Method and Apparatus for Enhancement of Audio
Reconstruction
Field of the invention
The present invention relates to techniques as to how to
improve the perception of a direction of origin of a
reconstructed audio signal. In particular, the present
invention proposes an apparatus and a method for
reproduction of recorded audio signals such that a
selectable direction of audio sources can be emphasized or
over-weighted with respect to audio signals coming from
other directions.
Background of the invention and prior art
Generally, in multi-channel reproduction and listening, a
listener is surrounded by multiple loudspeakers. Various
methods exist to capture audio signals for specific set-
ups. One general goal in the reproduction is to reproduce
the spatial composition of the originally recorded signal,
i.e. the origin of individual audio source, such as the
location of a trumpet within an orchestra. Several
loudspeaker set-ups are fairly common and can create
different spatial impressions. Without using special post-
production techniques, the commonly known two-channel
stereo set-ups can only recreate auditory events on a line
between the two loudspeakers. This is mainly achieved by
so-called "amplitude-panning", where the amplitude of the
signal associated to one audio source is distributed
between the two loudspeakers, depending on the position of
the audio source with respect to the loudspeakers. This is
usually done during recording or subsequent mixing. That
is, an audio source coming from the far-left with respect
to the listening position will be mainly reproduced by the
left loudspeaker, whereas an audio source in front of the
listening position will be reproduced with identical

amplitude (level) by both loudspeakers. However, sound
emanating from other directions cannot be reproduced.
Consequently, by using more loudspeakers that are
positioned around the listener, more directions can be
covered and a more natural spatial impression can be
created. The probably most well known multi-channel
loudspeaker layout is the 5.1 standard (ITU-R775-1), which
consists of 5 loudspeakers, whose azimuthal angles with
respect to the listening position are predetermined to be
0°, ±30° and ±110°. That means, that during recording or
mixing the signal is tailored to that specific loudspeaker
configuration and deviations of a reproduction set-up from
the standard will result in decreased reproduction quality.
Numerous other systems with varying numbers of loudspeakers
located at different directions have also been proposed.
Professional and special systems, especially in theaters
and sound installations, also include loudspeakers at
different heights.
According to the different reproduction set-ups, several
different recording methods have been designed and proposed
for the previously mentioned loudspeaker systems, in order
to record and reproduce the spatial impression in the
listening situation as it would have been perceived in the
recording environment. A theoretically ideal way of
recording spatial sound for a chosen multi-channel
loudspeaker system would be to use the same number of
microphones as there are loudspeakers. In such a case, the
directivity patterns of the microphones should also
correspond to the loudspeaker layout, such that sound from
any single direction would only be recorded with a small
number of microphones (1, 2 or more). Each microphone is
associated to a specific loudspeaker. The more loudspeakers
are used in reproduction, the narrower the directivity
patterns of the microphones have to be. However, narrow
directional microphones are rather expensive and typically

have a non-flat frequency response, degrading the quality
of the recorded sound in an undesirable manner.
Furthermore, using several microphones with too broad
directivity patterns as input to multi-channel reproduction
results in a colored and blurred auditory perception due to
the fact that sound emanating from a single direction would
always be reproduced with more loudspeakers than necessary,
as it would be recorded with microphones associated to
different loudspeakers. Generally, currently available
microphones are best suited for two-channel recordings and
reproductions, that is, these are designed without the goal
of a reproduction of a surrounding spatial impression.
From the point of view of microphone-design, several
approaches have been discussed to adapt the directivity
patterns of microphones to the demands in spatial-audio-
reproduction. Generally, all microphones capture sound
differently depending on the direction of arrival of the
sound to the microphone. That is, microphones have a
different sensitivity, depending on the direction of
arrival of the recorded sound. In some microphones, this
effect is minor, as they capture sound almost independently
of the direction. These microphones are generally called
omnidirectional microphones. In a typical microphone
design, a circular diaphragm is attached to a small
airtight enclosure. If the diaphragm is not attached to the
enclosure and sound reaches it equally from each side, its
directional pattern has two lobes. That is, such a
microphone captures sound with equal sensitivity from both
front and back of the diaphragm, however, with inverse
polarities. Such a microphone does not capture sound coming
from the direction coincident to the plane of the
diaphragm, i.e. perpendicular to the direction of maximum
sensitivity. Such a directional pattern is called dipole,
or figure-of-eight.
Omnidirectional microphones may also be modified into
directional microphones, using a non-airtight enclosure for

the microphone. The enclosure is especially constructed
such, that the sound waves are allowed to propagate through
the enclosure and reach the diaphragm, wherein some
directions of propagation are preferred, such that the
directional pattern of such a microphone becomes a pattern
between omnidirectional and dipole. Those patterns may, for
example, have two lobes. However, the lobes may have
different strength. Some commonly known microphones have
patterns that have only one single lobe. The most important
example is the cardioid pattern, where the directional
function D can be expressed as D = 1 + cos (9), 9 being the
direction of arrival of sound. The directional function
thus quantifies, what fraction of the incoming sound
amplitude is captured, depending on the direction.
The previously discussed omnidirectional patterns are also
called zeroth-order patterns and the other patterns
mentioned previously (dipole and cardioid) are called
first-order patterns. All previously discussed microphone
designs do not allow arbitrary shaping of the directivity
patterns, since their directivity pattern is entirely
determined by their mechanical construction.
To partly overcome this problem, some specialized
acoustical structures have been designed, which can be used
to create narrower directional patterns than those of
first-order microphones. For example, when a tube with
holes in it is attached to an omnidirectional microphone, a
microphone with narrow directional pattern can be created.
These microphones are called shotgun or rifle microphones.
However, they typically do not have a flat frequency
response, that is, the directivity pattern is narrowed at
the cost of the quality of the recorded sound. Furthermore,
the directivity pattern is predetermined by the geometric
construction and, thus, the directivity pattern of a
recording performed with such a microphone cannot be
controlled after the recording.

