DK178440B1

DK178440B1 - Configuring a plurality of sound zones in a closed compartment

Info

Publication number: DK178440B1
Application number: DKPA201400470A
Authority: DK
Inventors: Søren Borup Jensen; Martin Olsen; Martin Bo Møller
Original assignee: Bang & Olufsen As
Priority date: 2014-07-14
Filing date: 2014-08-22
Publication date: 2016-02-29
Also published as: DK201400470A1

Abstract

A method and a system to enable one or more active sound zones in a closed environment and in which individual sound sources are provided to individual persons located within each sound zone. The sound zones are created as 3D domains defined according to the aspects of "acoustically bright zone" and "acoustically dark zone", and with predefined threshold values related to a required "acoustical contrast" among the plurality of sound zones. Characteristics of the closed environment, this being e.g. a room or a car cabin, are analyzed and predefined via in situ measurement of the actual room/car cabin. A constraint based system controls the configuring of the individual sound zones according to relevant constraint parameters e.g., but not limited to: loudspeaker transducer configuration, number of sound channels, number and location of persons in the domain, room characteristics, (size, isolation, reflection, temperature), type of source material (speech, music genre, sound volume per zone).

Description

Configuring a plurality of sound zones in a closed compartment

The present invention is within the domain of multichannel sound systems and specifically in enabling and controlling individual sound zones to provide user selected sound channels to be rendered in a selected zone.

The creation of personal sound zones relies on simultaneous generation of mutual dependent sets of 3D spatial confined regions of high sound pressure (a bright zone) and low sound pressure relative to the bright zone (a dark zone). Two examples of zone configurations are given in Fig. Oa but not restricted to these scenarios.

Adequate separation of the bright and dark zone(s) in terms of acoustic contrast should be defined e.g. according to perceptual threshold values, as disclosed by the applicant in [US20130230175].

The general setup procedure includes thorough pre-analysis of the specific 3D listening space by means of microphone array measurements and head and torso simulator (HATS) binaural impulse room response measurements according to [US 8,175,286 B2], In addition, prediction of the sound field using sound field extrapolation algorithms and advanced simulations is also applied.

The system configuration procedure implicates particular processing steps in relation to the pre-analysis as illustrated in Fig. Ob.

Prior art in W02014/036121, discloses a method where creation of sound filed zones are based on a model derived from head and torso simulator (HATS) binaural impulse room response measurements.

Prior art US2013/0230175 discloses a method to establish a threshold of acceptability for an interfering audio programme on a target audio program. This method is partly included in the current invention as an object in the acoustical properties to be fulfilled.

A primary function of the current invention is that that every signal source is rendered to every sound transducer to generate a sound pressure (SP) which is recorded by every microphone in all positions in the microphone array, these data to be applied in the configuring process during run time.

A first aspect of the invention is a method and a system for creating and controlling personal sound zones in a closed space which is enabled with a plurality of loudspeaker transducers where the method is characterized by: o The creation of personal sound zones relies on simultaneous generation of mutual dependent sets of 3D spatial confined regions of high sound pressure (a bright zone) and low sound pressure relative to the bright zone (a dark zone); o Adequate separation of the bright and dark zone(s) in terms of acoustic contrast is defined according to perceptual threshold values; o A data model of the acoustical characteristic in the closed space is generated: a thorough pre-analysis of the specific 3D listening space by means of microphone array measurements from one or more signal sources and head and torso simulator (HATS) binaural impulse room response measurements; prediction of the sound field using sound field extrapolation algorithms and advanced simulations is applied; o The data model includes a set of parameters and actual values that depicts the acoustical properties of the target space e.g. being a car cabin.

In a second aspect of the invention the data model include results from in situ measurement of the acoustic transfer function between every loudspeaker transducer to a 3D grid of points defining a volume larger than a human head in a cubic grid of side length 5 cm.

Summary of Figures

Figure 0 displays the zone concept and pre-analysis steps Figure 1 displays reproduction of bass frequencies Figure 2 and 3 displays reproduction of midrange frequencies Figure 4 displays reproduction of high frequencies

Figure 5 displays the System Overview of one embodiment of the invention Figure 6 displays the Constraint Solver of one embodiment of the invention Figure 7 displays a sound zone configuration example Figure 8 displays a 3D microphone arrangement

Description

In a preferred embodiment reproduction methods are adjusted according to transducer frequency domain.

