Computer-implemented method for generating a head-related transfer function

Abstract

A computer-implemented method for generating a head-related transfer function, HRTF, for bringing about a three-dimensional, 3D, sound experience, the method comprising the following steps of: − simulating in a first simulation step a beam behaviour of one or more sound beams coming from one or more virtual loudspeaker sources depending on a head of a listener in a space; − calculating on the basis of the first simulation step one or more reflections of the one or more sound beams; − emulating in an emulation step one or more sound parameters representative of at least one of a time difference, an attenuation of the one or more sound beams, a delay of the one or more sound beams, a diffraction of the one or more sound beams and spectral information of the one or more sound beams on the basis of the calculated one or more reflections; and generating the HRTF on the basis of the one or more sound parameters emulated in the emulation step.

Description

[0001] Computer-implemented method for generating a head-related transfer function

[0002] Field of the invention

[0003] The present invention relates to the field of audio technology, and more particularly to techniques for generating a head-related transfer function, HRTF, for bringing about a three-dimensional, 3D, sound experience using a computer-implemented method.

[0004] Background

[0005] There are generated and adjustable Head-Related Transfer Functions, HRTFs, which are realistic and neutral in their reproduction of a 3D audio, binaural, experience, designed specifically as a tool for music producers and audio engineers.

[0006] The known techniques comprise of simulating HRTFs using a 3D model. The 3D model is used to solve wave equations. This is however a highly resource-intensive and impractical solution.

[0007] Other methods use a measured, generalized HRTF, such as the Neumann KU-100, and use the measured listener properties as a “deviation” from this model.

[0008] The existing HRTFs however have several drawbacks. They require expensive and uncomfortable measurements, cannot be adjusted, are arithmetically expensive and introduce undesirable sound coloration.

[0009] “ Ray acoustics using computer graphics technology” (Niklas Rober et al., 2007) relates to the modeling of room acoustics and sound wave propagation using raytracing techniques. “Room Impulse Response Simulation Based on Equal-area Ray Tracing” (Gu et al., 2014) describes a ray tracing method based on equal-area distribution for room impulse response simulation. “An adaptable ellifsoidal head model for the interauraltime difference” (Duda et al., 1999) discloses the use of an ellipsoidal head model, particularly for calculating interaural time differences. “ HRTF simulations through acoustic raytracing “ (Rober et al., 2006) describes earlier work of Rober on HRTF simulations through acoustic raytracing. US 10 003 906 B2 relates to determining and using room-optimized transfer functions. US 2024 / 177016 Al describes methods and systems for training and providing a machine learning model for audio compensation.

[0010] Summary of the invention

[0011] Embodiments of the invention have the object of providing a solution to the drawbacks of the existing HRTF techniques. According to a first aspect, a computer-implemented method is provided for generating a head- related transfer function, HRTF, for bringing about a three-dimensional, 3D, sound experience. An HRTF can be understood to mean a function which describes the transfer of sound from a source to the eardrums of a listener on the basis of the position of the source and the listener. The method comprises the following steps of:

[0012] - simulating in a first simulation step a beam behaviour of one or more sound beams coming from one or more virtual loudspeaker sources depending on a head of a listener in a space;

[0013] - calculating on the basis of the first simulation step one or more reflections of the one or more sound beams of the one or more loudspeaker sources;

[0014] - emulating in an emulation step one or more sound parameters representative of at least one of a time difference, an attenuation of the one or more sound beams, a delay of the one or more sound beams, a diffraction of the one or more sound beams and spectral information of the one or more sound beams on the basis of the calculated one or more reflections; and

[0015] - generating the HRTF on the basis of the one or more sound parameters emulated in the emulation step.

[0016] The computer-implemented method provides an efficient and cost-effective manner of generating an HRTF, particularly because the method requires no expensive or uncomfortable measurements and introduces no undesirable sound coloration. An advantage of the method is that simulating the beam behaviour of the sound beams in the first simulation step results in a realistic and more neutral reproduction of the 3D audio experience. Simulating beam behaviour means that the direction and intensity of the sound beams coming from virtual loudspeaker sources are simulated on the basis of the position of the source and the listener. This is done by making use of geometrical acoustics, i.e. modelling the sound waves as straight lines moving through the space and reflecting from surfaces. Simulating the beam behaviour is considerably less arithmetically intensive and requires a considerably smaller amount of computing power compared to solving for instance wave equations, enabling the method to work in real time. The emulation step moreover provides for a representative simulation of the sound parameters, this resulting in a more accurate HRTF. This results in an improved sound experience for the listener, wherein the HRTF is crucial for a realistic and accurate reproduction of sound. It is noted that simulating refers to imitating a determined phenomenon using a model or computer program. The object of simulation is to provide a realistic reproduction of reality and to understand how the phenomenon works.

[0017] Emulating refers to imitating a determined system using the simulation. The object of emulation is to replicate the functionality of the imitated system or device on the emulating system or device. According to a second aspect, computer-implemented method for generating a head- related transfer function, HRTF, for bringing about a three-dimensional, 3D, sound experience in real time is provided, the method comprising the following steps of simulating in a first simulation step a beam behaviour of one or more sound beams coming from one or more virtual loudspeaker sources depending on a position of a head of a listener in a space, wherein the simulating comprises simulating the head as an ellipsoid with a first virtual microphone and a second virtual microphone, wherein the first and the second virtual microphones are positioned opposite each other on the head and comprises determining a first plane defined by points where sound beams originating from the one or more virtual loudspeaker sources are intersecting the surface of the ellipsoidal head, determining a second plane defined by the one or more virtual loudspeaker sources, a center of the ellipsoidal head, and the first and second virtual microphones respectively; identifying an intersection line formed by the first plane and the second plane; determining two distinct intersection points of said intersection line with the surface of the ellipsoidal head; and utilizing one or more paths through both of said two distinct intersection points for determining said beam behavior of the one or more sound beams. The method further comprising calculating on the basis of the first simulation step one or more reflections of the one or more sound beams of the one or more loudspeaker sources, said reflections originating from one or more surfaces in the space; emulating, in an emulation step, one or more sound parameters representative of at least a time difference, an attenuation and, a delay of the one or more sound beams, optionally further representative of a diffraction of the one or more sound beams or spectral information of the one or more sound beams, on the basis of the calculated one or more reflections; and generating the HRTF on the basis of the one or more sound parameters emulated in the emulation step.