Therefore, other methods have been proposed to partly allow
to alter the directivity pattern after the actual
recording. Generally, this relies on the basic idea of
recording sound with an array of omnidirectional or
directional microphones and to apply signal processing
afterwards. Various such techniques have been recently
proposed. A fairly simple example is to record sound with
two omnidirectional microphones, which are placed close to
each other, and to subtract both signals from each other.
This creates a virtual microphone signal having a
directional pattern equivalent to a dipole.
In other, more sophisticated schemes the microphone signals
can also be delayed or filtered before summing them up.
Using beam forming, a technique also known from wireless
LAN, a signal corresponding to a narrow beam is formed by
filtering each microphone signal with a specially designed
filter and summing the signals up after the filtering
(filter-sum beam forming). However, these techniques are
blind to the signal itself, that is, they are not aware of
the direction of arrival of the sound. Thus, a
predetermined directional pattern must be defined, which is
independent of the actual presence of a sound source in the
predetermined direction. Generally, estimation of the
"direction of arrival" of sound is a task of its own.
Generally, numerous different spatial directional
characteristics can be formed with the above techniques.
However, forming arbitrary spatially selective sensitivity
patterns (i.e. forming narrow directional patterns)
requires a large number of microphones.
An alternative way to create multi-channel recordings is to
locate a microphone close to each sound source (e.g. an
instrument) to be recorded and recreate the spatial
impression by controlling the levels of the close-up
microphone signals in the final mix. However, such a system

demands a large number of microphones and a lot of user
interaction in creating the final down-mix.
A method to overcome the above problem has been recently
proposed and is called directional audio coding (DirAC),
which may be used with different microphone systems and
which is able to record sound for reproduction with
arbitrary loudspeaker set-ups. The purpose of DirAC is to
reproduce the spatial impression of an existing acoustical
environment as precisely as possible, using a multi-channel
loudspeaker system having an arbitrary geometrical set-up.
Within the recording environment, the responses of the
environment (which may be continuous recorded sound or
impulse responses) are measured with an omnidirectional
microphone (W) and with a set of microphones allowing to
measure the direction of arrival of sound and the
diffuseness of sound. In the following paragraphs and
within the application, the term "diffuseness" is to be
understood as a measure for the non-directivity of sound.
That is, sound arriving at the listening or recording
position with equal strength from all directions, is
maximally diffuse. A common way of quantifying diffusion is
to use diffuseness values from the interval [0,...,1],
wherein a value of 1 describes maximally diffuse sound and
a value of 0 describes perfectly directional sound, i.e.
sound arriving from one clearly distinguishable direction
only. One commonly known method of measuring the direction
of arrival of sound is to apply 3 figure-of-eight
microphones (XYZ) aligned with Cartesian coordinate axes.
Special microphones, so-called "SoundField microphones",
have been designed, which directly yield all desired
responses. However, as mentioned above, the W, X, Y and Z
signals may also be computed from a set of discrete
omnidirectional microphones.
In DirAC analysis, a recorded sound signal is divided into
frequency channels, which correspond to the frequency
selectivity of human auditory perception. That is, the

signal is, for example, processed by a filter bank or a
Fourier-transform to divide the signal into numerous
frequency channels, having a bandwidth adapted to the
frequency selectivity of the human hearing. Then, the
frequency band signals are analyzed to determine the
direction of origin of sound and a diffuseness value for
each frequency channel with a predetermined time
resolution. This time resolution does not have to be fixed
and may, of course, be adapted to the recording
environment. In DirAC, one or more audio channels are
recorded or transmitted, together with the analyzed
direction and diffuseness data.
In synthesis or decoding, the audio channels finally
applied to the loudspeakers can be based on the
omnidirectional channel W (recorded with a high quality due
to the omnidirectional directivity pattern of the
microphone used) , or the sound for each loudspeaker may be
computed as a weighted sum of W, X, Y and Z, thus forming a
signal having a certain directional characteristic for each
loudspeaker. Corresponding to the encoding, each audio
channel is divided into frequency channels, which are
optionally furthermore divided into diffuse and non-diffuse
streams, depending on analyzed diffuseness. If diffuseness
has been measured to be high, a diffuse stream may be
reproduced using a technique producing a diffuse perception
of sound, such as the decorrelation techniques also used in
Binaural Cue Coding. Non-diffuse sound is reproduced using
a technique aiming to produce a point-like virtual audio
source, located in the direction indicated by the direction
data found in the analysis, i.e. the generation of the
DirAC signal. That is, spatial reproduction is not tailored
to one specific, "ideal" loudspeaker set-up, as in the
prior art techniques (e.g. 5.1). This is particularly the
case, as the origin of sound is determined as direction
parameters (i.e. described by a vector) using the knowledge
about the directivity patterns on the microphones used in
the recording. As already discussed, the origin of sound in

