Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04S—STEREOPHONIC SYSTEMS 
  • H04S3/00—Systems employing more than two channels, e.g. quadraphonic
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04S—STEREOPHONIC SYSTEMS 
  • H04S5/00—Pseudo-stereo systems, e.g. in which additional channel signals are derived from monophonic signals by means of phase shifting, time delay or reverberation 
  • H04S5/005—Pseudo-stereo systems, e.g. in which additional channel signals are derived from monophonic signals by means of phase shifting, time delay or reverberation  of the pseudo five- or more-channel type, e.g. virtual surround
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03G—CONTROL OF AMPLIFICATION
  • H03G3/00—Gain control in amplifiers or frequency changers
  • H03G3/20—Automatic control
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04S—STEREOPHONIC SYSTEMS 
  • H04S3/00—Systems employing more than two channels, e.g. quadraphonic
  • H04S3/002—Non-adaptive circuits, e.g. manually adjustable or static, for enhancing the sound image or the spatial distribution
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04S—STEREOPHONIC SYSTEMS 
  • H04S7/00—Indicating arrangements; Control arrangements, e.g. balance control
  • H04S7/30—Control circuits for electronic adaptation of the sound field
  • H04S7/308—Electronic adaptation dependent on speaker or headphone connection
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04S—STEREOPHONIC SYSTEMS 
  • H04S2400/00—Details of stereophonic systems covered by H04S but not provided for in its groups
  • H04S2400/01—Multi-channel, i.e. more than two input channels, sound reproduction with two speakers wherein the multi-channel information is substantially preserved
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04S—STEREOPHONIC SYSTEMS 
  • H04S2400/00—Details of stereophonic systems covered by H04S but not provided for in its groups
  • H04S2400/11—Positioning of individual sound objects, e.g. moving airplane, within a sound field