[0018] The methods preferably comprises of raytracing the one or more sound beams for each of the one or more virtual loudspeaker sources during simulation of the beam behaviour. Raytracing is a technique for simulating the interaction of sound with objects in a space. Applying raytracing during simulation of sound beams provides a number of technical advantages. In this way a more realistic image of the reflections of the sound in the space can be obtained, this resulting in a more accurate HRTF. The use of raytracing can moreover make calculation of the sound beams more efficient, this resulting in a faster and more cost-effective method. A further improvement of the reproduction of the 3D audio experience can therefore be achieved by applying raytracing.

[0019] The methods preferably comprises of calculating the one or more reflections using the raytraced one or more sound beams. This means that the reflections of the sound waves are calculated on the basis of the simulated sound beams. The use of raytraced sound beams for calculating reflections provides a number of technical advantages. In this way the reflections of the sound in the space can for instance be calculated more accurately, this resulting in a more realistic and accurate HRTF. The use of raytracing can moreover make calculation of the reflections more efficient, this resulting in a faster and more cost-effective method. A further improvement of the reproduction of the 3D audio experience can therefore be achieved by using raytraced sound beams to calculate reflections.

[0020] The methods preferably comprises of raytracing according to the following steps of:

[0021] - Simulating each loudspeaker source of the one or more loudspeaker sources as a sphere with a predetermined diameter and a predetermined resolution, wherein the predetermined resolution is at least one parameter representative of the number of sound beams being emitted from the sphere.

[0022] - Distributing a number of points uniformly over the sphere, wherein the number of points corresponds with the number of sound beams.

[0023] - Emitting a respective sound beam from each point on the sphere.

[0024] Simulating each loudspeaker source as a sphere with a predetermined diameter and a predetermined resolution and distributing points uniformly over the sphere enables an efficient and accurate simulation of the sound waves to be achieved. Emitting a respective sound beam from each point on the sphere results in the sound waves being reproduced realistically, this resulting in a more accurate HRTF. Provision can therefore be made for the use of these steps in raytracing of the sound beams to enable an improved and efficient reproduction of the 3D audio experience.

[0025] The methods preferably comprises of simulating the head as an ellipsoid with a first virtual microphone and a second virtual microphone, wherein the first and the second virtual microphones are positioned opposite each other on the head. Simulating the head as an ellipsoid with virtual microphones opposite each other results in the sound recordings being reproduced more realistically, this resulting in a more accurate HRTF. Simulating the head as an ellipsoid instead of a sphere indeed provides for a more realistic simulation of the actual shape of the human head. This is further understood to mean that when the first and the second virtual microphones are positioned opposite each other on the head, the position of the virtual microphones corresponds with the position of the ears on a human head. This has the result that the sound recordings and the HRTF generated therefrom better match the actual sound experience of a human. Simulating the head as an ellipsoid, in combination with the above stated steps, enables a further improvement of the accuracy and efficiency of the method for generating an HRTF. The methods preferably comprises of positioning the first and the second virtual microphones at an acute angle relative to one of three perpendicular axes of symmetry of the ellipsoid head. Positioning the virtual microphones at an acute angle relative to one of the perpendicular axes of symmetry results in a more realistic reproduction of the human ear canal and the directional dependence of human hearing. This results in a more accurate and more realistic HRTF. The use of this step in simulating the beam behaviour of the sound beams, in combination with the above stated steps, enables a further improvement of the accuracy and efficiency of the method for generating an HRTF. The method preferably comprises of positioning the first and the second virtual microphones at an acute angle of at least 60° relative to one of the three perpendicular axes of symmetry of the ellipsoid head. This acute angle is preferably at least 65°. Positioning the virtual microphones at an acute angle of at least 60°, preferably at least 65°, results in a better approximation of the actual geometry of the human ear canal. The acute angle is preferably measured in a horizontal plane relative to one of the three perpendicular axes of symmetry which is directed forward, in a viewing direction of the ellipsoid head. This results in an even more accurate and more realistic HRTF.

[0026] The methods preferably comprises of calculating the number of reflections of a respective sound beam. This means that, for each sound beam, the number of reflections is calculated so as to simulate a realistic reproduction of the sound waves in the space. Calculating the number of reflections contributes to a more realistic and more accurate reproduction of the 3D audio experience.

[0027] The methods preferably comprises of calculating the angle of incidence of a respective sound beam relative to the ellipsoid head. This means that, for each sound beam, the angle of incidence relative to the head is calculated so as to simulate a realistic reproduction of the sound waves in the space. Calculating the angle of incidence contributes to an even more realistic reproduction of the 3D audio experience, and thereby an even better HRTF.

[0028] The methods preferably comprises of calculating a distance travelled by a respective sound beam between a virtual loudspeaker source of the one or more loudspeaker sources and the head. A “sound beam” can be understood to mean a bundle of one or more sound waves which are sent from a virtual loudspeaker source to the head of a listener. “Distance” refers to the length of the path travelled by the sound beam between the virtual loudspeaker source and the head of the listener. Calculating the distance travelled by a respective sound beam results in a more accurate calculation of the sound parameters which are used in generating the HRTF. This results in an improved sound experience for the listener, wherein the HRTF realizes a realistic and accurate reproduction of sound. In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, calculating the distance contributes to an even more realistic reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF.

[0029] Calculating the distance travelled by a respective sound beam preferably comprises of calculating both a first distance between the virtual loudspeaker source and the first virtual microphone and a second distance between the virtual loudspeaker source and the second virtual microphone. A “first distance” refers to the length of the path travelled by a sound beam between the virtual loudspeaker source and the first virtual microphone, while a “second distance” refers to the length of the path travelled by a sound beam between the virtual loudspeaker source and the second virtual microphone. Calculating both the first and the second distance contributes to an even more accurate and more realistic simulation of the sound waves, and therefore an even better HRTF. The use of virtual microphones opposite each other on the head results in a more realistic reproduction of the recorded sounds and calculating the distance between the virtual loudspeaker source and the virtual microphones contributes to a more accurate calculation of the sound parameters which are used in generating the HRTF. This results in an improved sound experience for the listener. In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, calculating both the first and the second distance contributes to an even more realistic reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF.

[0030] Emulating the one or more sound beams preferably comprises of emulating a delay of the one or more sound beams on the basis of the calculated distance. “Delay” refers to the difference in time between the moment at which a sound beam is emitted by a virtual loudspeaker source and the moment at which it is picked up by the virtual microphone on the head of the listener. Emulating the delay of the sound beams on the basis of the calculated distance contributes to an even more realistic and more accurate reproduction of sound. Emulating the delay of the sound beams on the basis of the distance between the virtual loudspeaker source and the virtual microphones on the head enables a more realistic reproduction of the sound sources to be achieved. This results in an improved sound experience for the listener, wherein the HRTF is crucial for a realistic and accurate reproduction of sound.