Reproduction method - bass frequencies FREQUENCY RANGE: approx. 20-300 Hz SOURCE LAYOUT: 4-8 loudspeakers (woofers) distributed around target zones and/or with single control loudspeakers in or nearby one or more of the target zones. All loudspeakers receive signals that are radiated in all target zones.

TARGET: the creation of one zone of high sound pressure level (bright zone) simultaneously reproduced with another zone of low sound pressure level (dark zone) relative to the bright zone; a zone being a 3D spatially confined region, that is larger than a human head, within an enclosed space in which monaural sound is reproduced.

CONTROL METHOD: Acoustic Contrast Control represents the energy cancellation approach, and the ratio of the spatially averaged sound pressure levels between the bright zone and the dark zone is maximized. The phase of the resulting sound field is not taken into account. The implementation of the control method is augmented with a number of performance related measures to improve the "quality" of the target for filter calculation. Alternative hybrid methods are described in [PCT/EP2014/050081] and [WO 2013/135819], SETUP: In situ measurement of the acoustic transfer function between every loudspeaker to a 3D grid of points defining a volume larger than a human head in a cubic grid of side length 5 cm. Advanced pre- and post-processing schemes are introduced in the calculation of the filter for each loudspeaker.

Reproduction method - midrange frequencies

FREQUENCY RANGE: approx. 200-7000 Hz SCENARIO #1: DIRECTIONAL RADIATION

SOURCE LAYOUT: 2x(7+2) loudspeakers mounted on front seat orientated towards target zones at the rear seats or mounted in ceiling angled towards the head of the listener. Individual array and signal flow in front of each target zone.

TARGET: Focusing of the sound field generated by each array towards the respective target zones. Each array distributes high energy towards the centre of the zone and attenuates the sound field towards the centre of the car compartment and side windows. Each zone is controlled by a specific array individually.

CONTROL METHOD: Acoustic Contrast Control, and/or least mean square optimization of the directivity and/or spatial sound field. Optimization made either based on anechoic measurements in the cube (directivity control) or utilizing in situ measurements (sound field control).

SETUP: Directional radiation measurements of arrays measured on e.g. a half circle with radius equal to target distance and/or in situ measurement of the acoustic transfer function between every loudspeaker to a 3D grid of points defining a volume larger than a human head in a cubic grid of side length 5 cm or 3D grid of points sampling volume around each ear of the target listener.

VARIATIONS: when each array produces two different lobes of high sound pressure aiming towards left and right ear respectively, binaural sound reproduction methods can be introduced, as illustrated in fig. 2.b.

Reproduction method - midrange frequencies

FREQUENCY RANGE: approx. 200-7000 Hz SCENARIO #2: SOUND FIELD CONTROL

SOURCE LAYOUT: 2x(7+2) loudspeakers mounted on front seat orientated towards target zones at the rear seats or mounted in ceiling angled towards the head of the listener and a spatial 3D distribution. All loudspeakers receive signals to all target zones.

TARGET: Optimization of the sound field generated by the total amount of loudspeakers towards the respective target zones.

CONTROL METHOD: Acoustic Contrast Control, and/or least mean square optimization of the sound field. The implementation of control method is augmented with a number of performance related measures to improve the "quality" of the target for filter calculation.

SETUP: In situ measurement of the acoustic transfer function between every loudspeaker to a 3D grid of points defining a volume larger than a human head in a cubic grid of side length 5 cm or 3D grid of points sampling volume around each ear of the target listener. Advanced pre- and post-processing schemes are introduced in the calculation of the filter for each loudspeaker.

Reproduction method - high frequencies FREQUENCY RANGE: approx. 7000-20000 Hz SOURCE LAYOUT: a single lens aiming towards the centre of the target zone or two lenses aiming towards each ear of the listener in each target zone for binaural reproduction. The width of the main lobe is optimized for each purpose individually.