Definitions

• the present invention relates to audio signal processing and, in particular, to an apparatus and a method for edge fading amplitude panning for 3D loudspeaker setups.
• the move towards 3D audio can be regarded as the next step in the evolution of movie and home cinema sound systems.
• a greater number of loudspeakers can extend the listening area and improve the spatial resolution of the reproduced sound field.
• a greater number of loudspeakers also means a greater demand, because more loudspeakers need to be placed where they are supposed to be.
• the placement and the number of involved loudspeakers is a compromise between sound quality, costs, aesthetics, spatial limitations, and also domestic/social aspects (see [20]).
• Object-based audio scenes do not require a specific loudspeaker configuration like channel-based content and thus have less demands on the placement of the loudspeakers.
• the rendering process involves a panning method where the object's sound signal is played back by more than one loudspeaker (see [7]).
• VBAP Vector Base Amplitude Panning
• VBAP uses a set of N unit vectors l.v which point at the loudspeakers of the 3D speaker set.
• a panning direction given by a Cartesian unit vector p is defined by a linear combination of those loudspeaker vectors according to formula (1 ): where g clergy denotes the scaling factor that is applied to l n .
• g clergy denotes the scaling factor that is applied to l n .
• a vector space is formed by 3 vector bases.
• Formula (1 ) can generally be solved by a matrix inversion, if the number of active speakers and thus the number of non-zero scaling factors is limited to 3. Practically, this is done by defining a mesh of triangles between the loudspeakers and by choosing those triplets for the area in between. This leads to the solution
• VBAP exhibits particular properties.
• the vector arithmetic based concepts of VBAP are in relation to the sound field which is created by the involved loudspeakers.
• the base vector that corresponds to a certain loudspeaker e.g., Gerzon's velocity vector (see [9])
• Gerzon's velocity vector coincides with the particle velocity that can be measured under free field conditions at the listener position.
• a linear combination of the sound fields created by two or more loudspeakers results in the linear combination of the particle velocity.
• VBAP reproduces under free field conditions the particle velocity at the sweet spot that results from a sound source at the panning position.
• Fig. 19 illustrates the VBAP panning gains for a common 5.1 surround setup (see [13]).
• a generalized VBAP method using an imaginary loudspeaker (light gray) and a downmix is depicted.
• VBAP always uses 3 base vectors depending on the chosen triangulation. If the 3D setup consists of two or more height layers stacked on top of each other with loudspeakers at the same azimuth angles, then there is no preference for a certain triangulation. For each section between two speakers of a layer, there are two possibilities for subdividing the rectangle between the middle and the upper layer speakers into two triangles. This arbitrary choice introduces an asymmetry even for perfectly symmetric setups. To illustrate this property, let us take a 5.1 setup as an example that has been extended by four height speakers above the M30, M-30, M1 10, and M-1 10 speakers i.e., U30, U-30, U1 10, and U-1 10 [14].
• the object of the present invention is to provide improved concepts for amplitude panning.
• the object of the present invention is solved by an apparatus according to claim 1 , by a method according to claim 26 and by a computer program according to claim 27, by an apparatus according to claim 28, by a method according to claim 45 and by a computer program according to claim 46.
• the apparatus comprises a panning gain determiner and a signal processor.
• the panning gain determiner is configured to determine a proper subset from a set of five or more loudspeaker positions, so that the proper subset comprises four or more of the five or more loudspeaker positions.
• the panning gain determiner is configured to determine the proper subset depending on a panning position and depending on the five or more loudspeaker positions.
• the panning gain determiner is configured to determine a panning gain for each of the four or more audio output signals by determining said panning gain depending on the panning position and depending on the four or more loudspeaker positions of the proper subset.
• the signal processor is configured to generate each audio output signal of the four or more audio output signals depending on the panning gain for said audio output signal and depending on an audio input signal.
• the method comprises:
• Determining a panning gain for each of the four or more audio output signals by determining said panning gain depending on the panning position and depending on the four or more loudspeaker positions of the proper subset.
• a computer program for implementing the above-described method when being executed on a computer or signal processor is provided.
• an apparatus for generating four or more audio output signals is provided. Each loudspeaker position of four or more loudspeaker positions is associated with exactly one of the four or more audio output signals, and wherein each of the four or more audio output signals is associated with exactly one of the four or more loudspeaker positions.
• the apparatus comprises a panning gain determiner, and a signal processor.
• the panning gain determiner is configured to determine, for each audio output signal of the four or more audio output signals, a group of associated loudspeaker positions, being associated with said audio output signal, depending on the loudspeaker position of each of the four or more audio output signals and depending on a panning position, so that said group of associated loudspeaker positions comprises the loudspeaker position being associated with said audio output signal and at least two further loudspeaker positions of the four or more loudspeaker positions, wherein at least one of the four or more loudspeaker positions is not comprised by said group of associated loudspeaker positions.
• the panning gain determiner is configured to calculate, for each audio output signal of the four or more audio output signals, the panning gain for said audio output signal depending on the panning position and depending on the loudspeaker positions of the group of associated loudspeaker positions being associated with said audio output signal.
• the signal processor is configured to generate each audio output signal of the four or more audio output signals depending on the panning gain for said audio output signal and depending on an audio input signal.
• the group of associated loudspeaker positions being associated with a first one of the four or more audio output signals is not equal to the group of associated loudspeaker positions being associated with a different second one of the four or more audio output signals.
• Each loudspeaker position of four or more loudspeaker positions is associated with exactly one of the four or more audio output signals, and wherein each of the four or more audio output signals is associated with exactly one of the four or more loudspeaker positions.
• the method comprises:
• the group of associated loudspeaker positions being associated with a first one of the four or more audio output signals is not equal to the group of associated loudspeaker positions being associated with a different second one of the four or more audio output signals.
• the provided concepts provide a requirement-driven concept for amplitude panning.
• Fig. 1 is an apparatus according to an embodiment
• Fig. 2 illustrates 17 loudspeakers at 17 loudspeaker positions
• Fig. 3 illustrates an example for a determination of a proper subset of loudspeaker positions according to an embodiment dependent on a first panning position
• Fig. 4 illustrates another example for a determination of a proper subset of loudspeaker positions according to an embodiment dependent on a second panning position
• Fig. 5 illustrates the determination of two proper subsets of loudspeaker positions according to an embodiment depending on the first and the second panning position
• Fig. 6 illustrates the determination of two proper subsets of loudspeaker positions according to an embodiment depending on the first and a third panning position
• Fig. 7 illustrates five loudspeaker positions and a panning position
• Fig. 8 illustrates a first triangle-subdivision of a polygon defined body dependent on a first loudspeaker position according to an embodiment
