleigh integral can be derived from, is Figure 3: The displacement components for a vibrating disk are shown for an on-axis and
not included, but these details are left to reader.) far-field observation point.

audioxpress.com | January 2023 | 55

ax Simulation Techniques

On the surface the displacement only has two Other very interesting results can fall out of
phases—“inward” and “outward”— and so the bigger phase decomposition, and once it is implemented,
of the two sections will be the in-phase component it will typically be used extensively. The additional
driving the pressure, and the smaller section will components such as Accumulated Acceleration
be in anti-phase, and the quadrature component Level (AAL) used within the Klippel environment
will be zero. can also be calculated within COMSOL Multiphysics,
While the total displacement stays the same, the and typically with a higher spatial resolution in the
decomposition will vary depending on the microphone simulation setup.
point in question as individual distances to the surface
will vary accordingly, and so very different results Axial-Symmetrical Decomposition
can be seen off-axis and in the near-field. The axial-symmetrical decomposition is slightly
For a more practical case, I have used an more straightforward when it comes to how it works,
example from one of my clients, Charalampos but not straightforward to implement. I recently
Ferekidis ([email protected]), which is shown discussed this implementation in the blog post
with his permission. Figure 4 shows the before that I wrote for Acculution [6]. This decomposition
(a) and after (b) sound pressure level for an initial does not rely on knowing the sound pressure at
driver and the subsequently modified driver. A dip all, as it works directly on the total displacement
was seen at 1kHz, and its origin was unclear. alone. The split into components now is between a
A phase decomposition analysis revealed the radial (or more accurately “meridional,” but we will
surround as the culprit, as its contribution was stick to “radial” going forward) component and a
mainly in-phase, meaning driving the resulting circumferential component, both to be calculated
sound pressure at that frequency, while the for each frequency of interest.
remaining displacement was largely quadrature, and The radial component has symmetry around
so not adding to the pressure (Figure 5). Essentially, the circumference, and so for example pistonic
the driver is less efficient at radiating sound, since motion would fall into this category, as would any
the effective area of radiation was located near displacement that can be described as a variation
the surround. along the radius which is then swept around the

Figure 4: The initial

a) b)
response, both simulated
and measured, (a), and
the simulated response for
the modified driver (b) are
shown here.

Figure 5: A phase
decomposition analysis
at 1kHz for a client case
revealed the surround as the
culprit, as its contribution
was mainly in-phase,
meaning driving the
resulting sound pressure at
that frequency, while the
remaining displacement was
largely quadrature, and so
not adding to the pressure.

56 | January 2023 | audioxpress.com

 Figure 6: This is an example
where the constructed
total displacement (left)
is combination of a
“pistonic” motion (radial
displacement in the
middle) and a “rocking”
motion (circumferential
displacement on the right).