TARGET: Passive control of the directivity of a single loudspeaker (tweeter) in order to focus energy towards the target zone and attenuate leakage sound energy towards the other zones and side windows.

CONTROL METHOD: 3D geometry optimization of acoustic lenses and influence of baffling.

SETUP: FEM simulation optimization and verification in situ.

In further aspects of the invention the data model includes: • The Acoustic Contrast Control that represents the energy cancellation approach and the ratio of the spatially averaged sound pressure levels between the bright zone and the dark zone is maximized.

• Definitions of one or more loudspeaker arrays configured to have the directional radiation towards individual target zones.

• Definitions to obtain binaural sound reproduction caused by two different lobes of high sound pressure aiming towards left and right ear respectively.

• Definitions of one or more loudspeaker arrays configured to have the sound field control towards individual target zones.

In yet other aspects of the invention: • The sound zone filters are represented in algorithms as a finite impulse response filter.

• The sound zone filters are procedures referred to by the data model as external constraints.

• A defined perceptual model (ref PA 2014 00083) is referred to by the data model.

Parameters that interact via relations, and are partly or fully included in the data model: • Acoustic contrast • Reproduction error • Planarity • Seat/zone configuration • Loudspeaker layout (location and type) • Privacy (VIP teleconference) - masking of sound events • In-car communication • Occupancy • Audio programme • Sound imaging / staging (Binaural reproduction / Stereo / surround sound) • Velocity / RPM of vehicle engine • External (road) noise • Internal noise • Reproduction level • Equalization (e.g. "mood") • Ambient conditions (e.g. temperature and static pressure) • "room gain" - adaptation to environment • Loudspeaker driver condition (e.g. voice coil temp., linear range)

In the preferred embodiment the data model is implemented as a constraint model in terms of a table.

This constraint table being accessed by a constraint solver to deduce legal combinations that constitutes one or more configurations of system components being at least: media sources, sound transducers, amplifiers, equalizers.

As room temperature may have influence on the acoustical properties in a room, different temperature intervals may be included as parameters in the constraint data model. Thus, the initial 3D analysis of the room is executed in different temperature conditions, with the result to be applied at run time to address legal adjustment parameters accordingly.

In the preferred embodiment a constraint solver handles parameters/variables of different types: • Boolean with a values set like On/off, True/false, Logical "1 or 0" • Integer, e.g. "10" • Integer intervals, e.g. "0 to 100" • Real, e.g. "12.88" • Real intervals, e.g. "17.5 to 21.5" • Symbolic, e.g. "Left", "Right" "Centre"

The attributes per variable are defined and interpreted as enumerated options, like:

One sound channel is defined: "Channel: Attrib(Left, Right, Centre)".

Numeric are defined and interpreted individually or in intervals, like:

Amplifier levels is defined: "AmpLevel: Integer (0, 1, 2, 3, 4, 5)", or "AmpLevel: Integer (0 -> 5)".

A defined variable may be the candidate for a resource calculation and/or an optimization procedure according to a specified performance requirement.

This enables for an implicit cost function to be enabled when a certain combinations of variables included in the solution space for the legal combination in which the addressed parameter is part.

A formula may interact with the constraint table variables in alternative modes of operation:

Examples of a set of generic constraint tables, in which a formula Zx is addressed, are:

Constraint table 1: PI and P2 and P3 and P4 and Z1 Or (PI and P5 and P6)

Or (P7 and P8)

Constraint table 2: PI and P2 and P3 Or (PI and P5 and P6)

Or (P7 and P8) and Z2 A formula example is: Zl= PI + P2 + n*P3 In which the variables P1,P2,P3 • are selectable as input/output parameters in the constraint table; • are included as variables in the calculation, and with the result Zl; • and Zl is input/output parameter in the constraint table.

Thus: P1,P2,P3,P4 may be selected unconditionally if the Zl variable is unselected/uncalculated.

P1,P2,P3 variable values will be constrained if Zl has been selected/calculated.

Alternative values for one or all of the P1,P2,P3 might be forced which may imply that Zl to be recalculated as the constraint table to be reevaluated.