[0031] In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, emulating the delay of the sound beams contributes to an even more realistic and more efficient reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF. Calculating the one or more reflections preferably comprises of calculating at least first-order reflections. “Reflections” refers to the reflection of sound waves from surfaces in the space, such as walls, floors and ceilings. “First-order reflections” refers to the reflections returning directly from the surface to the head of the listener without additional reflections. Calculating at least first- order reflections contributes to an even more accurate and more realistic simulation of the sound waves, and therefore an even better HRTF. Calculating the first-order reflections enables a more realistic reproduction of the acoustics of the space to be achieved. This results in an improved sound experience for the listener.

[0032] In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, calculating at least first-order reflections contributes to an even more realistic and more efficient reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF. Calculating the one or more reflections more preferably comprises of calculating at least first-order reflections and second- order reflections. “Second-order reflections” refers to reflections which are caused by the reflection of sound waves from surfaces which are indirect, i.e. via two surfaces. Calculating at least first- order and second-order reflections contributes to an even more realistic and more accurate simulation of the sound waves, and therefore an even better HRTF. Calculating the first-order and second-order reflections enables a more realistic reproduction of the acoustics of the space to be achieved, including the indirect reflections of sound waves. This results in an improved sound experience for the listener. In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, calculating at least first-order and second-order reflections contributes to an even more realistic and more efficient reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF. Calculating the one or more reflections still more preferably comprises of calculating at least first-order reflections, second-order reflections and third-order reflections. “Third-order reflections” refers to reflections caused by the reflection of sound waves from surfaces which are indirect, via two or more other surfaces. Provision can be made for the calculating of at least first-order, second-order and third-order reflections to contribute to an even more realistic and more accurate simulation of the sound waves, and therefore an even better HRTF. Calculating the first-order, second-order and third-order reflections enables a more realistic reproduction of the complex acoustics of the space to be achieved, including the indirect reflections of sound waves via a plurality of surfaces. This results in an improved sound experience for the listener. The step of calculating is preferably performed exclusively for the sound beams intersecting the head. Limiting the calculation only to the sound beams intersecting the head increases the efficiency of the method further. Calculating only the sound beams that intersect the head enables the calculation of the sound parameters to be limited to a smaller number of sound beams. This results in a more efficient and more cost-effective method for generating an HRTF, without having an adverse effect on the quality of the sound reproduction.

[0033] In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, limiting the calculation only to the sound beams intersecting the head contributes to an even more efficient and more cost-effective reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF.

[0034] According to a further aspect, a computer-implemented method is provided for generating a two- channel binaural signal for the purpose of bringing about a three-dimensional, 3D, sound experience. The method comprises the following steps of:

[0035] - simulating in a first simulation step a beam behaviour of one or more sound beams coming from one or more virtual loudspeaker sources depending on a head of a listener in a space;

[0036] - calculating on the basis of the first simulation step one or more reflections of the one or more sound beams of the one or more loudspeaker sources;

[0037] - emulating in an emulation step one or more sound parameters representative of at least one of a time difference, an attenuation of the one or more sound beams, a delay of the one or more sound beams, a diffraction of the one or more sound beams and spectral information of the one or more sound beams on the basis of the calculated one or more reflections; and

[0038] - applying the emulated one or more sound parameters to a plurality of audio signals in order to create a multi-channel audio signal, and

[0039] - converting the multi-channel audio signal to a two-channel binaural signal.

[0040] According to a further aspect, a computer program is provided, comprising a program, which can be run by a computer, with instructions for performing the steps of the method as described above when performed on a computer.

[0041] Brief description of the figures The above stated and other advantageous features and objects of the invention will become more apparent, and the invention better understood, on the basis of the following detailed description when read in combination with the accompanying drawings, in which:

[0042] Figure 1 shows a flow diagram of an exemplary embodiment of a computer-implemented method for generating a head-related transfer function; and

[0043] Figure 2 shows a flow diagram of a further exemplary embodiment of a step of simulating shown in figure 1.

[0044] Detailed embodiments

[0045] The following detailed description relates to determined specific embodiments. The teaching hereof can however be applied in different ways. The same or similar elements are designated in the drawings with the same reference numerals.

[0046] The present invention will be described with reference to specific embodiments. The invention is however not limited thereto, but solely by the claims.

[0047] As used here, the singular forms “a” and “the” comprise both the singular and plural references, unless clearly indicated otherwise by the context.

[0048] The terms “comprising”, “comprises” and “composed of’ as used here are synonymous with “including”. The terms “comprising”, “comprises” and “composed of’ when referring to stated components, elements or method steps also comprise embodiments which “consist of’ the components, elements or method steps.

[0049] The terms first, second, third and so on are further used in the description and in the claims to distinguish between similar elements and not necessarily to describe a sequential or chronological order, unless this is specified. It will be apparent that the thus used terms are mutually interchangeable under appropriate circumstances and that the embodiments of the invention described here can operate in an order other than described or illustrated here.

[0050] Reference in this specification to “one embodiment”, “an embodiment”, “some aspects”, “an aspect” or “one aspect” means that a determined feature, structure or characteristic described with reference to the embodiment or aspect is included in at least one embodiment of the present invention. The manifestations of the sentences “in one embodiment”, “in an embodiment”, “some aspects”, “an aspect” or “one aspect” in different places in this specification thus do not necessarily all refer to the same embodiment or aspects. As will be apparent to a skilled person in this field, the specific features, structures or characteristics can further be combined in any suitable manner in one or more embodiments or aspects. Although some embodiments or aspects described here comprise some but no other features which are included in other embodiments or aspects, combinations of features of different embodiments or aspects are further intended to fall within the context of the invention and to form different embodiments or aspects, as would be apparent to the skilled person. In the appended claims all features of the claimed embodiments or aspects can for instance be used in any combination.