3-dimensional space is parameterized in a frequency
selective manner. As such, the directional impression may
be reproduced with high quality for arbitrary loudspeaker
set-ups, as far as the geometry of the loudspeaker set-up
is known. DirAC is therefore not limited to special
loudspeaker geometries and generally allows for a more
flexible spatial reproduction of sound.
The US patent application 5,812,674 relates to a method for
the simulation of the acoustical quality produced by a
virtual sound source and for the localizing of this source
with respect to one or more listeners. To achieve the
desired natural reproduction, perceptual parameters
defining the space, the desired acoustical quality and the
localization of a virtual sound source are used. These
values are used to compute a pulse response, which is
described by its energy distribution as a function of time
and frequency. A context compensation is performed in order
to take account for the room effects and an artificial
reverberation of elementary signals is computed, based on
the description of the room. Once the room acoustics are
determined in the previously described manner, recorded
sound samples may be post processed in order to sound as if
they were recorded within the artificially created room.
Although numerous techniques have been developed to
reproduce multi-channel audio recordings and to record
appropriate signals for a later multi-channel reproduction,
none of the prior art techniques allows to influence an
already recorded signal such that a direction of origin of
audio signals can be emphasized during reproduction such
that, for example, the intelligibility of the signal from
one distinct desired direction may be enhanced.
Summary of the invention
According to one embodiment of the present invention, an
audio signal having at least one audio channel and

associated direction parameters indicating the direction of
origin of a portion of the audio qhannel with respect to a
recording position can be reconstructed allowing for an
enhancement of the perceptuality of the signal coming from
a distinct direction or from numerous distinct directions.
That is, in reproduction, a desired direction of origin
with respect to the recording position can be Selected.
While deriving a reconstructed portion of the reconstructed
audio signal, the portion of the audio channel is modified
such that the intensity of portions of the audio channel
having direction parameters indicating a direction of
origin close to the desired direction of origin are
increased with respect to other portions of the audio
channel having direction parameters indicating a direction
of origin further away from the desired direction of
origin. Directions of origin of portions of an audio

channel or a multi-channel signal can be emphasized, such
as to allow for a better perception of audio objects, which
were located in the selected direction during the
recording.
According to a further embodiment of the present invention,
a user may choose during reconstruction, which direction or
which directions shall be emphasized such that portions of
the audio channel or portions of multiple audio channels,
which are associated to that chosen direction are
emphasized, i.e. their intensity or amplitude is increased
with respect to the remaining portions. According to an
embodiment, emphasis or attenuation of sound from a
specific direction can be done with a much sharper spatial
resolution than with systems not implementing direction
parameters. According to a further embodiment of the
present invention, arbitrary spatial weighting functions
can be specified, which cannot be achieved with regular
microphones. Furthermore, the weighting functions may be
time and frequency variant, such that further embodiments
of the present invention may be used with high flexibility.
Furthermore, the weighting functions are extremely easy to
implement and to update, since these have only to be loaded
into the system instead of exchanging hardware (for
example, microphones).
According to a further embodiment of the present invention,
audio signals having associated a diffuseness parameter,
the diffuseness parameter indicating a diffuseness of the
portion of the audio channel, are reconstructed such that
an intensity of a portion of the audio channel with high
diffuseness is decreased with respect to another portion of
the audio channel having associated a lower diffuseness.
Thus, in reconstructing an audio signal, diffuseness of
individual portions of the audio signal can be taken into
account to further increase the directional perception of
the reconstructed signal. This may, additionally, increase

the redistribution of audio sources with respect to
techniques only using diffuse sound portions to increase
the overall diffuseness of the signal rather than making
use of the diffuseness information for a better
redistribution of the audio sources. Note that the present
invention also allows to conversely emphasize portions of
the recorded sound that are of diffuse origin, such as
ambient signals.
According to a further embodiment, at least one audio
channel is up-mixed to multiple audio channels. The
multiple audio channels might correspond to the number of
loudspeakers available for playback. Arbitrary loudspeaker
set-ups may be used to enhance the redistribution of audio
sources while it can be guaranteed that the direction of
the audio source is always reproduced as good as possible
with the existing equipment, irrespective of the number of
loudspeakers available.
According to another embodiment of the present invention,
reproductions may even be performed via a monophonic
loudspeaker. Of course, the direction of origin of the
signal will, in that case, be the physical location of the
loudspeaker. However, by selecting a desired direction of
origin of the signal with respect to the recording
position, the audibility of the signal stemming from the
selected direction can be significantly increased, as
compared to the playback of a simple down-mix.
According to a further embodiment of the present invention,
the direction of origin of the signal can be accurately
reproduced, when one or more audio channels are up-mixed to
the number of channels corresponding to the loudspeakers.
The direction of origin can be reconstructed as good as
possible by using, for example, amplitude panning
techniques. To further increase the perceptual quality,
additional phase shifts may be introduced, which are also
dependent on the selected direction.

Certain embodiments of the present invention may
additionally decrease the cost of the microphone capsules
for recording the audio signal without seriously affecting
the audio quality, since at least the microphone used to
determine the direction/diffusion estimate does not
necessarily need to have a flat frequency response.
Brief description of the drawings
Several embodiments of the present invention will in the
following be described referencing the enclosed drawings.
Fig. 1 shows an embodiment of a method for reconstructing
an audio signal;
Fig. 2 shows a block diagram of an apparatus for
reconstructing an audio signal; and
Fig. 3 shows a block diagram of a further embodiment;
Fig. 4 shows an example of the application of an inventive
method or an inventive apparatus in a teleconferencing
scenario;
Fig. 5 shows an embodiment of a method for enhancing a
directional perception of an audio signal;
Fig. 6 shows an embodiment of a decoder for reconstructing
an audio signal; and
Fig. 7 shows an embodiment of a system for enhancing a
directional perception of an audio signal.
Detailed description of preferred embodiments
Fig. 1 shows an embodiment of a method for reconstructing
an audio signal having at least one audio channel and