axis. The circumferential component is then any decomposition technique decomposes directly on
part of the total displacement that cannot be the complex surface displacement but is arguably
described this way. This could stem from modes more difficult to implement.
or asymmetries in the setup. The equation for the As the decomposition techniques shown here
radial component is: only consider the surface displacement, whatever
is happening behind the driver (spider behavior,
2π
wrad ( r ) = ∫ wtotal ( r , ϕ ) dϕ acoustic cavity modes) is seemingly not included in
0 the decomposition. But in fact, it is, in the sense that
if these effects affect the surface displacement, the
with r being the radiu s and φ being the decomposition will change accordingly.
circumferential angle. For each r on a continuum, Other decomposition techniques such as modal
a line integration is carried out along the associated can also be carried out, and it is up to the engineer
circumferential line, and the resulting displacement to figure out which approach will reveal the most
varies along the radius only. Knowing this along insight for the case at hand. In future articles we
with the total displacement, the circumferential may look at these alternatives. ax
component can be found as the difference between
the two:
About the Author
wcirc ( r , ϕ ) = wtotal ( r , ϕ ) − wrad ( r ) René Christensen (BSEE, MSc, PhD) has been working with
simulations in the loudspeaker and hearing aid industry for several
Figure 6 shows an example where the years, and in 2021 he started his own company, Acculution ApS,
with a focus on mathematical modelling of many different products,
constructed total displacement (left) is combination of
such as transducers. The overall aim is to use academic research
a “pistonic” motion (radial component in the middle) results for finding better engineering solutions faster, training
and a “rocking” motion (circumferential component engineers in this type of virtual prototyping, and working with
on the right). It is seen that the decomposition students across the world to progress the field.
correctly extracts the individual components. The
present decomposition can give insight into modal
behavior among other things. Again, having the
References
implementation makes it easy to employ whenever [1] W. Klippel and J. Schlechter, “Measurement and Visualization of Loudspeaker
relevant. Cone Vibration,” Klippel GmbH, Dresden, 01277, Germany and Faculty of Computer
Science, University of Technology, Dresden, Germany.
Concluding Remarks [2] “Cone Vibration and Radiation Diagnostics,” Application Note to the Klippel R&D
It has be en demons trate d how a tot al System, AN31, April 4, 2012.
displacement on a cone surface can be decomposed [3] R. Christensen, “Phase Decomposition Analysis of Loudspeaker Vibrations,”
into different components directly in the simulation Dynaudio A/S, COMSOL Blog, April 21, 2015,
software COMSOL Multiphysics, and how this can www.comsol.com/blogs/phase-decomposition-analysis-of-loudspeaker-vibrations
reveal some of the underlying mechanisms of the [4] R. Christensen, “Phase Decomposition for Loudspeaker Analysis,” GN Resound
driver. The phase decomposition decomposes A/S and Dynaudio A/S, 2016 COMSOL Conference in Munich,
relative to a complex sound pressure, and certain https://www.comsol.com/paper/download/356591/christensen_paper.pdf
aspects about limitations of the Rayleigh integral [5] U. Skov and R. Christensen, “An Investigation of Loudspeaker Simulation
have not been touched upon here but can be found Efficiency and Accuracy using A Conventional Model, A Near-To-Far-Field
in that AES paper that I co-wrote with Skov [5]. Transformation and The Rayleigh Integral,” 136th Audio Engineering Society (AES)
Conference, Berlin, Germany, 2014.
Another consideration is whether to use axial
or normal displacements for non-flat surfaces, and [6] R. Christensen, “Axial Symmetrical Decomposition… In COMSOL Multiphysics,”
Acculution Blog Post #036, August 29, 2021, www.acculution.com/single-post/
that will be discussed in an upcoming publication
axial-symmetrical-decomposition-in-comsol-multiphysics.
with Charalampos Ferekidis. The axial-symmetrical

audioxpress.com | January 2023 | 57

ax Audio Electronics

Power Transformer Parameters,

Selection, and Testing
Part 3 — Insulation Materials,
Winding Bobbins, and Testing Methods

In the previous two articles in this series, Chuck Hansen detailed the history of transformer cores
and their various construction methods. This article explores transformer insulation materials and
winding bobbins, and testing methods for the insulation system.

By
Chuck Hansen
Commercial transformers are often rated in Electrotechnical Commission (IEC) 60601 standard
terms of the maximum operating temperature rise and other internationally recognized safety agency
(temperature increase above ambient) according to standards.
a letter-class system: A, B, F, or H. These letter codes Medical-grade transformers use toroidal
are arranged in order of the lowest temperature construction (Photo 8), which has the following
rise to the highest: advantages: low stray electric and magnetic fields;
high efficiency, tight regulation factor (ε), some as
• Class A: No more than 55°C winding low as 0.007 (0.7%); and Faraday shielding between
temperature rise, at 40°C maximum ambient primary and secondary windings.
air temperature The effectiveness of an electrostatic field
• Class B: No more than 80°C winding shield is largely independent of the geometry of
temperature rise, at 40°C maximum ambient the conductive material. Solid foil is better than a
air temperature layer of wire. Silver is better than copper, copper
• Class F: No more than 115°C winding is better than aluminum, and so on.
temperature rise, at 40°C maximum ambient However, fixed or low-frequency magnetic fields
air temperature can penetrate the conductive shield completely.
• Class H: No more than 150°C winding Electromagnetic shielding requires soft magnetic
temperature rise, at 40°C maximum ambient materials.
air temperature In the case of varying electromagnetic fields, the
higher the frequency the better the material resists
Medical-Grade Transformers magnetic field penetration. As with electrostatic
Medical equipment has to be extremely sensitive shielding, electromagnetic shielding also depends
to deal with low patient signal levels, and also be on the electrical conductivity as well as the thickness
highly precise to enable accurate medical diagnosis. of the shield.
Transformers that are considered to be medical Medical power isolation transformers (Photo 9)
grade need to satisfy a number of additional safety use medical-grade toroidal transformers installed
guidelines and rules beyond those for non-medical- in an enclosure. They are designed for a wide
grade transformers. They are designed to protect frequency range of 47Hz to 63Hz, with output
caregivers, patients, and the equipment used in their voltage taps of 90, 100, 120, 208, 220, and
care, so medical-grade transformers are required 240 Vrms, to accommodate both 50Hz and 60Hz
to maintain strict adherence to the International power sources.