[0051] Figure 1 shows a flow diagram of a computer-implemented method 1000 for generating a Head-Related Transfer Function, HRTF, for bringing about a three-dimensional, 3D, sound experience. The term computer-implemented means that the method is performed by means of a computer or processor. An example hereof is the use of software which performs the method in order to generate the HRTF. A head-related transfer function, HRTF, is a function which describes the transfer of sound from a source to the eardrums of a listener on the basis of the position of the source and the listener. An HRTF is unique to each individual due to the differences in the shape of the head and the eardrums. An example hereof is the difference in the way in which sound is transferred to the eardrums of a person who is tall or short. A three-dimensional 3D sound experience means that the sound is presented in three dimensions, making it seem as if the sound is coming from different directions and creating the sensation that the listener is present in a determined space. An example hereof is the experience of hearing a concert as if the listener were present in the concert hall. In order to apply an HRTF to a sound output the HRTF is implemented in a software program which processes the sound before it is played. When a 3D sound experience is created in a video game, the HRTF of the player is for instance used to process the sound, so that the sound appears to be coming from different directions. This is achieved by adjusting the amplitude and phase of the sound on the basis of the HRTF of the player, whereby the sound appears to be coming from different directions and the listener is made to feel as if the sound is present in a determined space, as will be further elucidated. Another possible application of HRTFs is in the development of headphones offering a 3D sound experience. In this case the HRTF of the user can be used to process the sound being sent to the headphones, so that the sound appears to be coming from different directions and the listener is made to feel as if the sound is present in a determined space. In short, the HRTF is used to process the sound so that the sound appears to be coming from different directions and the listener is given the sensation of the sound being present in a determined space.

[0052] The computer-implemented method 1000 comprises the following steps of:

[0053] - simulating 100 in a first simulation step a beam behaviour of one or more sound beams (also referred to as rays) coming from one or more virtual loudspeaker sources depending on a head of a listener in a space;

[0054] - calculating 200 on the basis of the first simulation step one or more reflections of the one or more sound beams of the one or more loudspeaker sources;

[0055] - emulating 300 in an emulation step one or more sound parameters representative of at least one of a time difference, an attenuation of the one or more sound beams, a delay of the one or more sound beams, a diffraction of the one or more sound beams and spectral information of the one or more sound beams on the basis of the calculated one or more reflections; and

[0056] - generating 400 the HRTF on the basis of the one or more sound parameters emulated in the emulation step.

[0057] The first step, the simulating 100, involves the behaviour, i.e. the beam behaviour, of a determined phenomenon, i.e. the sound beams of the one or more loudspeaker sources, being imitated using a model. The object of the simulation 100 is to obtain a realistic reproduction of the reality around a listener and simulate the behaviour around the listener. In the context of the method 1000 simulating the beam behaviour of sound beams means that the direction and intensity of the sound beams coming from virtual loudspeaker sources are simulated on the basis of the position of the source and the listener. This is done by making use of for instance geometrical acoustics, i.e. modelling the sound waves as straight lines moving through the space and reflecting from surfaces. An example of simulating sound beams is the use of a computer program for calculating how the sound spreads in a room from a virtual loudspeaker source. Taken into consideration here are the distance between the source and the listener, and the presence of furniture and other obstacles in the room. The term virtual loudspeaker sources refers to imaginary sources of sound which are used to simulate the sound. It will be apparent that one or more virtual loudspeaker sources can be present. These sources are not physically present in the space, but are used in order to provide a realistic reproduction of how the sound spreads in the space. An example of a virtual loudspeaker source is an imaginary loudspeaker which is used to imitate the sound of a concert in a video game. A head of a listener refers to the head of a person listening to the sound to which the HRTF is being applied. It is mainly the head and the ears of the listener that determine the acoustic properties. The HRTF is generated on the basis of the specific properties of the head of the listener and is applied to the sound in order to create a realistic and accurate reproduction of the sound.

[0058] In a second step the method comprises of calculating 200 on the basis of the first simulation step one or more reflections of the one or more sound beams of the one or more loudspeaker sources. The term reflections refers to the phenomenon that sound waves reflect from surfaces in a space and are reflected in different directions. This results in a complex sound environment which can be used to create a realistic sound experience. In the context of the method 1000 calculating reflections means that the computer program calculates the reflections of the sound beams from surfaces in the space on the basis of the position of the virtual loudspeaker sources and the listener. An example of reflections is when sound beams hit a wall and are reflected to the listener, which results in an echo. This can be used to create a realistic acoustic environment, for instance in a video game or a film. The sound beams move from the loudspeaker to different parts of the room and are reflected from different surfaces, depending on the shape and size of the room.

[0059] In a subsequent step the method comprises of emulating 300 in an emulation step one or more sound parameters representative of at least one of a time difference, an attenuation of the one or more sound beams, a delay of the one or more sound beams, a diffraction of the one or more sound beams and spectral information of the one or more sound beams on the basis of the calculated one or more reflections. A parameter is a variable in a function or formula which affects the value of that function or formula. In this respect the term emulating means to imitate a determined sound parameter on the basis of the calculated reflections of sound beams from surfaces in the space. The object of the emulation step is to simulate and to model the properties of the sound as accurately as possible so that the HRTF is as realistic as possible. In other words, the emulation step 300 includes the computer using the calculated reflections of sound waves to simulate and to calculate different sound parameters. Emulating these parameters enables the computer to generate an HRTF which is as realistic as possible and which provides an accurate reproduction of the sound in the space. The term sound parameters refers to different properties of the sound, such as time difference, attenuation, delay, diffraction and spectral information. Time difference refers to the difference in arrival time of the sound beams at the ears of the listener. Attenuation is the decrease in amplitude of the sound as it spreads further away from the source. Delay refers to the difference in time between the arrival of the sound beams at the ears of the listener. Diffraction is the bending of sound beams when the sound beams encounter an obstacle. Spectral information refers to the frequency content of the sound. An example of time difference is when a sound wave coming from a determined direction arrives at one ear before arriving at the other ear, this contributing to the 3D audio experience. An example of attenuation is when the sound from a source diminishes as it spreads further away from the source. An example of delay is when the sound arrives at the ears of the listener from different directions at different speeds. An example of diffraction is when the sound wave encounters an obstacle, such as a wall, and is bent. An example of spectral information is the use of frequency filters for processing the sound and changing the frequency content thereof. In the emulation step of the method 1000 these sound parameters are emulated on the basis of the calculated reflections of sound beams from surfaces in the space. This is done in order to make the HRTF as accurate as possible, so that the 3D audio experience is as realistic as possible for the listener. It is noted that simulating refers to imitating a determined phenomenon using the method, i.e. a sound system in a determined space with a listener in the space. The object of simulation is to provide a realistic reproduction of reality and to understand how the phenomenon works. In the context of the patent application simulating thus refers to imitating the beam behaviour of the sound beams and calculating the reflections of the sound waves from surfaces in the space. Emulating refers to imitating a determined system using the simulation. The object of emulation is to replicate the functionality of the imitated system or device on the emulating system or device. Emulating thus for instance refers to imitating the sound parameters which are representative of a delay of the sound beams, on the basis of the calculated distance between the virtual loudspeaker source and the virtual microphones on the head of the listener.