associated direction parameters indicating a direction of
origin of a portion of the audio channel with respect to a
recording position. In a selection step 10, a desired
direction of origin with respect to the recording position
is selected for a reconstructed portion of the
reconstructed audio signal, wherein the reconstructed
portion corresponds to a portion of the audio channel. That
is, for a signal portion to be processed, a desired
direction of origin, from which signal portions shall be
clearly audible after reconstruction, is selected. The
selection can be done directly by a user input or
automatically, as detailed below.
The portion may be a time portion, a frequency portion, or
a time portion of a certain frequency interval of an audio
channel. In a modification step 12, the portion of the
audio channel is modified for deriving the reconstructed
portion of the reconstructed audio signal, wherein the
modification comprises increasing an intensity of a portion
of the audio channel having direction parameters indicating
a direction of origin close to the desired direction of
origin with respect to another portion of the audio channel
having direction parameters indicating a direction of
origin further away from the desired direction of origin.
That is, such portions of the audio channel are emphasized
by increasing their intensity or level, which can, for
example, be implemented by the multiplication of a scaling
factor to the portion of the audio channel. According to an
embodiment, portions originating from a direction close to
the selected (desired) direction are multiplied by large
scale factors, to emphasize these signal portions in
reconstruction and to improve the audibility of those audio
recorded objects, in which the listener is interested in.
Generally, in the context, of this application, increasing
the intensity of a signal or a channel shall be understood
as any measure which renders the signal to be better
audible. This could for example be increasing the signal
amplitude, the energy carried by the signal or multiplying

the signal with a scale factor greater than unity.
Alternatively, the loudness of competing signals may be
decreased to achieve the effect.
The selection of the desired direction may be directly
performed via a user interface by a user at the listening
site. However, according to alternative embodiments, the
selection can be performed automatically, for example, by
an analysis of the directional parameters, such that
frequency portions having roughly the same origin are
emphasized, whereas the remaining portions of the audio
channel are suppressed. Thus, the signal can be
automatically focused on the predominant audio sources,
without requiring an additional user input at the listening
end.
According to further embodiments, the selection step is
omitted, since a direction of origin has been set. That is,
the intensity of a portion of the audio channel having
direction parameters indicating a direction of origin close
to the set direction is increased. The set direction may,
for example be hardwired, i.e. the direction may be
predetermined. If, for example only the central talker in a
teleconferencing scenario is of interest, this can be
implemented using a predetermined set direction.
Alternative embodiments may read the set direction from a
memory which may also have stored a number of alternative
directions to be used as set directions. One of these may,
for example, be read when turning on an inventive
apparatus.
According to an alternative embodiment, the selection of
the desired direction may also be performed at the encoder
side, i.e. at the recording of the signal, such that
additional parameters are transmitted with the audio
signal, indicating the desired direction for reproduction.
Thus, a spatial perception of the reconstructed signal may

already be selected at the encoder without the knowledge on
the specific loudspeaker set-up used for reproduction.
Since the method for reconstructing an audio signal is
independent of the specific loudspeaker set-up intended to
reproduce the reconstructed audio signal, the method may be
applied to monophonic as well as to stereo or multi-channel
loudspeaker configurations. That is, according to a further
embodiment, the spatial impression of a reproduced
environment is post-processed to enhanced the
perceptibility of the signal.
When used for monophonic playback, the effect may be
interpreted as recording the signal with a new type of
microphone capable of forming arbitrary directional
patterns. However, this effect can be fully achieved at the
receiving end, i.e. during playback of the signal, without
changing anything in the recording set-up.
Fig. 2 shows an embodiment of an apparatus (decoder) for
reconstruction of an audio signal, i.e. an embodiment of a
decoder 20 for reconstructing an audio signal. The decoder
20 comprises a direction selector 22 and an audio portion
modifier 24. According to the embodiment of Fig. 2 a multi-
channel audio input 26 recorded by several microphones is
analyzed by a direction analyzer 28 which derives direction
parameters indicating a direction of origin of a portion of
the audio channels, i.e. the direction of origin of the
signal portion analyzed. According to one embodiment of the
present invention, the direction, from which most of the
energy is incident to the microphone is chosen. The
recording position is determined for each specific signal
portion. This can, for example, be also done using the
DirAC-microphone-techniques previously described. Of
course, other directional analysis method based on recorded
audio information may be used to implement the analysis. As
a result, the direction analyzer 28 derives direction
parameters 30, indicating the direction of origin of a

portion of an audio channel or of the multi-channel signal
26. Furthermore, the directional analyzer 28 may be
operative to derive a diffuseness parameter 32 for each
signal portion (for example, for each frequency interval or
for each time-frame of the signal).
The direction parameter 30 and, optionally, the diffuseness
parameter 32 are transmitted to the direction selector 22
which is implemented to select a desired direction of
origin with respect to a recording position for a
reconstructed portion of the reconstructed audio signal.
Information on the desired direction is transmitted to the
audio portion modifier 24. The audio portion modifier 24
receives at least one audio channel 34, having a portion,
for which the direction parameters have been derived. The
at least one channel modified by audio portion modifier
may, for example, be a down-mix of the multi-channel signal
26, generated by conventional multi-channel down-mix
algorithms. One extremely simple case would be the direct
sum of the signals of the multi-channel audio input 26.
However, as the inventive embodiments are not limited by
the number of input channels, in an alternative embodiment,
all audio input channels 26 can be simultaneously processed
by audio decoder 20.
The audio portion modifier 24 modifies the audio portion
for deriving the reconstructed portion of the reconstructed
audio signal, wherein the modifying comprises increasing an
intensity of a portion of the audio channel having
direction parameters indicating a direction of origin close
to the desired direction of origin with respect to another
portion of the audio channel having direction parameters
indicating a direction of origin further away from the
desired direction of origin. In the example of Fig. 2, the
modification is performed by multiplying a scaling factor
36 (q) with the portion of the audio channel to be
modified. That is, if the portion of the audio channel is
analyzed to be originating from a direction close to the