• Fig. 9 illustrates a second triangle-subdivision of the polygon defined body dependent on a second loudspeaker position according to an embodiment, illustrates a third triangle-subdivision of the polygon defined body dependent on a third loudspeaker position according to an embodiment, illustrates a fourth triangle-subdivision of the polygon defined body dependent on a fourth loudspeaker position according to an embodiment, illustrates a fifth triangle-subdivision of the polygon defined body dependent on a fifth loudspeaker position according to an embodiment, illustrates triangle-subdivisions of another polygon-defined according to an embodiment, wherein the polygon-defined body is a quad, illustrates a triangle-subdivision of a further polygon-defined according to an embodiment, wherein the polygon-defined body is a hexagon, illustrates a subdivision of a further polygon-defined according to an embodiment, wherein the polygon-defined body is a octagon, which is subdivided into quads, illustrates a determination of a
• Fig. 24 illustrates a subdivision of polygon-defined body into triangles according to an embodiment
• Fig. 25 indicates trajectories reproduced by test signals in a listening test
• Fig. 26 shows the average and the 95% confidence interval of test results for a first listening test where the timbre was rated
• Fig. 27 illustrates a difference plot for the first listening test where the timbre was rated
• Fig. 28 shows the test results for a second test where the location accuracy and smoothness of movement was rated
• Fig. 29 illustrates a difference plot for the second listening test where the location accuracy and smoothness of movement was rated
• Fig. 30 shows the test results for a third test where the source extension and focus was rated
• Fig. 31 illustrates a difference plot for the third listening test where the source extension and focus was rated
• Fig. 32 illustrates the results for the overall quality
• Fig. 33 illustrates a difference plot for the results for the overall quality.
• Fig. 1 illustrates an apparatus for generating four or more audio output signals according to an embodiment.
• the apparatus comprises a panning gain determiner 1 10 and a signal processor 120.
• the panning gain determiner 1 10 is configured to determine a proper subset from a set of five or more loudspeaker positions, so that the proper subset comprises four or more of the five or more loudspeaker positions, wherein the panning gain determiner 1 10 is configured to determine the proper subset depending on a panning position and depending on the five or more loudspeaker positions.
• the panning gain determiner 1 10 is configured to determine a panning gain for each of the four or more audio output signals by determining said panning gain depending on the panning position and depending on the four or more loudspeaker positions of the proper subset.
• the signal processor 120 is configured to generate each audio output signal of the four or more audio output signals depending on the panning gain for said audio output signal and depending on an audio input signal.
• a proper subset of a set of five or more loudspeaker positions is a subset of the five or more loudspeaker positons which does not comprise at least one of the five or more loudspeaker positions.
• the panning gain determiner is configured to determine a proper subset from a plurality of five or more loudspeaker positions, so that at least four loudspeaker positions are comprised by the subset.
• Fig. 2 illustrates 17 loudspeakers at 17 loudspeaker positions 201 - 217.
• the 17 loudspeaker positions 201 - 217 define five pentagons 221 , 222, 223, 224 and 225.
• the pentagon 221 is defined by a polygon with the vertices 201 , 202, 203, 204 and 205.
• the pentagon 222 is defined by a polygon with the vertices 201 , 205, 206, 207 and 208.
• the pentagon 223 is defined by a polygon with the vertices 201 , 208, 209, 210 and 21 1 .
• the pentagon 224 is defined by a polygon with the vertices 208, 212, 213, 214 and 209. And, the pentagon 225 is defined by a polygon with the vertices 209, 214, 215, 216 and 217.
• the loudspeaker positions are positions within a two- dimensional coordinate system.
• the horizontal axis 231 of the two-dimensional coordinate system may, e.g., indicate an azimuth angle ⁇ of the loudspeaker position
• the vertical axis 232 of the coordinate system may, e.g., indicate an elevation angle ⁇ of the coordinate system.
• loudspeaker positions that are described only by an azimuth angle or an elevation angle may be positions (assumed to be) located on a sphere in the real three- dimensional world.
• the horizontal axis 231 of the coordinate system may, e.g., indicate an abscissa (x-axis) coordinate value of the loudspeaker position
• the vertical axis 232 of the coordinate system may, e.g., indicate an ordinate (y-axis) of a Cartesian coordinate system.
• all loudspeakers may be located in a plane.
• the bodies defined by the polygons are convex.
• the body defined by the polygon with the vertices 201 , 202, 203, 204 and 205 is convex.
• the body defined by the polygon with the vertices 201 , 205, 208, 207 and 208 is convex.
• the five polygons which define the five pentagons do not enclose any other loudspeaker position which does not belong to the respective polygon.
• the polygon with the vertices 201 , 202, 203, 204 and 205 does not enclose any of the loudspeaker positions 206 - 217.
• a panning position 241 is indicated.
• the loudspeaker arrangement shall now reproduce an audio input signal, as if the source emitting the audio input signal would be located at the panning position.
• the panning gain determiner 1 10 of Fig. 1 may, e.g., be configured to determine one of the polygons mentioned-above that enclose the panning position to determine the subset of loudspeaker positions.
• this is the (subset-specific) polygon with the vertices 201 , 202, 203, 204 and 205.
• the loudspeaker positions 201 , 202, 203, 204 and 205 are the (only) elements of the proper subset of loudspeaker positions.
• the polygon that defines the subset is subset-specific for the subset and can thus be referred to as subset-specific polygon.
• the panning gain determiner is now configured to determine a panning gain for each of the audio output signals depending on the panning position and depending on the loudspeaker positions of the (preselected) subset.
• Embodiments are based on the finding that to reproduce the audio input signal as if it would originate from the panning position 241 , it is sufficient that only the loudspeakers 201 , 202, 203, 204 and 205 output an output signal. The other loudspeakers are not needed.
• each audio output signal is generated for a particular loudspeaker position (or, in other words, for a loudspeaker which is associated with, or, e.g., positioned at, that particular loudspeaker position), it is sufficient to generate only the audio output signals for the loudspeakers at the loudspeaker positions of the proper subset, to reproduce an audio output signal, as if it would be emitted from the panning position.
• embodiments are based on the finding that at least four loudspeaker positions shall be in the subset, as at least four speakers should be employed to represent an audio input signal in the panning position.
• embodiments are also based on the finding that a reproduction of the audio input signal by only three speakers or less has disadvantages compared to using four or more speakers, in particular, when the panning position moves, as will be described further below. Therefore, the subset is a proper subset and does therefore not comprise all existing loudspeaker positions, but also, the subset comprises four or more loudspeaker positions.
• Fig. 4 again illustrates the 17 loudspeakers at the 17 loudspeaker positions 201 - 217.
• a new panning position is located at position 242.
• the new panning position 242 of Fig. 4 is different from the old panning position 241 of Fig. 3.
• the reason for this may for example be, that in the recording scene, a person which emits sound waves causing the audio input signal may have moved, so that at a later point in time, the panning position also moves from position 241 to position 242.
• panning position 242 may relate to the same point-in-time but to a further audio input signal.
• the panning position 241 may relate to a first audio input signal which may comprise the sound part of a violin in an orchestra.
• the panning position 242 may relate to a second audio input signal which may comprise the sound port of a trumpet in the orchestra. Then, in the reproduction scene, panning position 241 indicates that the violin shall be virtually positioned at panning position 241 , and the panning position 242 indicates that the trumpet shall be virtually positioned at panning position 242.