Another formula example is: Z2= PI + P4

In which the variables P1,P4 are from different constraint tables.

Thus:

The processing is in principle as disclosed above, just including the two constraint tables to be reevaluated in parallel.

Figure 5 displays a System Overview, in which there are two functional entities and related subsystems: a) Analyze & Simulate and b) Run Time environment.

Analyze and Simulate is the subsystem applied to analyze the space (room) of the target, e.g. a car cabin: • The system is operated based on functional requirements, specifications and parameter pre-sets, defined for the current application, being e.g. a car audio/video rendering system including individual sound zones enabled as 3D-sound spaces and rendered via a multi-channel audio system.

• The acoustical characteristics and the behavior of the specific target space is analyzed via an advanced robotic system enabled with a plurality of microphones and sound signals, that automatically scans the space, and identify one set of parameters for this specific target.

The parameters are according to the requirements defined for e.g., but not limited to: the number of passengers and location in the car, positions of the heads of the passengers, the number and location of sound channels (transducer units), the number and placement of sound zones, type of media (music/speech) and alike.

• The result of the Analyze and Simulate function is a complete parameterized data model including combination of parameters that define the "state space" for one actual target (room) enabled with: a physical configuration, a number of sound channels (transducers), one or more listeners, one or more sound zones, one or more of 3D sound spaces, sound zone characteristics (gain, threshold limits).

• The parameterized data model is converted (compiled and compressed) and mapped into a constraint table that relate any relevant parameter with any other relevant parameter. The structure of the table being an n-dimensional state space with direct real time access to any of the defined parameter combinations.

Run Time is the subsystem applied to control the acoustical system functionality in the room of the target, e.g. a car cabin: • Based on the sensor readings and other inputs (user given of system generated) the acoustical parameters are controlled via stimuli to the system hardware. This, as example, to select an audio source, and request a specific sound zone configuration, and provide it to the user at a certain volume level.

• The audio system is reconfigured dynamically according to the defined constraints/relations via the Deduce function, that communication with the relation via the state vector, which acts as an input/output list of parameters and their attributes.

• Certain problem domains may include an optimize function, in which a linear constraint or formula is computed.

• Media information as channels or streams with audio, video or tele is accessed via the local devices or virtual sources residing on the Internet or in the Cloud.

Figure 6 displays the concept of the constraint solver, and how this interrogates with the application.

The State Vector (20) is the data interface that connects the system parameters to the application program (22).

The system parameters include input values from physical sensors, user given commands and control commands from utility software.

The system parameters include output values to be applied for setting actual variables as deduced (23) from the defined constraints (21). The constraints define a solution space including legal combinations according to the defined relations; thus the solution space is consistent and complete and without any contradictions. In a preferred embodiment the sequence of processing is: • The state vector is initialized with preset values.

• A change of the state in the State Vector [SV](20), e.g. a user triggered command/event generates new inputs into the SV.

• The application (22) accesses the SV via the Deduce function (23) which interrogates with the Constraints to determine the consequence of the given input as validated against the defined parameter relations.

• The consequence of this deduction might be new parameter values and updates of the SV accordingly.

• As required in actual application additional constraints, represented in linear constraints, procedures or formulas might be supported; this supported via the Optimize function (24).

Figure 7 displays an example of a multi-channel sound system including five loudspeaker transducers, each having individual controllable amplifier-, equalizer-, and delay means. The system is configured according to the invention to be enabled with two sound zones for the pleasure of two users.

The configuring procedure transforms the physical domain to a virtual sound space enabled access and control by a user.

The basic primitives for the configuring are, but limited to: • The number of the sound zones to be configured; • The physical location Χ,Υ,Ζ relative in a room/space of the zones; • The threshold, i.e. acceptable interference, among the one or more zones; • The type of sound source, e.g. speech or music, which is selected to be rendered in a selected sound zone; • The number of sound transducers and type (tweeter, midrange, woofer); • The physical location Χ,Υ,Ζ relative in a room/space of each of the transducers; • The characteristics for each amplifier-, equalizer-, and delay means.

• The characteristics of the Χ,Υ,Ζ room/space, like: size, acoustical impedance, room gain, reflections, isolation, reverberations and alike in given positions.