selected desired direction, a large scaling factor 36 is
multiplied with the audio portion. Thus, at its output 38,
the audio portion modifier outputs a reconstructed portion
of the reconstructed audio signal corresponding to the
portion of the audio channel provided at its input. As
furthermore indicated by the dashed lines at the output 38
of the audio portion modifier 24, this may not only be
performed for a mono-output signal, but also for multi-
channel output signals, for which the number of output
channels is not fixed or predetermined.
In other words, the embodiment of the audio decoder 20
takes its input from such directional analysis as, for
example, used in DirAC. Audio signals 26 from a microphone
array may be divided into frequency bands according to the
frequency resolution of the human auditory system. The
direction of sound and, optionally, diffuseness of sound is
analyzed depending on time in each frequency channel. These
attributes are delivered further as, for example, direction
angles azimuth (azi) and elevation (ele), and as
diffuseness index Psi, which varies between zero and one.
Then, the intended or selected directional characteristic
is imposed on the acquired signals by using a weighting
operation on them, which depends on the direction angles
(azi and/or ele) and, optionally, on the diffuseness (Psi).
Evidently, this weighting may be specified differently for
different frequency bands, and will, in general, vary over
time.
Fig. 3 shows a further embodiment of the present invention,
based on DirAC synthesis. In that sense, the embodiment of
Fig. 3 could be interpreted to be an enhancement of DirAC
reproduction, which allows to control the level of sound
depending on analyzed direction. This makes it possible to
emphasize sound coming from one or multiple directions, or
to suppress sound from one or multiple directions. When
applied in multi-channel reproduction, a post-processing of

the reproduced sound image is achieved. If only one channel
is used as output, the effect is equivalent to the use of a
directional microphone with arbitrary directional patterns
during recording of the signal. In the embodiment shown in
Fig. 3, the derivation of direction parameters, as well as
the derivation of one transmitted audio channel is shown.
The analysis is performed based on B-format microphone
channels W, X, Y and Z, as, for example, recorded by a
sound field microphone.
The processing is performed frame-wise. Therefore, the
continuous audio signals are divided into frames, which are
scaled by a windowing function to avoid discontinuities at
the frame boundaries. The windowed signal frames are
subjected to a Fourier transform in a Fourier transform
block 40, dividing the microphone signals into N frequency
bands. For the sake of simplicity, the processing of one
arbitrary frequency band shall be described in the
following paragraphs, as the remaining frequency bands are
processed equivalently. The Fourier transform block 40
derives coefficients describing the strength of the
frequency components present in each of the B-format
microphone channels W, X, Y, and Z within the analyzed
windowed frame. These frequency parameters 42 are input
into audio encoder 44 for deriving an audio channel and
associated direction parameters. In the embodiment shown in
Fig. 3, the transmitted audio channel is chosen to be the
omnidirectional channel 4 6 having information on the signal
from all directions. Based on the coefficients 42 for the
omnidirectional and the directional portions of the B-
format microphone channels, a directional and diffuseness
analysis is performed by a direction analysis block 48.
The direction of origin of sound for the analyzed portion
of the audio channel 4 6 is transmitted to an audio decoder
50 for reconstructing the audio signal together with the
omnidirectional channel 46. When diffuseness parameters 52
are present, the signal path is split into a non-diffuse

path 54a and a diffuse path 54b. The non-diffuse path 54a
is scaled according to the diffuseness parameter, such
that, when diffuseness ψ is high, most of the energy or of
the amplitude will remain in the non-diffuse path.
Conversely, when the diffuseness is high, most of the
energy will be shifted to the diffuse path 54b. In the
diffuse path 54b, the signal is decorrelated or diffused
using decorrelators 56a or 56b. Decorrelation can be
performed using conventionally known techniques, such as
convolving with a white noise signal, wherein the white
noise signal may differ from frequency channel to frequency
channel. As long as the decorrelation is energy preserving,
a final output can be regenerated by simply adding the
signals of the non-diffuse signal path 54a and the diffuse
signal path 54b at the output, since the signals at the
signal paths have already been scaled, as indicated by the
diffuseness parameter ψ. The diffuse signal path 54b may be
scaled, depending on the number of loudspeakers, using an
appropriate scaling rule. For example, the signals in the
diffuse path may be scaled by when N is the number
of loudspeakers.
When the reconstruction is performed for a multi-channel
set-up, the direct signal path 54a as well as the diffuse
signal path 54b are split up into a number of sub-paths
corresponding to the individual loudspeaker signals (at
split up positions 58a and 58b) . To this end, the split up
at the split position 58a and 58b can be interpreted to be
equivalent to an up-mixing of the at least one audio
channel to multiple channels for a playback via a
loudspeaker system having multiple loudspeakers. Therefore,
each of the multiple channels has a channel portion of the
audio channel 46. The direction of origin of individual
audio portions is reconstructed by redirection block 60
which additionally increases or decreases the intensity or
the amplitude of the channel portions corresponding to the
loudspeakers used for playback. To this end, redirection
block 60 generally requires knowledge about the loudspeaker