58 | January 2023 | audioxpress.com

ax Audio Electronics

They are designed to certain standards and have • Continuous noise filtering, enhanced common
several protective characteristics including: mode surge protection, filters out EMI/
RFI noise, utility switching transients, load-
• Low leakage current to ground, to equipment generated harmonics and ground loops
enclosures, and to the patient • Residual-current circuit breaker (RCCB) or
• Low patient auxiliary current (measured ground fault circuit interrupter GFCI), which
between any single patient connection and directly detect leakage current
all other patient connections of the same • Built-in thermal protection for overload and
module or function, such as electrocardiogram short circuit conditions
electrodes) • NEMA 5-20R hospital-grade output receptacles
• Built in accordance with ISO 9001:2015
Photo 8: Medical-grade
certification
toroidal transformers
• UL/IEC 60601-1 compliant medical-grade listing
use toroidal construction.
(Image Source: Triad
• Leakage current less than 100µA
Magnetics) • Double-reinforced insulation to withstand
5,000 Vrms Hi-Pot
• Efficiency: 90%-98%
• Low magnetic stray field
• Low mechanical noise
• Minimums on creepage distance and air
clearance
• Cool operation with a maximum temperature
rise between 25ºC to 55ºC.

Photo 9: A medical power

There are a couple different types of medical-
isolation transformer
assembly, shown here, uses
grade transformers. Safety ground transformers
medical-grade toroidal utilize standard transformer isolation. A safety
transformers installed in an shield between the primary and the secondary coils,
enclosure. (Image Source: with insulation between the shield and each coil,
Standex) limits the maximum leakage current. If the isolation
fails, the AC input current is immediately redirected
to ground through the shield, which will trip the
overcurrent device. The other type are double or
reinforced insulation transformers, which can be
larger than standard transformers, have two or
more means of insulation. Should one insulation
layer fail, the others will maintain isolation. If
this second insulation layer fails, the overcurrent
protection device will trip (fail safe).

Military-Grade Transformers
Military transformers are designed in accordance
with MIL-PRF-27F (Photo 10). They have specific
construction grades, and are designed to operate
over the wide milit ar y maximum ambient
Photo 10: This is a MIL-PRF- temperatures, temperature rises, and electrical
27F Grade 5 encapsulated overloads. The 115V AC 400Hz power is used in
PC board-mounted power aircraft, Navy submarines, space programs (Bendix
transformer. This particular made the 400Hz static inverters for the Gemini
one is a Pico Electronics
program) and even the AC power for some computer
88660, 15VA toroidal
server rooms.
transformer designed
to MIL-PRF-27F Grade 5
The transformers are available in two different
Class S, with PC board 400Hz primary voltages, and with numerous
pins (Image Source: Pico secondary voltages from 5V AC to 115V AC. The
Electronics, Inc.) 115V AC transformers are used to provide various

