Zoom out Search Issue

Sven Nordholm, Walter Kellermann,
Simon Doclo, Vesa Välimäki,
Shoji Makino, and John R. Hershey
[
from the
GUEST EDITORS
]
IEEE SIGNAL PROCESSING MAGAZINE [16] MARCH 2015
Signal Processing Techniques for Assisted Listening
N
atural hearing is a desirable
goal in many electronic
communication applica-
tions, such as hearing aids,
audio conferencing, gaming,
and virtual reality applications. The era of
low-power, high-complexity electronics
supports the implementation of computa-
tionally complex algorithms as needed to
provide a more natural listening environ-
ment for the advanced augmentation of
virtual reality and natural content.
As such, assisted listening techniques
provide the means to communicate audio
information from devices to human listen-
ers. The main objective is to provide the
user with a listening experience through
the device that resembles natural hearing
of the sound information. Prominent ap-
plications include virtual augmented au-
dio, hearing aids, and cochlear implants.
But the same techniques are also applica-
ble in other communication applications
such as monaural voice communication,
where additional spatial information can
greatly enhance the listening experience.
In the realm of hearing aids, current de-
vices aim to be more natural both for the
hearing impaired and profoundly deaf by
using new processing techniques. They
even promise to enhance the listening ex-
perience of so-called normal hearing us-
ers. Along with increasingly affordable and
growing computer power, a large variety
of elaborate algorithms for overlayed au-
dio or so-called augmented audio continu-
ously strive toward new applications in
gaming and telepresence to provide a “be-
ing there” experience.
The articles in this special issue of
IEEE Signal Processing Magazine (SPM)
focus on three main aspects of signal
processing in this domain: audio en-
hancement, presentation/rendering, and
evaluation. To limit the scope in this spe-
cial issue, machine-learning techniques
have been excluded. While it is under-
stood that future systems for assisted lis-
tening will greatly be influenced by
machine-learning-based algorithms, an-
other special issue dedicated to this de-
velopment can already be envisioned.
Audio signal enhancement, particular-
ly of speech signals, has a long research
tradition and still dominates the scene,
and, consequently, is also the main topic
of this special issue. Techniques for sin-
gle-channel and multichannel signal en-
hancement techniques play a preeminent
role in telecommunication, hearing aids,
and augmented headsets. Accordingly,
fundamental problems and state-of-the-
art techniques are presented in the article
“Multichannel Signal Enhancement Algo-
rithms for Assisted Listening Devices” by
Doclo et al. Beyond the description of the
generic algorithms, this article emphasiz-
es the specific problems and solutions for
hearing aids and headsets addressing both
the signal acquisition and the binaural
rendering aspect.
As a special technique for capturing and
describing the spatial information relevant
for assisted listening, multichannel tech-
niques that estimate the direct path infor-
mation and suppress a combination of
reverberation and diffuse noise are provid-
ed in the article “Parametric Spatial Sound
Processing” by Kowalczyk et al.
Two articles provide overviews on high-
ly relevant aspects of single-channel en-
hancement techniques: “Optimizing
Speech Intelligibility in a Noisy Environ-
ment” by Kleijn et al. focuses on techniques
for improving speech intelligibility using
perceptual criteria and auditory modeling,
and “Phase Processing for Single-Channel
Speech Enhancement” by Gerkmann et al.
provides a survey of techniques that utilize
both amplitude and phase information for
speech enhancement.
Processing and coding signals for cochle-
ar implants is addressed in the article “Sound
Coding in Cochlear Implants” by Wouters
et al. This article describes signal processing
techniques used in cochlear implants to
map the information extracted from an au-
dio signal onto cochlea excitation that a
profoundly deaf person can understand.
Rendering of audio aims at providing
an immersive, undisturbed listening expe-
rience for recorded information via loud-
speakers or headsets with typical applica-
tions demanding high-quality sound
reproduction, such as, e.g., home theaters,
gaming, or telepresence systems. Betlehem
et al. provide an overview of techniques to
deliver audio information to multiple lis-
teners via loudspeakers in their article
“Personal Sound Zones.” These techniques
also have applications in providing audio in
public areas without disturbing the sur-
roundings. Then the natural sound in the
environment is augmented by the rendered
audio. A similar concept to augment out-
side information using personal headsets is
presented by Välimäki et al. in “Assisted
Listening Using a Headset,” which also
Digital Object Identifier 10.1109/MSP.2014.2375972
Date of publication: 12 February 2015
THE MAIN OBJECTIVE
OF ASSISTED LISTENING
TECHNIQUES IS TO
PROVIDE THE USER
WITH A LISTENING
EXPERIENCE THROUGH
THE DEVICE THAT
RESEMBLES NATURAL
HEARING OF THE SOUND
INFORMATION.
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page
q
q
M
M
q
q
M
M
q
M
THE WORLD’S NEWSSTAND
®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page
q
q
M
M
q
q
M
M
q
M
THE WORLD’S NEWSSTAND
®