[0060] After emulating 300 the sound parameters the HRTF is generated 400 on the basis of the one or more sound parameters emulated in the emulation step. The HRTF is an important factor in creating a realistic and immersive audio experience because the HRTF describes the acoustic transfer function between a sound source and the ear of a listener. The HRTF is applied to adjust the sound to the acoustic properties of a space and / or to the anatomy of the listener. This is done in applications such as audio processing, virtual reality and augmented reality. In the context of audio processing the HRTF is for instance used to compensate the sound for the acoustic properties of the space in which it is being played. This may for instance be necessary in a concert hall where the acoustics are different than in a recording studio. Applying the HRTF to a sound signal enables the sound to be adjusted, for instance in order to maintain the correct timbre and clarity.

[0061] In applications such as virtual reality and augmented reality the HRTF is for instance used to adjust the sound to the anatomy of the listener.

[0062] Figure 2 shows a further exemplary embodiment of the simulating of the beam behaviour of the sound waves. This is because figure 2 shows that, during the simulation 100 of the beam behaviour, the method 1000 preferably comprises of raytracing 110 one or more sound beams for each of the virtual loudspeaker sources. Raytracing is a technique for simulating the interaction of sound with objects in a space. Raytraced sound beams are sound waves which have been simulated using raytracing. Applying raytracing during simulation of sound beams provides several technical advantages. An advantage is that a more realistic image of the reflections of the sound in the space can be obtained, this resulting in a more accurate HRTF. Another advantage is that the use of raytracing can make calculation of the sound beams more efficient, this resulting in a faster and more cost-effective method. A further improvement of the reproduction of the 3D audio experience can therefore be achieved by applying raytracing.

[0063] The method 1000 preferably comprises of raytracing 110 the sound beams according to the following steps of:

[0064] - Simulating 111 each loudspeaker source of the one or more loudspeaker sources as a sphere with a predetermined diameter and a predetermined resolution, wherein the predetermined resolution is at least one parameter which is representative of the number of sound beams being emitted from the sphere. - Distributing 112 a number of points uniformly over the sphere, wherein the number of points corresponds with the number of sound beams.

[0065] - Emitting 113 a respective sound beam from each point on the sphere.

[0066] When simulating 100 sound waves using raytracing 111, it is advantageous to simulate 111 each loudspeaker source of the one or more loudspeaker sources as a sphere with a predetermined diameter and resolution. The predetermined resolution is seen here as a parameter which is representative of the number of sound beams being emitted from the sphere. This means that a determined number of points is distributed over the sphere in order to emit sound beams. An example of simulating a loudspeaker source as a sphere with a predetermined diameter and resolution is simulating a surround sound system in a home cinema. Simulating 111 the loudspeaker sources as spheres with a predetermined diameter and resolution enables the sound reproduction to be calculated more efficiently and accurately.

[0067] When simulating sound waves using raytracing it is advantageous to make 112 a uniform distribution of a number of points over the sphere. The number of points distributed corresponds with the number of sound beams being emitted. An example of distributing points uniformly over a sphere is simulating the acoustics of a concert hall. Distributing points uniformly over the sphere and emitting sound beams enables a realistic image of the acoustics of the concert hall to be obtained. A technical advantage of the use of this method is that it results in a more accurate simulation of the sound waves. Distributing points uniformly over the sphere prevents holes from occurring in the simulation. This results in a more realistic reproduction of the sound waves and a more accurate HRTF. By simulating each loudspeaker source as a sphere with a predetermined diameter and resolution and distributing points uniformly over the sphere an efficient and accurate simulation of the sound waves can be achieved. Emitting a respective sound beam from each point on the sphere results in the sound waves being reproduced realistically, this resulting in a more accurate HRTF. The use of these steps in raytracing of the sound beams enables an improved and efficient reproduction of the 3D audio experience.

[0068] When simulating 111 sound waves using raytracing, it is advantageous to simulate the head of the listener as an ellipsoid with virtual microphones opposite each other. This results in a more realistic simulation of the actual shape of the human head, this resulting in a more accurate HRTF. An example of simulating the head as an ellipsoid with virtual microphones opposite each other is the use of a virtual assistant in a room with a plurality of loudspeakers. Simulating the head of the listener as an ellipsoid with virtual microphones opposite each other enables the virtual assistant to reproduce the sound recordings more realistically, this resulting in a better sound experience for the listener. An ellipsoid is a three-dimensional figure which is formed by a flat, oval shape rotating about an axis. It is deemed a 3D version of an ellipse or oval. An ellipsoid has three different axes which can be used to define it: the long axis, the short axis and the central axis. A technical advantage of using this method is that it results in a more realistic simulation of the actual shape of the human head. This enables a more accurate reproduction of the sound recordings and a more accurate HRTF. This can result in a better 3D audio experience for the listener. The head can also be simulated as a sphere, although this approach is less accurate than simulating the head as an ellipsoid. This is because the shape of the human head is not perfectly round, but rather has oval and irregular shapes. Simulating the head as an ellipsoid enables a more realistic image of the actual shape of the human head to be obtained, this resulting in a more accurate HRTF. This can result in a better 3D audio experience for the listener.

[0069] The method 1000 preferably comprises of positioning the first and the second virtual microphones at an acute angle relative to one of three perpendicular axes of symmetry of the ellipsoid head. Positioning the virtual microphones at an acute angle relative to one of the perpendicular axes of symmetry results in a more realistic reproduction of the human ear canal and the directional dependence of human hearing. When positioning the first and the second virtual microphones at an acute angle relative to one of the three perpendicular axes of symmetry, the actual geometry of the human ear canal and the directional dependence of human hearing which is created by the ears are taken into consideration. This is because the human ear canal curves and does not enter in a straight line. Positioning the virtual microphones at an acute angle enables this curve in the ear canal to be simulated better, this resulting in a more realistic reproduction of the directional dependence of human hearing. This is further understood to mean that when the first and the second virtual microphones are positioned opposite each other on the head, the position of the virtual microphones corresponds with the position of the ears on a human head. A technical advantage of the use of this method is that it results in a more realistic reproduction of the human ear canal and the directional dependence of human hearing, this resulting in a more accurate and more realistic HRTF. This can result in a better 3D audio experience for the listener.