• the audio input signal relating to the violin is only reproduced by the loudspeakers at loudspeaker positions 201 , 202, 203, 204 and 205 and the further audio input signal relating to the trumpet is only reproduced by the loudspeakers at loudspeaker positions 208, 212, 213, 214 and 209 (see Fig. 5).
• panning gains for amplifying or attenuating the audio input signal representing the sounds from the violin are only calculated for the loudspeakers at loudspeaker positions 201 , 202, 203, 204 and 205.
• gains for amplifying or attenuating the further audio input signal representing the sounds from the trumpet are only calculated for the loudspeakers at loudspeaker positions 208, 212, 213, 214 and 209.
• gains g-, g 2 1 , 93,1 , 94,1 and g 5i1 for the loudspeakers at loudspeaker positions 201 , 202, 203, 204 and 205, respectively, are calculated by the panning gain determiner 1 10, and the signal processor 120 applies the calculated panning gains gi ⁇ , , g 2, i, g 3, i, g 4 .i and g 5 1 on the audio input signal ais 1 to obtain the audio output signals aosi , aos 2 , aos 3l aos 4 and aos 5 for the loudspeakers at the loudspeaker positions 201 , 202, 203, 204 and 205, respectively, for example, according to:
• gains g 8j2 , gi 2l2 , 9i3,2, 9i4,2 and g 9 2 for the loudspeakers at loudspeaker positions 208, 212, 213, 214 and 209, respectively, are calculated by the panning gain determiner 1 10, and the signal processor 120 applies the calculated panning gains g S 2 , g 12,2 , g i3 2 , 9i 2 and g 9 2 on the audio input signal ais 2 to obtain the audio output signals aos 9 , aos 12 , aos 13 , aos i4 and aos 9 for the loudspeakers at the loudspeaker positions 208, 212, 213, 214 and 209, respectively, for example, according to:
• the audio input signal comprises a plurality of audio input samples.
• the signal processor 120 may, e.g., be configured to generate each audio output signal of the four or more audio output signals by multiplying each of one or more of the audio input samples of the audio input signal with the panning gain for said audio output signal to obtain one or more audio output samples of the audio output signal.
• the audio input signal comprises a plurality of audio input samples
• the signal processor 120 is configured to generate each audio output signal of the four or more audio output signals by multiplying each of one or more of the audio input samples of the audio input signal with a square root of the panning gain for said audio output signal to obtain one or more audio output samples of the audio output signal.
• more than one audio output signal shall be reproduced by the same loudspeaker.
• the panning position 243 relating to an audio input signal ais 3 is located within a pentagon defined by the polygon with the vertices 201 , 208, 209, 210 and 21 1.
• panning gains g 1 3 , g 8 3 , g 8 3 . 9io ,3 and g 11 i3 relating to ais 3 are calculated by the panning gain determiner 1 10, and the signal processor 120 applies the calculated panning gains g 1i3 , g 8i3 , g 9,3 , g 1 0r 3 and g 11 i3 on the audio input signal ais 3 .
• the signal processor 120 may, e.g., be configured to obtain such an audio output signal by applying the respective gains on the respective audio input signals and by combining the respectively amplified or attenuated audio input signals.
• the calculated panning gain g 1 (1 is applied on aisi to obtain the amplified or attenuated ⁇ ais-
• the calculated panning gain g 3 is applied on ais 3 to obtain the amplified or attenuated g 1 i3 ⁇ ais 3 .
• ⁇ aisi and g 1 j3 ⁇ ais 3 are combined.
• the audio input signal may, e.g., be a first audio input signal, wherein the panning position is a first panning position, wherein the panning gain is a first input-signal-dependent panning gain, and wherein the proper subset is a first proper subset.
• the panning gain determiner 1 10 may, e.g., be configured to determine one or more further proper subsets from a set of five or more loudspeaker positions, so that each of the one or more further proper subsets comprises four or more of the five or more loudspeaker positions.
• the panning gain determiner 1 10 may, e.g., be configured to determine each of the one or more further proper subsets depending on one of one or more further panning positions and depending on the five or more loudspeaker positions, Moreover, the panning gain determiner 1 10 may, e.g., be configured to determine one or more further input-signal-dependent panning gains for each of the four or more audio output signals by determining each of the one or more further panning gains depending on one of the one or more further panning positions and depending on the four or more loudspeaker positions of one of the one or more further proper subsets.
• the signal processor 120 may, e.g., be configured to generate each audio output signal of the four or more audio output signals depending on the first input-signal-dependent panning gain for said audio output signal, depending on the one or more further input-signal-dependent panning gains for said audio output signal, depending on the audio input signal, and depending on the one or more further audio input signals.
• the panning position is sometimes also called a panning direction.
• the term panning direction originates from that for example, in an azimuth, elevation coordinate system, the panning position in the two- dimensional coordinate system is, in the real three-dimensional setup a direction information pointing from a central point, e.g., from a sweet spot to the direction of the loudspeaker.
• a direction information pointing from a central point, e.g., from a sweet spot to the direction of the loudspeaker.
• This aspect relates to how panning is realized between the loudspeaker positions of the determined subset, for example between the loudspeaker positions 208, 212, 213, 214 and 209 of Figs. 4 and 5.
• Fig. 7 illustrates the setup, showing the 5 loudspeaker positions 208, 212, 213, 214 and 209 of the loudspeakers and the panning position 242.
• Fig. 8 illustrates panning gain determination for the audio output signal for loudspeaker position 209.
• the body enclosed by the polygon with the vertices 208, 212, 213, 214 and 209 is subdivided into three triangles, namely a first triangle with vertices 209, 208, 212, a second triangle with vertices 209, 212, 213, and a third triangle with vertices 209, 213, 214, so that the subdivision of the body resulted in triangles that have loudspeaker position 209 (for which the panning gain is determined) as vertex.
• the panning gain for the audio output signal for loudspeaker position 209 is then calculated depending on loudspeaker positions 209, 212, 213, and not by the remaining loudspeaker positions 208 and 214. This simplifies computation and heips to save processor time compared to using all loudspeaker positions when calculating the panning gain associated with the audio output signal to be generated for loudspeaker position 209.
• the panning gain determiner has determined a group of associated loudspeaker positions comprising the loudspeaker positions 209, 212, 213, wherein the group of associated loudspeaker positions is associated the audio output signal for the loudspeaker at loudspeaker position 209 and determines which of the loudspeaker positions are taken into account when calculating the panning gain to obtain the output signal for (associated with) the loudspeaker position 209.
• the group of associated loudspeaker signals defines a triangle that is group- specific for the group of associated loudspeaker signals. In more general, the triangle 209,
• 212, 213 can be considered as a group-specific polygon with the vertices 209, 212, 213.
• FIG. 9 illustrates panning gain determination for the audio output signal for loudspeaker position 208.
• 213, 214 and 209 is subdivided into three triangles, namely a first triangle with vertices 208, 212, 213, a second triangle with vertices 208, 213, 214, and a third triangle with vertices 208, 214, 209, so that the subdivision of the body resulted in triangles that have loudspeaker position 209 (for which the panning gain is determined) as vertex.
• the panning gain for the audio output signal for loudspeaker position 208 is then calculated depending on loudspeaker positions 208, 212, 213, and not by the remaining loudspeaker positions 209 and 214.
• the panning gain determiner has determined a group of associated loudspeaker positions comprising the loudspeaker positions 208, 212, 213, wherein the group of associated loudspeaker positions is associated the audio output signal for the loudspeaker at loudspeaker position 208.