setup used for playback. The actual redistribution
(redirection) and the derivation of the associated
weighting factors can, for example, be implemented using
techniques as vector based amplitude panning. By supplying
different geometric loudspeaker setups to the
redistribution block 60, arbitrary configurations of
playback loudspeakers can be used to implement the
inventive concept, without a loss of reproduction quality.
After the processing, multiple inverse Fourier transforms
are performed on frequency domain signals by inverse
Fourier transform blocks 62 to derive a time domain signal,
which can be played back by the individual loudspeakers.
Prior to the playback, an overlap and add technique must be
performed by summation units 64 to concatenate the
individual audio frames to derive continuous time domain
signals, ready to be played back by the loudspeakers.
According to the embodiment of the invention shown in Fig.
3, the signal processing of Dir-AC is amended in that an
audio portion modifier 66 is introduced to modify the
portion of the audio channel actually processed and which
allows to increase an intensity of a portion of the audio
channel having direction parameters indicating a direction
of origin close to a desired direction. This is achieved by
application of an additional weighting factor to the direct
signal path. That is, if the frequency portion processed
originates from the desired direction, the signal is
emphasized by applying an additional gain to that specific
signal portion. The application of the gain can be
performed prior to the split point 58a, as the effect shall
contribute to all channel portions equally.
The application of the additional weighting factor can, in
an alternative embodiment, also be implemented within the
redistribution block 60 which, in that case, applies
redistribution gain factors increased or decreased by the
additional weighting factor.

When using directional enhancement in reconstruction of a
multi-channel signal, reproduction can, for example, be
performed in the style of DirAC rendering, as shown in Fig.
3. The audio channel to be reproduced is divided into
frequency bands equal to those used for the directional
analysis. These frequency bands are then divided into
streams, a diffuse and a non-diffuse stream. The diffuse
stream is reproduced, for example, by applying the sound to
each loudspeaker after convolution with 30ms wide noise
bursts. The noise bursts are different for each
loudspeaker. The non-diffuse stream is applied to the
direction delivered from the directional analysis which is,
of course, dependent on time. To achieve a directional
perception in multi-channel loudspeaker systems, simple
pair-wise or triplet-wise amplitude panning may be used.
Furthermore, each frequency channel is multiplied by a gain
factor or scaling factor, which depends on the analyzed
direction. In general terms, a function can be specified,
defining a desired directional pattern for reproduction.
This can, for example, be only one single direction, which
shall be emphasized. However, arbitrary directional
patterns are easily implementable with the embodiment of
Fig. 3.
In the following approach, a further embodiment of the
present invention is described as a list of processing
steps. The list is based on the assumption that sound is
recorded with a B-format microphone, and is then processed
for listening with multi-channel or monophonic loudspeaker
set-ups using DirAC style rendering or rendering supplying
directional parameters, indicating the direction of origin
of portions of the audio channel. The processing is as
follows:
1. Divide microphone signals into frequency bands and
analyze direction and, optionally, diffuseness at
each band depending on frequency. As an example,

direction may be parameterized by an azimuth and an
elevation angle (azi, ele).
2. Specify a function F, which describes the desired
directional pattern. The function may have an
arbitrary shape. It typically depends on direction.
It may, furthermore, also depend on diffuseness, if
diffuseness information is available. The function
can be different for different frequencies and it
may also be altered depending on time. At each
frequency band, derive a directional factor q from
the function F for each time instance, which is used
for subsequent weighting (scaling) of the audio
signal.
3. Multiply the audio sample values with the q values
of the directional factors corresponding to each
time and frequency portion to form the output
signal. This may be done in a time and/or a
frequency domain representation. Furthermore, this
processing may, for example, be implemented as a
part of a DirAC rendering to any number of desired
output channels.
As previously described, the result can be listened to
using a multi-channel or a monophonic loudspeaker system.
Fig. 4 shows an illustration as to how the inventive
methods and apparatuses may be utilized to strongly
increase the perceptibility of a participant within in a
teleconferencing scenario. On the recording side 100, four
talkers 102a-102d are illustrated which have a distinct
orientation with respect to a recording position 104. That
is, an audio signal originating from talker 102c has a
fixed direction of origin with respect to the recording
position 104. Assuming the audio signal recorded at
recording position 104 has a contribution from talker 102c
and some "background" noise originating, for example, from
a discussion of talkers 102a and 102b, a broadband signal

recorded and transmitted to a listening site 110 will
comprise both signal components.
As an example, a listening set-up having six loudspeakers
112a-112f is sketched which surround a listener located at
a listening position 114. Therefore, in principle, sound
emanating from almost arbitrary positions around the
listener 114 can be reproduced by the set-up sketched in
Fig. 4. Conventional multi-channel systems would reproduce
the sound using these six speakers 112a-112f to reconstruct
the spatial perception experienced at the recording
position 104 during recording as closely as possible.
Therefore, when the sound is reproduced using conventional
techniques, also the contribution of talker 102c as the
"background" of the discussing talkers 102a and 102b would
be clearly audible, decreasing the intelligibility of the
signal of talker 102c.
According to an embodiment of the present invention, a
direction selector can be used to select a desired
direction of origin with respect to the recording position
which is used for a reconstructed version of a
reconstructed audio signal which is to be played back by
the loudspeakers 112a-112f. Therefore, a listener 114 can
select the desired direction 116, corresponding to the
position of talker 102c. Thus, the audio portion modifier
can modify the portion of the audio channel to derive the
reconstructed portion of the reconstructed audio signal
such that the intensity of the portions of the audio
channel originating from a direction close to the selected
direction 116 are emphasized. The listener may, at the
receiving end, decide which direction of origin shall be
reproduced. Having made this selection, only those signal
portions are emphasized which originate from the direction
of talker 102c and thus, the discussing talkers 102a and
102b will become less disturbing. Apart from emphasizing
the signal from the selected direction, the direction may
be reproduced by amplitude panning, as symbolically