[0070] The method 1000 preferably comprises of calculating 120 one or more reflections using the raytraced sound beams 110. This means that the reflections of the sound waves are calculated on the basis of the simulated sound beams. Reflections refer to the phenomenon that sound waves reflect from surfaces in a space and are reflected in different directions. The use of raytraced sound beams for calculating reflections provides several technical advantages. An advantage is that the reflections of the sound in the space can be calculated more accurately, this resulting in a more realistic and more accurate HRTF. When simulating sound waves using raytracing, it is advantageous to calculate the number of reflections of a respective sound beam. This means that, for each sound beam, the number of reflections is calculated so as to simulate a realistic reproduction of the sound waves in the space. A technical advantage of using this method is that it contributes to a more realistic and more accurate reproduction of the 3D audio experience. This is because calculating the number of reflections improves the simulation of the sound waves in the space and results in a more realistic reproduction of the sound waves.

[0071] When simulating 111 sound waves using raytracing, it is further advantageous to calculate the angle of incidence of a respective sound beam relative to the ellipsoid head. This means that, for each sound beam, the angle of incidence relative to the head is calculated so as to simulate a realistic reproduction of the sound waves in the space. Calculating the angle of incidence of each sound beam relative to the head enables simulation of a realistic reproduction of the sound waves in the virtual environment, this resulting in a better 3D audio experience for the user. A technical advantage of using this method is that it contributes to an even better HRTF, and thereby an even more realistic reproduction of the 3D audio experience. This is because calculating the angle of incidence results in a more realistic reproduction of the directional dependence of human hearing and the response of the head to sounds from different angles of incidence. When simulating sound waves using raytracing, it is advantageous to calculate the distance travelled by a respective sound beam between a virtual loudspeaker source of the one or more loudspeaker sources and the head. A sound beam is understood to mean a bundle of one or more sound waves which are sent from a virtual loudspeaker source to the head of a listener. Calculating the distance travelled by a respective sound beam between a virtual loudspeaker source and the head of the listener enables simulation of a realistic reproduction of the sound waves in the concert hall, this resulting in a better 3D audio experience for the listener. A technical advantage of using this method is that it results in a more accurate calculation of the sound parameters which are used in generating the HRTF. This results in an improved sound experience for the listener, wherein the HRTF realizes a realistic and accurate reproduction of sound. In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, calculating the distance contributes to an even more realistic reproduction of the 3D audio experience.

[0072] The step of calculating reflections preferably comprises of performing calculations of at least first-order reflections. Reflections are the reflection of sound waves from surfaces such as walls, floors and ceilings in a space. First-order reflections are the sound waves which return directly to the listener from the surface, without additional reflections. Calculating these reflections enables a more realistic simulation of the sound waves to be obtained, this resulting in an improved HRTF. This contributes to a more accurate reproduction of the acoustics of the space, and provides an improved sound experience for the listener. In combination with other steps, such as simulating beam behaviour of sound beams and the use of raytracing, calculating at least first-order reflections results in a more efficient and more realistic 3D audio experience. Calculating the reflections more preferably also comprises of calculating at least first-order and second-order reflections. Second- order reflections are the reflections caused by the reflection of sound waves from indirect surfaces, i.e. via two or more surfaces. Calculating these reflections contributes to an even more realistic and more accurate simulation of the sound waves, this resulting in an even better HRTF. This improves the reproduction of the acoustics of the space, including the indirect reflections of sound waves, and improves the sound experience of the listener. In an even more preferred variant calculating the reflections also comprises of calculating at least first-order, second-order and third-order reflections. Third-order reflections are the reflections resulting from the reflection of sound waves from indirect surfaces, via two or more other surfaces. Calculating these reflections contributes to an even more realistic and more accurate simulation of the sound waves, this resulting in an even better HRTF. This results in a more realistic reproduction of the complex acoustics of the space, including the indirect reflections of sound waves via a plurality of surfaces, and ultimately improves the sound experience of the listener.

[0073] The step of calculating is preferably performed exclusively for the sound beams intersecting the head. Limiting the calculation only to the sound beams intersecting the head increases the efficiency of the method further. Calculating only the sound beams that intersect the head enables the calculation of the sound parameters to be limited to a smaller number of sound beams. This results in a more efficient and more cost-effective method for generating an HRTF, without having an adverse effect on the quality of the sound reproduction.

[0074] In combination with the other steps of the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, limiting the calculation only to the sound beams intersecting the head contributes to an even more efficient and more cost-effective reproduction of the 3D audio experience. This provides a further technical improvement of the method for generating an HRTF. Still more preferably, the calculating step is performed exclusively for the sound beams intersecting the head. Limiting the calculation only to these sound beams improves the efficiency of the method further. Limiting the calculations to the sound beams that intersect the head enables the calculation of the sound parameters to be limited to a smaller number of beams. This results in a more efficient and more cost-effective method for generating an HRTF, without any loss of sound quality. In combination with the other steps in the method, such as simulating the beam behaviour of the sound beams and the use of raytracing, limiting the calculations only to the sound beams crossing the head contributes to an even more efficient and more cost-effective reproduction of the 3D audio experience. This results in a further technical improvement of the method for generating an HRTF.

[0075] The following is an example of how a three-dimensional, 3D, sound experience can be realized using the above stated method. In other words, simulating beam behaviour, also referred to as raytracing, is the collecting of relevant data for each sound beam. In the above stated situation one or more sources, also referred to as loudspeakers, are placed in a virtual space together with a head of a listener. Each loudspeaker emits a predetermined number of beams in random directions. This is a standardised sampling of a unit sphere, for instance 1000 beams. Each beam is followed or raytraced until this beam hits the head or until said beam exceeds the stated maximum number of reflections. In this latter case the beam is rejected. The data collected for a sound beam from the simulation comprise the launch angle from the loudspeaker and the number of reflections, i.e. the number of times a surface is hit. In addition, the overall distance is measured to the beam which hits the head, wherein the final part of the surface toward the head is not counted, which will be important later. The coordinates on the surface of the final reflection point are also determined in order to construct the final beam from this point to the head. Finally, the angle of incidence relative to the head of the beam that crossed the head is registered.