60 | January 2023 | audioxpress.com

lower voltages, such as 26V AC, 36V AC, and 28V Table 1: The class of
Symbol Maximum Operating
military-grade transformer
DC by means of transformer-rectifier units or Temperature
is identified by a single
switching power supplies. The 26V and 36V AC Q 85°C letter in accordance with
transformers are used to supply essential AC power R 105°C MIL-PRF-27 and denotes
for inertial navigation systems, autopilot, cockpit the maximum operating
S 130°C
instrumentation, and low voltage lighting. temperature of temperature
V 155°C
The Pico Electronics 88660 is fully specified. The rise plus maximum ambient
no-load exciting current is 20mA. The primary and T 170°C temperature, and overloads
secondary DC resistances are 27.4Ω and 4.15Ω. The U 170°C, as specified of 12% for 48 hours.
open circuit primary and secondary inductances at
400Hz are 694mH and 78.2mG. Typical efficiency turn in an adjacent layer above or below it can be
and load regulation are 90% and 7%, respectively. significant. This determines the thickness of the
They also measured some typical winding interwinding insulation required to prevent a flash-
parameters for us. For instance, the primary over failure.
leakage inductance (400Hz secondary shorted) is Commercial E-I transformers use cellulose
1.83mH. The secondary leakage inductance (400Hz impregnated paper to insulate the windings from
primary shorted) is 216µH. The primary-secondary the transformer core and metal frame, and the
capacitance is 473pF. The class is identified by a individual windings layers from each other. There are
single letter in accordance with MIL-PRF-27 as shown also standard-sized nylon bobbins, which completely
in Table 1, and denotes the maximum operating separate the primary and secondary windings
temperature of temperature rise plus maximum with barrier walls that save a lot of winding and
ambient temperature, and overloads of 12% for construction time. They also reduce the capacitance
48 hours. and mutual inductance between windings, and allow
For example, Class R has a maximum operating for a higher dielectric breakdown voltage.
temperature designation of 105°C at sea level; this
is normally at 65°C ambient plus a 40°C rise. If
the temperature rise was determined to be 30°C,
this same transformer could operate in an ambient
as high as 75°C (75°C +30°C = 105°C). At higher
altitudes, the temperature rise will increase and
should be compensated for in the equipment design.
It is not recommended that a higher operating
ambient temperature be specified than that to which
the transformer or inductor will actually be exposed.
To do so may result in a larger and heavier unit
than is needed (violating SWaP-c). Transformers
and inductors designed in accordance with this
specification are capable of withstanding an overload
of 12% for 48 hours.
Military-grade transformers are also subjected
by the manufacturer to a barrage of MIL-PRF-27
qualification tests and standardization, most of
which commercial transformers do not have to
endure.
Once qualified, the transformer is assigned a
MIL-PRF-27 qualified parts list (QPL) number and
does not have to undergo any further quality testing
unless there is a change in the design’s form, fit,
and/or function (FFF).

Transformer Coils
As transformer coils are being wound, the
difference in voltage between adjacent turns in the
same layer is quite small. However, the voltage
difference between the start turn and the finish

audioxpress.com | January 2023 | 61

ax Audio Electronics

a) b) Photo 11 shows a group of E-I lamination

transformers, left to right on top; (a) a Hammond
166G36 transformer rated 18 VA with a steel frame
and nylon bobbin; (b) a RadioShack RS 273-1512B
transformer rated 50.4 VA. It has a steel case with
steel end bells for electromagnetic shielding, and
a nylon bobbin.
c)
On the left bottom (c) is a Triad F-264U transformer
d) rated 120 VA. The laminations are exposed all around
(“bare”), and the four mounting brackets hold the
assembly together. The coils are layer wound with
the primary on the bottom. Layer winding can result
in a higher capacitance between the primary and
secondary than bobbin-wound transformers.
I also took apart an old doorbell transformer (d),
shown with the top end bell removed, and found it
interesting that all the E laminations were stacked
on the left side of the core, then the bobbin-wound
coil was placed on the middle leg. Finally, all the
I laminations were stacked on the other end, and
the mating edges of the I and E laminations were
Photo 11: Layer-wound and bobbin-wound E-I core transformers. Left to right on top: welded together. That has to increase the core losses
(a) a Hammond 166G36 transformer; (b) a RadioShack RS 273-1512B transformer; left and reduce the efficiency. The terminal board on the
bottom (c) a Triad F-264U transformer; and lower right (d) an old doorbell transformer outside of the lower end bell is for the low-voltage
with the top end bell removed. secondary connections.

EMPOWER THE MUSIC

Read.
Design.
Build.
ENJOY!

With this how-to loudspeaker book, you will be

able to crank up the volume on a first-rate system
that you designed and built yourself.