• Fig. 10 illustrates that the group of associated loudspeaker positions, being associated with the audio output signal for the loudspeaker at loudspeaker position 212, comprises the loudspeaker positions 212, 213, 214 and the panning gain to obtain said audio output signal is calculated depending on these loudspeaker positions 212, 213, 214.
• Fig. 1 1 illustrates that the group of associated loudspeaker positions, being associated with the audio output signal for the loudspeaker at loudspeaker position 213, comprises the loudspeaker positions 213, 208, 212 and the panning gain to obtain said audio output signal is calculated depending on these loudspeaker positions 213, 208, 212.
• Fig. 12 illustrates that the group of associated loudspeaker positions, being associated with the audio output signal for the loudspeaker at loudspeaker position 214, comprises the loudspeaker positions 214, 212, 213 and the panning gain to obtain said audio output signal is calculated depending on these loudspeaker positions 214, 212, 213.
• the triangle that encloses the panning position defines the group of associated loudspeaker positions.
• some embodiments choose one of the two triangles for calculating the panning gain.
• Other embodiments for example, calculate a first intermediate panning gain for a first one of the two triangles and further calculate a second intermediate panning gain for a second one of the two triangles, and then calculate the average of the first and the second intermediate panning gain as the final panning gain.
• the body defined by the polygon (here the polygon with the edges 208, 212, 213, 214, 209, which here defines a pentagon) it is preferable that the body is convex. Moreover, it is preferred that the body defined by the polygon is subdivided into triangles, such that a triangle does not enclose loudspeaker positions different from the loudspeaker positions that define the vertices of the triangle.
• the polygon with the loudspeaker positions as vertices does not define a pentagon, but defines any other kind of body with four or more vertices, for example, a quad, a hexagon, etc.
• Fig. 13 illustrates the subdivision of a quad for the audio output signals for loudspeakers at each of the loudspeaker positions 301 , 302, 303, 304 and for a panning position 305.
• Fig. 14 illustrates the subdivision of a hexagon with vertices 401 , 402, 403, 404, 405, 406 for the audio output signal for the loudspeaker at the loudspeaker position 401 and for a panning position 407.
• the group of associated loudspeaker positions for the audio output signal for the loudspeaker at the loudspeaker position 401 comprises the loudspeaker positions 401 , 403 and 404.
• Fig. 15 illustrates an example according to an embodiment, wherein the octagon with the vertices 501 , 502, 503, 504, 505, 506, 507 and 508 is subdivided into three quads, when the group of associated loudspeaker positions for the audio output signal for loudspeaker position 501 shall be determined, namely a first quad with the vertices 401 , 402. 403 and 404, a second quad with the vertices 401 , 404, 405 and 406 and a third quad with the vertices 401 , 406, 407, 408.
• the panning gain determiner calculates the panning gain associated with loudspeaker position 401 dependent on the panning position 409, and depending on the loudspeaker positions 401 , 404, 405, 406 of the group of associated loudspeaker positions for the audio output signal for loudspeaker position 401.
• the panning gain determiner 1 10 is configured to determine a group-specific polygon which encloses the panning position. Such a polygon is group specific for the group of associated loudspeaker signals.
• these concepts are based on the finding, that by determining a gain factor for each audio output signal for each of the loudspeaker positions creates a more realistic sound impression compared to only determining gain factors and thus audio output signals for the loudspeaker positions of a single triangle. Instead, embodiments determine gain factors for each loudspeaker position of the subset, although by only taking for each of the gain factors the loudspeaker positions of a gain-factor-specific triangle into account. However, as the corresponding triangles (or, more generally: sub-bodies) for determining the panning gains for the audio output signals differ for at least some of the audio output signals, this ensures, that all loudspeaker positions are taken into account for determining at least one of the gain factors. This is advantageous compared to always taking the same triangle into account for determining all panning gains.
• the panning gain for an audio output signal for a loudspeaker at a loudspeaker position may, for example, be determined depending on the panning position and depending on the loudspeaker positions of the group of associated loudspeaker positions.
• Fig. 16 illustrates a corresponding example showing the loudspeaker positions 501 , 502, 503, 504, 505 and panning position 506.
• the panning position 508 is located within the triangle of the loudspeaker positions 501 , 503 and 504, only the loudspeaker positions 501 , 503 and 504 belong to the group of associated loudspeaker positions, and only these loudspeaker positions 501 , 503, 504 are taken into account for determining the panning gain for the audio output signal for loudspeaker position 501 ⁇ and not the loudspeaker positions 502 and 505, which do not belong to this group of associated loudspeaker signals).
• Line 51 1 indicates a first distance being a shortest distance between the panning position 507 and a first straight line through the two further loudspeaker positions 503, 504 of the group of associated loudspeaker positions.
• Line 512 indicates a second distance being a shortest distance between the loudspeaker position 501 (for the audio output signal of which, the panning gain is determined) and a second straight line 515 through the panning position, wherein said second straight line is parallel to said first straight line 510.
• the panning gain determiner 1 10 may, for example, be configured to determine the panning gain depending of the ratio of the first distance 51 1 and a sum of the first distance 51 1 and the second distance 512. For example, assuming that in Fig. 16 the first distance 51 1 is 0.6 and the second distance 512 is 0.2, then, the panning gain /> 50! may, e.g., be calculated to be
• Fig. 17 illustrates another embodiment, wherein the group of associated loudspeaker positions comprises four loudspeaker positions 601 , 602, 603 and 604.
• the panning position is indicated by 605.
• the panning gain for the audio output signal for loudspeaker position 601 shall be determined.
• Mathematical concepts of the state of the art may be employed to determine a curve 608 through the loudspeaker positions 602, 603, 604.
• a dashed straight line 610 through loudspeaker position 601 and panning position 605 is illustrated.
• the intersection of the dashed straight line 610 and curve 608 defines intersection point 609.
• a first distance 61 1 is defined by the distance between panning position 605 and intersection point 609.
• a second distance 612 is defined by the distance between panning position 605 and loudspeaker position 601 .
• the panning gain determiner 1 10 may, for example, be configured to determine the panning gain depending of the ratio of the first distance 51 1 and a sum of the first distance 5 1 and the second distance 512. Assuming that in Fig. 17, the first distance 61 1 is 0.25 and that the second distance 612 is 0.3, the panning gain may, e.g., be
• the panning gain /3 ⁇ 4oi is slightly below 0.5 and this reflects that the loudspeaker position 601 is slightly farer away from the panning position 605 than the intersection point 609.
• no group of associated loudspeaker positions for determining each of the panning gains is determined. Instead, all loudspeaker positions of the proper subset are taken into account for calculating each gain.
• each loudspeaker position of the four or more loudspeaker positions of the proper subset is associated with exactly one of the four or more audio output signals, and wherein each of the four or more audio output signals is associated with exactly one of the four or more loudspeaker positions of the proper subset.
• the panning gain determiner 1 10 may, e.g., be configured to calculate, for each audio output signal of the four or more audio output signals, the panning gain for said audio output signal depending on a panning position and depending on the loudspeaker position of each of the four or more audio output signals.