indicated by wave forms 120a and 120b. As talkers 102c
would be located closer to loudspeaker 112d than to
loudspeaker 112c, amplitude panning will lead to a
reproduction of the emphasized signal via loudspeakers 112c
and 112d, whereas the remaining loudspeakers will be nearly
quiet (eventually playing back diffuse signal portions).
Amplitude panning will increase the level of loudspeaker
112d with respect to loudspeaker 112c, as talker 102c is
located closer to loudspeaker 112d.
Fig. 5 illustrates a block diagram of an embodiment of a
method for enhancing a directional perception of an audio
signal. In a first analysis step 150, at least one audio
channel and associated direction parameters indicating a
direction of origin of a portion of the audio channel with
respect to a recording position are derived.
In a selection step 152, a desired direction of origin with
respect to the recording position is selected for a
reconstructed portion of the reconstructed audio signal,
the reconstructed portion corresponding to a portion of the
audio channel.
In a modification step 154, the portion of the audio
channel is modified to derive the reconstructed portion of
the reconstructed audio signal, wherein the modification
comprises increasing an intensity of a portion of the audio
channel having direction parameters indicating a direction
of origin close to the desired direction of origin with
respect to another portion of the audio channel, having
direction parameters indicating a direction of origin
further away from the desired direction of origin.
Fig. 6 illustrates an embodiment of an audio decoder for
reconstructing an audio signal having at least one audio
channel 160 and associated direction parameters 162
indicating a direction of origin of a portion of the audio
channel with respect to a recording position.

The audio decoder 158 comprises a direction selector 164
for selecting a desired direction of origin with respect to
the recording position for a reconstructed portion of the
reconstructed audio signal, the reconstructed portion
corresponding to a portion of the audio channel. The
decoder 158 further comprises an audio portion modifier 166
for modifying the portion of the audio channel for deriving
the reconstructed portion of the reconstructed audio
signal, wherein the modification comprises increasing an
intensity of a portion of the audio channel having
direction parameters indicating a direction of origin close
to the desired direction of origin with respect to another
portion of the audio channel having direction parameters
indicating a direction of origin further away from the
desired direction of origin.
As indicated in Fig. 6, a single reconstructed portion 168
may be derived or multiple reconstructed portions 170 may
simultaneously be derived, when the decoder is used in a
multi-channel reproduction set-up. The embodiment of a
system for enhancement of a directional perception of an
audio signal 180, as shown in Fig. 7 is based on decoder
158 of Fig. 6. Therefore, in the following, only the
additionally introduced elements will be described. The
system for enhancement of a directional perception of an
audio signal 180 receives an audio signal 182 as an input,
which may be a monophonic signal or a multi-channel signal
recorded by multiple microphones. An audio encoder 184
derives an audio signal having at least one audio channel
160 and associated direction parameters 162 indicating a
direction of origin of a portion of the audio channel with
respect to a recording position. The at least one audio
channel and the associated direction parameters are,
furthermore, processed as already described for the audio
decoder of Fig. 6, to derive a perceptually enhanced output
signal 170.

Although the invention has been described mainly in the
field of multi-channel audio reproduction, different fields
of application can profit from the inventive methods and
apparatuses. As an example, the inventive concept may be
used to focus (by boosting or attenuating) on specific
individuals speaking in a teleconferencing scenario. It can
be furthermore used to reject (or amplify) ambient
components as well as for de-reverberation or reverberation
enhancement. Further possible application scenarios
comprise noise canceling of ambient noise signals. A
further possible use could be the directional enhancement
for signals of hearing aids.
Depending on certain implementation requirements of the
inventive methods, the inventive methods can be implemented
in hardware or in software. The implementation can be
performed using a digital storage medium, in particular a
disk, DVD or a CD having electronically readable control
signals stored thereon, which cooperate with a programmable
computer system such that the inventive methods are
performed. Generally, the present invention is, therefore,
a computer program product with a program code stored on a
machine readable carrier, the program code being operative
for performing the inventive methods when the computer
program product runs on a computer. In other words, the
inventive methods are, therefore, a computer program having
a program code for performing at least one of the inventive
methods when the computer program runs on a computer.
While the foregoing has been particularly shown and
described with reference to particular embodiments thereof,
it will be understood by those skilled in the art that
various other changes in the form and details may be made
without departing from the spirit and scope thereof. It is
to be understood that various changes may be made in
adapting to different embodiments without departing from
the broader concepts disclosed herein and comprehended by
the claims that follow.

We-Claim
1. Method for reconstructing an audio signal having at
least one audio channel and associated direction
parameters indicating a direction of origin of a
portion of the audio channel with respect to a
recording position, the method comprising:
selecting a set direction of origin with respect to
the recording position; and
modifying the portion of the audio channel for
deriving a reconstructed portion of the reconstructed
audio signal, wherein the modification comprises
increasing an intensity of the portion of the audio
channel having direction parameters indicating a
direction of origin close to the set direction of
origin with respect to another portion of the audio
channel having direction parameters indicating a
direction of origin further away from the set
direction of origin.
2. The Method of claim 1, wherein selecting comprises:
reading the set direction from a memory.
3. The Method of claim 1, in which the modification
comprises modifying a frequency domain representation
of the portion of the audio channel.
4. The Method of claim 1, in which the modification
comprises modifying a time domain representation of
the portion of the audio channel.
5. The Method of claim 1, in which the modification
comprises deriving a scaling factor for each portion
of the audio channel such that a scaled portion of the