cc-webshop.com
62 | January 2023 | audioxpress.com
 MIL-Spec transformers use Dupont Kapton C core. Custom bobbins and terminal blocks can be
and glass-epoxy impregnated tape for electrical made using additive 3D manufacturing with high-
insulation. All insulating materials are controlled by temperature insulating polymers.
MILI24768. For C and E laminated cores, bobbins or Some commercial E-I core transformers use
rectangular core tubes are pre-wound with heavy a cellulose paper that has outstanding electrical
ML wire, placed over the legs of the core, then the insulation properties. It is applied in layers and
core halves are mated and held together with a extends slightly past the copper end turns of each
316 stainless steel band using custom stainless- winding (see the arrow in Photo 13). Large utility
steel retaining clips. power transformers are oil-cooled and use oil-
Photo 12 shows four typical C and E core tube, impregnated cellulose insulation.
or core box, materials. The largest one on the left (a)
as well as the third from left (c) are MIL-I-24768 type a) b) c) d)
GEBF glass-filled epoxy resin, high-temperature arc
and flash resistant insulations. The typical industry
identification is G10.
The second from left, (b) in Photo 12, uses a special
nomex-kapton-nomex construction. Finally, the core
tube on the right (d) is M24768 type PBG, paper-based
phenolic-resin, general purpose. PBG is generally used
for prototype designs or custom lab transformers that
will not be undergoing environmental, vibration, or
shock testing. The rectangular core tubes are used Photo 12: These are typical C and E core tube, or core box, materials. The largest one
for shell-type transformers using two C cores with a on the left (a) as well as the third from left (c) are MILI24768-type GEBF insulations. The
single winding in the center. Square core tubes are second from left (b) uses a special nomex-kapton-nomex construction, and the core tube
used for small transformers that use just a single on the right (d) is M24768 type PBG.

Ready for any Challenge!

Discover an Audio Analyzer you can
trust for your application:
• Sound Level Meter class 1
• Level RMS / THD+N
• 1/1 or 1/3 octave RTA
(up to 1/12 optional)

Distortion Analyzer RealTime Analyzer 1/12 octave Analyzer

I NTi Audio Inc.

[email protected] (503) 684 7050 www.nti-audio.com/XL2

audioxpress.com | January 2023 | 63

ax Audio Electronics

Newer commercial E-I core transformers use

dual-section nylon or polypropylene bobbins that
provide excellent dielectric isolation between the
primary and secondary windings, and between the
windings and the core (Photo 14).
The number of volts-per-turn of wire (V/N) is
proportional to the lowest AC frequency (f), the
peak continuous induction (B) we design for, and
the effective core cross-sectional area (A). B and f
are usually fixed values, leaving the core area as a
trade-off. This depends on the core areas of available
standard cores. The optimum core area per unit
weight for C and E core transformers is a square,
where the steel strip width is equal to the effective
metallic buildup height of the core laminations.
The required core weight is proportional to the
required output power in watts divided by the lowest
AC frequency (f), using the effective core cross-
section and the magnetic length of the core as the
variables used to select a core weight as close as
Photo 13: Cellulose paper insulation between windings (arrow) is shown on a nylon possible to the calculated required weight. Core
bobbin transformer. catalogs publish the core weights, and watts per
pound for each of their standard core part numbers.
There are times in aerospace when we will use
a custom core to achieve the desired SwaP-c. There

Advanced
200 pounds of subwoofer
in a one-ounce module Voice Coil-Based
Tactile Transducers

SALES CONTACT
Dave Lindberg
tactiontechnology.com [email protected]
+852-6821-1651

64 | January 2023 | audioxpress.com

is usually a non-recurring cost involved to tool up Photo 14: Newer commercial
E-I core transformers
for a custom core.
use dual-section nylon or
polypropylene bobbins.
Magnet Wire and Connection Leads Shown here is a tranformer
Magnet wire for commercial transformers bobbin T08. (Image Source:
and inductors is usually coated with one or more Bhavna Plastics)
thin layers of a polyvinyl formal acetate. Polyvinyl
insulation is rated for operation from -40°C to
105°C. The insulated connecting leads are usually
thermoplastic PVC.
Aerospace magnet wire is coated with one or more
thin layers of polyimide resin, which can operate at
temperatures up to +260°C. It is available in round,
square, and rectangular cross-sections. The silver-
plated stranded copper connecting leads are insulated
with MIL-W-16878/4 Type E Dupont Teflon (PTFE). consumer and commercial audio applications, the
The voltage increase between each turn of wire use of materials and processes found in aerospace
in a layer depends on the calculated volts-per-turn. applications would probably be cost prohibitive.
At the ends of each layer is a cumulative increase in
voltage to adjacent layers depending on the mean Peak Operating Voltage
magnetic path length in the chosen core and the Operating voltage, sometimes referred to as
number of layers of wire. The inter-layer insulation working voltage, is defined in standards such as
has to account for this. Three-phase 115V AC 400Hz Underwriters Laboratories UL 60950. It is the
aerospace power transformers can be made with a highest voltage to which the insulation or the
secondary voltage as high as 750V AC. Of course, in component under consideration is, or can be,

Consultation • Analysis • Sales

We represent and utilize Klippel and Microflown products,
the most sophisticated analysis and measurement
equipment available.