[0076] In the next step the final beam (mentioned in the raytracing step) is used in order to obtain individual distances for both ears. In the case of the original source (number of reflections = 0) the “final coordinate” is the original source position. The path of the coordinates from the final reflection point to the centre of the head is followed here, and the same path is followed independently to both the left and the right ear. The coordinate on the head where the beam lies tangent to the head is then determined for each ear. The arc length to the ear is determined from this tangent point, the locus of all of these tangent points forming an ellipse in the case of an ellipsoid and a circle in the case of a sphere. If the straight path from point to ear is shorter than the sum of the arc length and the length of the tangent, the direct path is used. The length from the final reflection point to each ear and the length of the arc used is determined from this calculation step.

[0077] A further refinement in determining the paths of sound beams around the ellipsoidal head for each ear involves a geometric construction utilizing two planes. To accurately determine the distinct sound paths to each ear, particularly considering the head model, a geometric construction is employed. A first plane is determined by the collection of points where the sound beams originating from the virtual loudspeaker source are tangent to the surface of the ellipsoidal head. Concurrently, a second plane is defined by three points: the virtual loudspeaker source, the origin (i.e. center) of the ellipsoidal head, and the particular ears for which the sound path is being calculated. It is understood that this second plane will be distinct for each of the ears.

[0078] These two planes, intersect along a common line. This intersection line, in turn, intersects the surface of the ellipsoidal head at two distinct points. Typically, one of these points will be located closer to the ear, and the other further away. When simulating the sound beams and their propagation towards the ears, both paths corresponding to these two intersection points on the ellipsoid can be utilized for a given ear. Employing both these paths allows for a more correct and comprehensive representation of the sound radiation pattern as it interacts with and propagates around the listener's head. Furthermore, depending on the relative arc lengths of these two distinct paths to the ear, the sound rays can constructively interfere (reinforce) or destructively interfere (cancel out) at certain frequencies, which significantly contributes to the spectral shaping of the resulting HRTF. It is noted that, while precise, an exact analytical formula for the arc length along the surface of an ellipsoid may not always be available, and therefore, these arc lengths may be calculated using an appropriate approximation method.

[0079] Based on these determined paths, the coordinate on the head where the beam lies tangent to the head is determined for each ear. The arc length to the ear is determined from this tangent point, the locus of all of these tangent points forming an ellipse in the case of an ellipsoid and a circle in the case of a sphere. If the straight path from point to ear is shorter than the sum of the arc length and the length of the tangent, the direct path is used. The length from the final reflection point to each ear and the length of the arc used is determined from this calculation step.

[0080] Applying this further refinement, a computer-implemented method for generating a Head- Related Transfer Function (HRTF) to create a realistic three-dimensional (3D) sound experience in real time may be realized. This method significantly enhances the accuracy and efficiency of 3D audio reproduction.

[0081] The refined method begins with a first simulation step focused on determining the beam behavior of sound originating from one or more virtual loudspeaker sources. This simulation is dependent on the position of a listener's head within the simulated space. To achieve a highly realistic representation of sound interaction with the human anatomy, the head is simulated as an ellipsoid. This ellipsoidal head model is equipped with a first virtual microphone and a second virtual microphone, strategically positioned opposite each other on the head to correspond with the listener's ears. To further enhance realism and directional perception, these virtual microphones are preferably positioned at an acute angle relative to one of the three perpendicular axes of symmetry of the ellipsoid, with an angle of at least 60°, and more preferably at least 65°, yielding a better approximation of the actual geometry of the human ear canal and directional hearing.

[0082] A further refinement in determining the paths of sound beams around the ellipsoidal head for each ear involves a geometric construction utilizing two planes. To more accurately determine the distinct sound paths to each ear, particularly considering the head model, a geometric construction is employed. A first plane is determined by the collection of points where the sound beams originating from the virtual loudspeaker source are intersecting the surface of the ellipsoidal head. Concurrently, a second plane is defined by three points: the virtual loudspeaker source, the origin (i.e. center) of the ellipsoidal head, and the particular ears for which the sound path is being calculated. It is understood that this second plane will be distinct for each of the ears.

[0083] These two planes, thus defined, intersect along a common line. This intersection line, in turn, intersects the surface of the ellipsoidal head at two distinct points. Typically, one of these points will be located closer to the ear, and the other further away. When simulating the sound beams and their propagation towards the ears, both paths corresponding to these two intersection points on the ellipsoid can be utilized for a given ear. Employing both these paths allows for a more correct and comprehensive representation of the sound radiation pattern as it interacts with and propagates around the listener's head. Furthermore, depending on the relative arc lengths of these two distinct paths to the ear, the sound rays can constructively interfere (reinforce) or destructively interfere (cancel out) at certain frequencies, which significantly contributes to the spectral shaping of the resulting HRTF. It is noted that, while precise, an exact analytical formula for the arc length along the surface of an ellipsoid may not always be available, and therefore, these arc lengths may be calculated using an appropriate approximation method.

[0084] Based on these determined paths, the coordinate on the head where the beam intersect the head is determined for each ear. The arc length to the ear is determined from this tangent point, the locus of all of these tangent points forming an ellipse in the case of an ellipsoid and a circle in the case of a sphere. If the straight path from point to ear is shorter than the sum of the arc length and the length of the tangent, the direct path is used. The length from the final reflection point to each ear and the length of the arc used is determined from this calculation step.

[0085] The simulation of these sound beams, including their interaction with the ellipsoidal head and the determination of their behavior through the aforementioned geometric construction, is preferably performed using raytracing as previously discussed.

[0086] Following the initial beam simulation, the method proceeds to a calculation step to determine one or more reflections of these sound beams. These reflections are understood to originate from one or more surfaces within the simulated space (e.g., walls, floors, and ceilings). The calculation of these reflections is also preferably performed using the raytraced sound beams, ensuring consistency in the simulation methodology. To capture the full acoustic environment, the method may calculate various orders of reflections: at least first-order reflections (direct reflections from a single surface), optionally extending to include second-order reflections (reflections from two surfaces), and even third-order reflections (reflections from three or more surfaces), thereby progressively increasing the realism of the simulated acoustics. For efficiency, these calculations are preferably performed exclusively for the sound beams that are determined to intersect the head.