• the panning gain determiner 1 10 may, e.g. , be configured to determine the panning gain for each audio output signal of the four or more audio output signals for a first point in time as a first time-dependent panning gain for said audio output signal. Moreover, the panning gain determiner 1 10 may, e.g., be configured to determine a further panning gain for each audio output signal of the four or more audio output signals for a different second point in time as a second time-dependent panning gain for said audio output signal, said second time-dependent panning gain being different from the first time-dependent panning gain for said audio output signal.
• different panning gains may, e.g., be determined for different frequencies.
• the panning gain determiner 1 0 may. e.g., be configured to determine the panning gain for each audio output signal of the four or more audio output signals for a first frequency as a first frequency-dependent panning gain for said audio output signal.
• the panning gain determiner 10 may, e.g., be configured to determine a further panning gain for each audio output signal of the four or more audio output signals for a different second frequency as a second frequency-dependent panning gain for said audio output signal, said second frequency-dependent panning gain being different from the first frequency-dependent panning gain for said audio output signal.
• Fig. 18 illustrates a system according to an embodiment.
• the system comprises five or more loudspeakers, and an apparatus as described above with reference to Fig. 1 .
• the apparatus is indicated by reference sign 100.
• the system of Fig. 18 comprises 17 loudspeakers at the loudspeaker positions 201 - 217.
• Each of the loudspeakers is associated with exactly one loudspeaker position 201 - 217 of the set of five or more loudspeaker positions.
• Each of the four or more audio output signals is associated with exactly one loudspeaker position of the loudspeaker positions 201 - 217 of the proper subset.
• each of the four or more audio output signals is associated with exactly one of the loudspeaker positions 201 - 217 of the proper subset.
• the system is configured to output each audio output signal of the four or more audio output signals by the loudspeaker which is associated with the same loudspeaker position as said audio output signal.
• the system may, e.g., be configured to output none of the four or more audio output signals by any of the four or more loudspeakers not being associated with the same loudspeaker position as said audio output signal.
• the system may, e.g., be configured to not output any of the four or more loudspeaker signals by at least one of the five or more loudspeakers.
• each loudspeaker position of four or more loudspeaker positions is associated with exactly one of the four or more audio output signals, and wherein each of the four or more audio output signals is associated with exactly one of the four or more loudspeaker positions.
• 1 is configured to determine, for each audio output signal of the four or more audio output signals, a group of associated loudspeaker positions, being associated with said audio output signal, depending on the loudspeaker position of each of the four or more audio output signals and depending on the panning position, so that said group of associated loudspeaker positions comprises the loudspeaker position being associated with said audio output signal and at least two further loudspeaker positions of the four or more loudspeaker positions, wherein at least one of the four or more loudspeaker positions is not comprised by said group of associated loudspeaker positions.
• the panning gain determiner 1 10 is configured to calculate, for each audio output signal of the four or more audio output signals, the panning gain for said audio output signal depending on the panning position and depending on the loudspeaker positions of the group of associated loudspeaker positions being associated with said audio output signal.
• the signal processor 120 is configured to generate each audio output signal of the four or more audio output signals depending on the panning gain for said audio output signal and depending on an audio input signal.
• the group of associated loudspeaker positions being associated with a first one of the four or more audio output signals is not equal to the group of associated loudspeaker positions being associated with a different second one of the four or more audio output signals.
• the provided concepts relate to 3D loudspeaker setups, as the 3D loudspeaker setup may be projected in the above-described two-dimensional coordinate system.
• Embodiments provide Edge Fading Amplitude Panning (EFAP) concepts for 3D loudspeaker setups. Similar to other panning methods like Vector Base Amplitude Panning (VBAP), it can be used to create phantom sources between the loudspeaker positions.
• the proposed method features symmetric panning gains for symmetric loudspeaker setups, N-wise panning by using polygons instead of triangles, and a better behavior for large opening angles between loudspeakers while involving a computational complexity that is in the same order of magnitude as VBAP,
• a solution would require the usage of polygons instead of triangles as boundary, resulting in N-wise panning.
• VBAP supports only triangles due to its fundamental principle, it can be generalized to yield N-wise panning as illustrated in Fig. 2. In doing so, an imaginary loudspeaker [1 ] is added in the middle of the polygon and its VBAP gain is then downmixed to its neighbors - a solution that is simpler than previously proposed solutions [2].
• Fig. 21 illustrates VBAP triangles in spherical coordinates for a 5.1 +4 setup (squares: loudspeaker positions).
• the squares mark the loudspeaker positions of the setup that was already used in the previous example.
• the solid lines result from the vectorial linear combination of the loudspeaker pairs which specify the edges of the VBAP triangles.
• the apparent geometric distortion of the triangles can be explained as follows:
• the triangles are the subdivided surfaces of the polyhedron which is defined by the loudspeakers at constant radius, By projecting the triangle edges onto the sphere surface, one yields the azimuth and the elevation angle as part of their spherical coordinates. Consequently, if the user wanted an object to be pair-wise panned between the U1 10 and the U-1 10 speaker located at 35° elevation, he would have to follow a trajectory that goes beyond 60° elevation. For a trajectory with constant elevation of 35°, VBAP would result in significant amplitudes for the M1 10 and M-1 10 loudspeaker channels.
• Panning concepts are provided that conform to these requirements. 2D considerations are extended for 3D setups.
• a 0 denotes the opening angle between the involved loudspeaker pair and a paragraph denotes the angle between the respective loudspeaker and the panning direction.
• Fig. 22 illustrates panning gains for a stereo setup (solid: VBAP; dotted: linear cross- fading; dashed: power normalized cross-fading).
• Fig. 22 depicts for an opening angle of 80° these cross-fading functions and the target curves which are given by the 2D VBAP panning gains, If in a second step the energy normalization (3) is applied to the linear cross-fading gains, like for VBAP, one can observe that the result closely approximates the given target curves.
• Fig. 23 illustrates a top view of an angular deviation between VBAP (+) and linear cross- fading ( ⁇ ).
• Fig. 23 illustrates the underlying approximation principle for the 60° example.
• the loudspeakers are the boundary of the shown angular range.
• the crosses mark the given target directions which consist of a set of equi-angular intermediate positions. If the two base vectors are multiplied by the corresponding cross- fading gains, the results are obtained that are marked by the squares.
• the angular deviations between the target directions and the result of the cross-fading approach are illustrated by the solid lines. From this geometric consideration it can be concluded a) that the approximation is the closer the smaller the loudspeaker opening angle is and b) that an opening angle of 80° can still be considered as being small.
• the cross-fading gain only depends on the ratio between the panning angle and the opening angle between the loudspeakers. Hence, a greater opening angle results in the dashed graph shown in Fig. 22 scaled along the x-axis (azimuth angle). This is a desired property as it complies with the seventh design goal.
• Power normalization may be conducted, e.g., by employing the formula:
• the parameter space is one-dimensional in the 2D case and only consists of the azimuth angle, it is two-dimensional in the 3D case and is spanned by the azimuth and the elevation angle.
• Fig. 24 illustrates a subdivision of polygon-defined body into triangles according to an embodiment.
• Fig, 24 illustrates a contour plot of the panning gain for the topmost loudspeaker of an exemplary polygon.
• the x-axis indicates the azimuth angle
• the y-axis indicates the elevation angle
• squares indicate loudspeaker vertices
• solid lines indicate gain contour
• dashed lines indicate polygon edges
• arrows indicate normal vectors.
• the linear cross-fading method can be transferred to the 3D case by defining linear cross-fading functions between the loudspeakers of a polygon.
• Fig. 24 illustrates this by a contour plot for the topmost loudspeaker which is part of a polygon defined by five loudspeakers.
• the loudspeaker directions which are the vertices of the polygon shown by the dashed line, are marked by the squares.
• the gain for the topmost loudspeaker as a function of the panning direction within the polygon is shown by means of the solid contour lines.
• the polygon In order to compute the cross-fading gain for a loudspeaker, the polygon first needs to be sub-divided into triangles specified by the loudspeaker vertex and the edges of the polygon. This sub-division is indicated in Fig. 24 by the dotted lines.
• the final panning gains are then obtained by applying the energy normalization (3) to the cross-fading gains.
• power normalization is conducted, e.g., by applying the formula
• a special feature of the used coordinate system is the existence of the poles at ⁇ 90° elevation.
• a method like the generalized VBAP approach needs to be applied to solve this issue. In doing so, an additional vertex is added at ⁇ 90° elevation and the polygons which contain the poles are split.
• the gains for the imaginary pole loudspeakers are downmixed to their physical neighbors.
• the poles are not points in the azimuth-elevation parameter space but lines, it is reasonable for the computation of the cross-fading gains to set the azimuth angle of the pole vertices to the azimuth angle of the panning direction.
• a square root of the gain e.g., a square root of the gain of formula (1 1 ) may, e.g., be applied on the samples of the audio input signal.
• Fig. 25 indicates these trajectories.
• the conducted listening test was not a MUSHRA test.
• the "Ref signal is a reference with regard to all quality features besides location accuracy.
• the test signals shall reproduce one of the trajectories. Participants were encouraged to slightly move their head within ⁇ 30° azimuth/elevation angle. The timbre, the location accuracy / smoothness of movement, the source extension / focus, and the overall quality of all test signals were judged and commented.
• Each test item contained a single object at constant velocity which was rendered with an elevation angle that was linearly interpolated between 0° and 35° and an azimuth angle that was linearly interpolated as follows: Trajectory I (front): -30° to 30°; trajectory II (back): -1 10° to 1 10°; trajectory III (front-left): 30° to 90°; and trajectory IV (front-left): 90° to 30°.
• test signals For the generation of the test signals, three kinds of mono signals were used, which were then rendered along the four trajectories, namely 1 : “Speech”; 2: “Pink Noise”; and 3: “Beat”.
• the "Speech” signal was a 6.7s long sentence from a female speaker.
• the "Pink Noise” signal contained 8s of stationary pink noise.
• the "Beat” signal also lasted 8s and contained a beat of a woodblock and a castanet struck in turn at 160bpm.
• the three input signals were manually adjusted to similar loudness.
• the two VBAP variants only differed by the triangulation i.e., the diagonal within the rectangular loudspeaker arrangement, whereas the diagonal of "vbap A" coincided with the trajectories I, II, and III.
• Fig. 26 shows the average and the 95% confidence interval of the test results for the first listening test where the timbre was rated.
• Fig. 27 illustrates a difference plot for the first listening test where the timbre was rated.
• Fig. 28 shows the test results for a second test where the location accuracy and smoothness of movement was rated.
• the corresponding difference plot shown in Fig. 29 reveals that the EFAP method results in a smoother movement / better location accuracy than VBAP.
• Fig. 30 shows the test results for the third test where the source extension and focus was rated.
• EFAP performs either equally well or slightly worse than VBAP with respect to the source extension. This observation can be explained by the fact that "vbap A" mostly resulted in pair-wise panning and thus caused a smaller perceived source extension compared to the other triangulation variant or EFAP.
• symmetric panning gains for symmetric setups by N-wise panning defined via polygons are realized.
• the listening test compared the provided concepts with VBAP, gives evidence that the proposed concepts result in a better location accuracy.
• the greater number of active loudspeakers stabilizes the position and trajectory of the phantom source, but it also produces a slightly stronger bass boost and a slightly greater source extension. While some subjects preferred the improved spatial accuracy, others put more emphasize on the timbre, resulting in a balanced overall preference.
• the proposed concepts are beneficial in applications where the location accuracy and smoothness of movement is of importance. This property is further improved by frequency-dependent exponentiation of the calculated cross-fading gains while the timbre could be compensated by means of equalization.
• inventive decomposed signal can be stored on a digital storage medium or can be transmitted on a transmission medium such as a wireless transmission medium or a wired transmission medium such as the Internet.
• embodiments of the invention can be implemented in hardware or in software.
• the implementation can be performed using a digital storage medium, for example a floppy disk, a DVD, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed.
• a digital storage medium for example a floppy disk, a DVD, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed.
• Some embodiments according to the invention comprise a non-transitory data carrier having electronically readable control signals, which are capable of cooperating with a programmable computer system, such that one of the methods described herein is performed.
• embodiments of the present invention can be implemented as a computer program product with a program code, the program code being operative for performing one of the methods when the computer program product runs on a computer,
• the program code may for example be stored on a machine readable carrier.
• inventions comprise the computer program for performing one of the methods described herein, stored on a machine readable carrier.
• an embodiment of the inventive method is, therefore, a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.
• a further embodiment of the inventive methods is, therefore, a data carrier (or a digital storage medium, or a computer-readable medium) comprising, recorded thereon, the computer program for performing one of the methods described herein.
• a further embodiment of the inventive method is, therefore, a data stream or a sequence of signals representing the computer program for performing one of the methods described herein.
• the data stream or the sequence of signals may for example be configured to be transferred via a data communication connection, for example via the Internet,
• a further embodiment comprises a processing means, for example a computer, or a programmable logic device, configured to or adapted to perform one of the methods described herein.
• a further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.
• a programmable logic device for example a field programmable gate array
• a field programmable gate array may cooperate with a microprocessor in order to perform one of the methods described herein.
• the methods are preferably performed by any hardware apparatus.
• ITU-R. Rec. ITU-R BS.775-1 Multichannel Stereophonic Sound System with and without Accompanying Picture, 2006.
• Pulkki, V Uniform Spreading of Amplitude Panned Virtual Sources. In IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, New Paltz, NY, USA, Oct. 1999.
• Pulkki, V. Karjalainen, M., and Vaiimaki, V. Localization, Coloration, and Enhancement of Amplitude-Panned Virtual Sources. In AES 16th Int. Conference, Rovaniemi, Finland, 1999,