audio channel, the scaled portion derived by
multiplying the portion of the audio channel with the
scaling factor, having associated direction parameters
indicating a direction of origin close to the desired
direction of origin has an increased intensity with
respect to another scaled portion of the audio channel
having associated direction parameters indicating a
direction of origin further away from the desired
direction of origin
6. The Method of claim 1, further comprising:
deriving a frequency representation of the at least
one audio channel.
7. The Method of claim 6, in which the deriving comprises
deriving a representation of a first and a second
finite width frequency interval of the at least one
audio channel, wherein the width of the first
frequency interval is different than the width of the
second frequency interval.
8. The Method of claim 1, wherein selecting of the
desired direction of origin comprises receiving input
parameters indicating the desired direction as a user
input.
9. The Method of claim 1, wherein selecting the desired
direction comprises receiving direction parameters
associated to the audio signal, the direction
parameters indicating the desired direction.
10. The Method of claim 1, wherein selecting the desired
direction comprises determining the direction of
origin of a finite width frequency interval of the at
least one audio channel.
11. The Method of claim 1, further comprising:

receiving a diffuseness parameter associated to the
audio channel, the diffuseness parameter indicating a
diffuseness of the portion of the audio channel; and
wherein the modifying of the portion of the audio
channel comprises decreasing an intensity of the
portion of the audio channel having a diffuseness
parameter indicating a high diffuseness with respect
to another portion of the audio channel having a
diffuseness parameter indicating a lower diffuseness.
12. The Method of claim 1, further comprising:
up-mixing the at least one audio channel to multiple
channels for playback via a loudspeaker system having
multiple loudspeakers, wherein each of the multiple
channels has a channel portion corresponding to the
portion of the at least one audio channel.
13. The Method of claim 12, in which the modification
comprises increasing the intensity of each of the
channel portions up-mixed from the portion of the
audio channel having direction parameters indicating a
direction of origin close to the desired direction of
origin with respect to other channel portions of the
multiple channels up-mixed from another portion of the
audio channel having direction parameters indicating a
direction of origin further away from the desired
direction of origin.
14. The Method of claim 13 or 14, further comprising:
panning the amplitude of the channel portions such
that a perceived direction of origin of reconstructed
channel portions corresponds to the direction of
origin when played back using a predetermined
loudspeaker set-up.

15. The Method for enhancing a directional perception of
an audio signal, the method comprising:
deriving at least one audio channel and associated
direction parameters indicating a direction of origin
of a portion of the audio channel with respect to a
recording position;
selecting a set direction of origin with respect to
the recording position; and
modifying a portion of the audio channel for deriving
a portion of an enhanced audio signal, wherein the
modification comprises increasing an intensity of a
portion of the audio channel having direction
parameters indicating a direction of origin close to
the set direction of origin with respect to another
portion of the audio channel having direction
parameters indicating a direction of origin further
away from the set direction of origin.
16. Audio decoder for reconstructing an audio signal
having at least one audio channel and associated
direction parameters indicating a direction of origin
of a portion of the audio channel with respect to a
recording position, comprising:
a direction selector adapted to select a set direction
of origin with respect to the recording position; and
an audio portion modifier for modifying the portion of
the audio channel for deriving a reconstructed portion
of the reconstructed audio signal, wherein the
modification comprises increasing an intensity of the
portion of the audio channel having direction
parameters indicating a direction of origin close to
the set direction of origin with respect to another

portion of the audio channel having direction
parameters indicating a direction of origin further
away from the set direction of origin.
17. Audio encoder for enhancing a directional perception
of an audio signal, the audio encoder comprising:
a signal generator for deriving at least one audio
channel and associated direction parameters indicating
a direction of origin of a portion of the audio
channel with respect to a recording position;
a direction selector adapted to select a set direction
of origin with respect to the recording position; and
a signal modifier for modifying the portion of the
audio channel for deriving a portion of an enhanced
audio signal, wherein the modification comprises
increasing an intensity of a portion of the audio
channel having direction parameters indicating a
direction of origin close to a set direction of origin
with respect to another portion of the audio channel
having direction parameters indicating a direction of
origin further away from the set direction of origin.
18. System for enhancement of a reconstructed audio
signal, the system comprising:
an audio encoder for deriving an audio signal having
at least one audio channel and associated direction
parameters indicating a direction of origin of a
portion of the audio channel with respect to a
recording position;
a direction selector adapted to select a set direction
of origin with respect to the recording position; and

an audio decoder having an audio portion modifier for
modifying the portion of the audio channel for
deriving a reconstructed portion of the reconstructed
audio signal, wherein the modifying comprises
increasing an intensity of the portion of the audio
channel having direction parameters indicating a
direction of origin close to a set direction of origin
with respect to another portion of the audio channel
having direction parameters indicating a direction of
origin further away from the set direction of origin.
>. A computer program for, when running on a computer,
implementing the method of claim 1.

An audio signal having at least one audio channel and
associated direction parameters indicating a direction
of origin of a portion of the audio channel with
respect to a recording position is reconstructed to
derive a reconstructed audio signal. A desired
direction of origin with respect to the recording
position is selected. The portion of the audio channel
is modified for deriving a reconstructed portion of
the reconstructed audio signal, wherein the modifying
comprises increasing an intensity of the portion of
the audio channel having direction parameters
indicating a direction of origin close to the desired
direction of origin with respect to another portion of
the audio channel having direction parameters
indicating a direction of origin further away from the
desired direction of origin.