Contact Warkwyn Today to Learn More!

www.warkwyn.com [email protected]

audioxpress.com | January 2023 | 65

ax Audio Electronics

Photo 15: After transformer

construction is finished,
a high-potential (hi-pot)
voltage test is applied with
special insulation resistance
and high-voltage test
equipment. Shown here is a
Bendix Type 60B4-1-A AC/DC
Insulation and hi-pot tester.

subjected when the equipment is operating under The hi-pot requirement for all commercial
conditions of normal use. transformers with 120V AC or 240V AC primary
Steady-state voltages above 42.4V peak or 60V windings is 1,250V AC to 1,500V AC. The commercial
DC, in computers and similar office equipment, are specs allow for a DC hi-pot voltage, but it must be
generally considered hazardous. increased by a factor of √2 times (1.414) to equal
the AC peak voltage, and applied in each polarity.
High-Potential Testing Insulation resistance tests are performed with at
After transformer construction is finished, a high- least 500V DC.
potential (hi-pot) voltage test is applied with special The UL/CSA/CE conformity marks are required
insulation resistance and high-voltage test equipment to sell all household consumer equipment in the
(Photo 15) that limits the current and disconnects US, Canada, and Europe. UL 1585 determines the
the high voltage in case of a hi-pot failure. dielectric voltage-withstanding test for Class 2 and
This minimizes subsequent arc damage and allows Class 3 transformers.
for possible winding repairs. The test is performed Earlier commercial 115V AC transformer designs
with all the windings shorted together, with the test with secondary windings of 50V AC or less could
voltage applied from windings to metal chassis or place the secondary winding closest to the core
case ground. Hi-pot between secondary and primary since they could be hi-pot tested at the lower 600
windings is applied with the wire end turns of each +50/0V AC maximum limit for 12 seconds allowed at
winding individually shorted to each other. that time. This compromised the primary winding
somewhat in that it had higher resistance than if it
About the Author was placed closest to the transformer core.
Chuck Hansen is an Electrical Engineer and holds five patents in his The MIL-Spec hi-pot limit is 60Hz, 1,000V AC plus
field of electric power engineering. He has worked in both the electric twice the RMS operating voltage. The test current
utility and aerospace electric power industries. He was assigned as the limit level has to be adjusted for the capacitive
Instrument and Controls Startup Group Lead Engineer for two nuclear coupling between the windings. There is no DC
power generating units. He was the Electrical Systems and Controls
hi-pot test specified or required. The insulation
Supervisor for a number of military programs as well as commercial
and business jet electric power programs. He has written two books for resistance test voltage is at least 600V DC.
Audio Amateur publications, and more than 260 magazine articles. He began building vacuum-
tube audio equipment in college, and enjoys restoring and modifying audio equipment and test Next Month
equipment. In his most recent roles as Senior Design Engineer and Lab Calibration Manager, he Part 4 of this series will discuss the various
designed specialized test equipment particular to the aerospace electric power sector.
parameters of transformers and their losses. ax

66 | January 2023 | audioxpress.com

 See What You've
Been Hearing
Dayton Audio AS-1 Audiostrobe Woofer Analyzer

Hand-Held Diagnosis

The Dayton Audio AS-1 can help identify the

cause of response defects and other useful
information of individual drivers. These
include, but are not limited to:
• Non-linearity i.e. rocking motions of the cone
• Cone, surround and dust cap deformations
• Resonance modes in spider and tinsel leads
• Lead deformation
• Spider deformation
• Phase relationships in multi-speaker systems
• Tuning frequency and behavior of passive radiators
• Excursion limits of active and passive drivers

Dayton Audio’s AS-1 provides intermittent illumination of an object for visual analysis up to a frequency of
665Hz. By adjusting the frequency of illumination, objects in vibrational or rotational movement can appear
stationary or in slow motion. This optical illusion, called visual persistence, allows for a non-contact inspection
and status of moving object surfaces well beyond the ability of the naked eye.

Distributed by:
LEARN MORE
Scan here for more
information about
the Dayton Audio
parts-express.com 1-800-338-0531 Audiostrobe or visit
daytonaudio.com
Custom OEM inquiries welcome daytonaudio.com