Figure 8 displays the concept of a 3D microphone arrangement (30).

A plurality of microphones are installed in mechanical support means that constitute a 3-dimensional grid (30) within which one or more microphones can be positioned relative to a given X,Y and Z position in a 3D coordinate system.

A number of microphones, e.g. five are considered as a set, which is fixated at the same mean.

In a preferred embodiment the movement of a set of microphones (36,37,38,39) are enabled with slider means to move along the X axis (35); e.g. one set of microphones is slide from one position (36) to another position (37), or moved from one level (38) to another level (39).

The grid arrangement is arranged with a number of levels e.g. three (31,32,33).

The movement/slide of the set of microphones may be manually operated or partly or fully motorized.

The physical dimensions of the 3D microphone arrangement, are determined by the closed room dimension to analyzed and the required to precision of the measured acoustical parameters.

An application example having a car cabin, which must be in defined in 3D cube grid of side/length 5cm, the dimension are: number of microphones in a set is 20, length along the X-axis, approx. 2.5m, width along the Z-axis approx. 1.5m, and height along the Y axis being in three layers 10cm.

The invention is very applicable in advanced multi channels sound systems in which individual sound zone shall be enabled; this being closed spaces/room in cars, houses, boats, airplane and alike.

The control systems include means for adaptive configuring and reconfiguring of sound zones and where the control is based on the information residing in the data model include relations among parameters, and where the data model is accessed by a constraint solver that deduces legal solutions according to control inputs given by e.g.: the user, sensor readings, system software generated events, application generated events and alike.

Claims

1. A method for generating and controlling personal sound zones in a confined space, the method is provided with a plurality of speaker transducers, and the method is characterized by: a. Forming personal sound zones based on simultaneous generation of interdependent sets of 3D spatially delineated regions with a high-sound pressure region and a low-sound pressure region with a low sound pressure relative to the high-sound pressure region; b. separating the high-noise region and low-noise region in terms of acoustic contrast, defined in accordance with perceptual threshold values; c. generates a data model of acoustic characteristic in the confined space by: i. pre-analyzing the specific 3D listening compartment complete using microphone array measurements and head and torso simulator (HATS) binaural impulse response measurements in space; ii. predict the sound field using sound field extrapolation algorithms and apply simulations; in which d. The data model includes a set of parameters and current values that depict the acoustic properties of a target area, such as a car booth.

A method according to claim 1, wherein the data model includes results from in situ measurements of the acoustic transfer function between each speaker transducer to a 3D grid of points defining a volume larger than a human head, and in a cube-shaped grid having a length of 5 cm.

A method according to claim 2, wherein a plurality of low frequency transducers, having a frequency range 20 - 300Hz, are distributed around target zones and / or with single control speakers in or close to one or more target zones.

A method according to claim 2, wherein a plurality of intermediate frequency transducers, with frequency range 200 -7000Hz, are mounted on front seat (s) oriented towards target zones at rear seat (s) or mounted in the ceiling angled to the head of the listener (s). .

A method according to claim 2, wherein one or more high frequency transducers, with a frequency range of 7000 -20000Hz, is directed to the center of one or more target zones.

A method according to claim 3, 4 or 5, wherein the data model includes Acoustic Contrast Control representing an energy suppression method and wherein the ratio of the spatial average sound pressure between the high sound pressure region and the low sound pressure region is maximized.

A method according to claim 6, wherein the data model includes definitions of one or more speaker arrays configured to have directional radiation toward individual target zones.

A method according to claim 6, wherein the data model includes definitions for obtaining binaural sound reproduction caused by two different high-pressure loops targeted respectively to the left and right ear.

A method according to claim 7 or 8, wherein the data model includes definitions of one or more speaker arrays configured to have sound field control against individual target zones.

A method according to claim 9, wherein sound zone filters are represented in algorithms such as FIR, i.e. finite impulse response filter.

A method according to claim 10, wherein sound zone filters are procedures referred to by the data model as external coercive bonds.

A method according to claim 10, wherein a defined perceptual mode is referenced via the data model.