[0087] Within this calculation step, various specific parameters are determined for each sound beam. This includes calculating the number of reflections a respective sound beam undergoes, which contributes to simulating the realistic reproduction of sound waves. Additionally, the angle of incidence of a respective sound beam relative to the head is calculated, which is vital for simulating the directional dependence of human hearing. Furthermore, the distance travelled by a respective sound beam between a virtual loudspeaker source and the head is calculated. This overall distance is refined by calculating a first distance to the first virtual microphone and a second distance to the second virtual microphone, crucial for binaural processing. The method then enters an emulation step, where one or more sound parameters are emulated based on the calculated reflections and other determined characteristics. These emulated parameters are representative of at least a time difference, an attenuation of the one or more sound beams, and a delay of the one or more sound beams. Optionally, the emulated parameters may further include representations of diffraction of the one or more sound beams and spectral information of the one or more sound beams. The emulation of delay, in particular, is based on the previously calculated distances travelled by the sound beams, directly translating path length into temporal differences. These parameters are essential for accurately shaping the HRTF.

[0088] Finally, the method concludes by generating the HRTF itself. This HRTF is constructed directly from the array of emulated sound parameters, resulting in a highly accurate, realistic, and real-time capable HRTF that delivers an immersive 3D sound experience to the listener.

[0089] In order to realize a binauralization of all sources the distance of the raytracing and the distance calculation are combined. A total time delay is deduced therefrom. The total air filtering is also deduced therefrom. In this context air filtering refers to the adjustments in frequency and loudness which occur when sound travels through the air. This comprises of damping determined frequencies and reducing the sound intensity as the distance increases. The number of reflections is used as a multiplier of the absorption coefficient of the room, which functions as a frequency filter. The launch angle from the loudspeaker is applied in a polar pattern, which serves both as frequency filter and loudness filter. The angle of incidence to the head is recalculated to the ears, and likewise used in a polar pattern, which also acts as frequency filter and loudness filter. The length of the arc is additionally used as a multiplier in a frequency filter, wherein high-frequency roll-off occurs because these frequencies bend less. Adding up the different beams results in early reflections. An additional convolution reverb can additionally also be added in order to create the late reverb tail. Once this has been performed for each beam, the beams for this source are added up. This results in a binaural reproduction of this source. This is performed for each source, all these signals being added together so that the end result is a binauralization of all sources.

[0090] The skilled person will appreciate on the basis of the above description that the invention can be embodied in different ways and on the basis of different principles. The invention is not limited here to the above described embodiments. The above described embodiments and the figures are purely illustrative and serve only to increase understanding of the invention. The invention will not therefore be limited to the embodiments described herein, but is defined in the claims.

Claims

22Claims1. A computer-implemented method for generating a head-related transfer function, HRTF, for bringing about a three-dimensional, 3D, sound experience in real time, the method comprising the following steps of:- simulating in a first simulation step a beam behaviour of one or more sound beams coming from one or more virtual loudspeaker sources depending on a position of a head of a listener in a space, wherein the simulating comprises- simulating the head as an ellipsoid with a first virtual microphone and a second virtual microphone, wherein the first and the second virtual microphones are positioned opposite each other on the head and comprises determining a first plane defined by points where sound beams originating from the one or more virtual loudspeaker sources are intersecting the surface of the ellipsoidal head;- determining a second plane defined by the one or more virtual loudspeaker sources, a center of the ellipsoidal head, and the first and second virtual microphones respectively;- identifying an intersection line formed by the first plane and the second plane;- determining two distinct intersection points of said intersection line with the surface of the ellipsoidal head; and- utilizing one or more paths through both of said two distinct intersection points for determining said beam behavior of the one or more sound beams;- calculating on the basis of the first simulation step one or more reflections of the one or more sound beams of the one or more loudspeaker sources, said reflections originating from one or more surfaces in the space;- emulating, in an emulation step, one or more sound parameters representative of at least a time difference, an attenuation and, a delay of the one or more sound beams, optionally further representative of a diffraction of the one or more sound beams or spectral information of the one or more sound beams, on the basis of the calculated one or more reflections; and- generating the HRTF on the basis of the one or more sound parameters emulated in the emulation step.

2. The computer-implemented method according to the foregoing claim, wherein the simulation step comprises of raytracing the one or more sound beams for each of the one or more virtual loudspeaker sources during simulation of the beam behaviour.

3. The computer-implemented method according to the foregoing claim, wherein calculating the one or more reflections is performed using the ray traced one or more sound beams.

4. The computer-implemented method according to any one of the foregoing claims 2-3, wherein the raytracing comprises the following steps of:- simulating each loudspeaker source of the one or more loudspeaker sources as a sphere with a predetermined diameter and a predetermined resolution, wherein the predetermined resolution is at least one parameter representative of the number of sound beams being emitted from the sphere;- distributing a number of points uniformly over the sphere, wherein the number of points corresponds with the number of sound beams;- emitting a respective sound beam from each point on the sphere.

5. The computer-implemented method according to any one of the foregoing claims, wherein the first and the second virtual microphones have an acute angle relative to one of three perpendicular axes of symmetry.

6. The computer-implemented method according to the foregoing claim, wherein the acute angle is at least 60°, preferably at least 65°.

7. The computer-implemented method according to any one of the foregoing claims 4-6, wherein the step of calculating comprises of calculating the number of reflections of a respective sound beam.

8. The computer-implemented method according to any one of the foregoing claims 4-7, wherein the step of calculating comprises of calculating an angle of incidence of a respective sound beam relative to the head.

9. The computer-implemented method according to any one of the foregoing claims 4-8, wherein the step of calculating comprises of calculating a distance travelled by a respectivesound beam between a virtual loudspeaker source of the one or more loudspeaker sources and the head.

10. The computer-implemented method according to the claims 5 and 10, wherein calculating a distance comprises of calculating a first distance between the virtual loudspeaker source and the first virtual microphone and calculating a second distance between the virtual loudspeaker source and the second virtual microphone.

11. The computer-implemented method according to any one of the foregoing claims 9-10, wherein the emulating comprises of emulating a delay of the one or more sound beams on the basis of the calculated distance.

12. The computer-implemented method according to any one of the foregoing claims, wherein calculating the one or more reflections comprises of calculating at least first-order reflections.

13. The computer-implemented method according to the foregoing claim, wherein calculating the one or more reflections comprises of calculating at least first-order reflections and second-order reflections.

14. The computer-implemented method according to the foregoing claim, wherein calculating the one or more reflections comprises of calculating at least first-order reflections, second- order reflections and third-order reflections.

15. The computer-implemented method according to any one of the foregoing claims 4-14, wherein the step of calculating is performed exclusively for the sound beams intersecting the head.

16. A computer program comprising a program, which can be run by a computer, with instructions for performing the steps of the method according to any one of the foregoing claims when performed on a computer.

Images