Landscapes

• Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• Acoustics & Sound (AREA)
• Signal Processing (AREA)
• Stereophonic System (AREA)
• Circuit For Audible Band Transducer (AREA)

Priority Applications (7)

Applications Claiming Priority (2)

Related Child Applications (1)

Publications (1)

Family

ID=50771413

Family Applications (1)

Country Status (8)

Families Citing this family (7)

Citations (1)

Family Cites Families (19)

• 2014
  • 2014-10-07 RU RU2016148647A patent/RU2666248C2/ru active
  • 2014-10-07 ES ES14781537T patent/ES2833424T3/es active Active
  • 2014-10-07 JP JP2016566628A patent/JP6513703B2/ja active Active
  • 2014-10-07 CN CN201480079058.8A patent/CN106465027B/zh active Active
  • 2014-10-07 EP EP14781537.7A patent/EP3143779B1/en active Active
  • 2014-10-07 WO PCT/EP2014/071449 patent/WO2015172854A1/en active Application Filing
  • 2014-10-07 BR BR112016026283-2A patent/BR112016026283B1/pt active IP Right Grant
• 2016
  • 2016-11-09 US US15/347,493 patent/US10021499B2/en active Active

Patent Citations (1)

Non-Patent Citations (24)

Also Published As

Similar Documents

Legal Events