Method and apparatus for compressing and decompressing a higher order ambisonics representation

The method optimizes HOA compression by adaptively assigning directional and ambient components to channels based on perceptual relevance, reducing bit rates and improving decoding quality for streaming.

CA3190346CActive Publication Date: 2026-11-10DOLBY INTERNATIONAL AB
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CA3190346
Authority / Receiving Office
CA · CA
Patent Type
Patents
Current Assignee / Owner
Priority Date
2013-04-29
Filing Date
2014-04-24
Publication Date
2026-11-10
Estimated Expiration
2034-04-24

AI Technical Summary

Technical Problem

Existing Higher Order Ambisonics (HOA) compression methods are inefficient in optimizing the use of channels for perceptual coding, leading to suboptimal decoding processes and high bit rates, especially for streaming applications.

Method used

A method and apparatus that dynamically assign directional and ambient HOA components to a fixed number of channels based on perceptual relevance, using a criterion that minimizes perceptible error by considering the spatial power distribution of quantization noise and modeling errors, allowing for adaptive channel usage in frames with varying dominant sound sources.

Benefits of technology

This approach optimizes the use of channels for perceptual coding, reducing bit rates and improving the quality of decoded HOA representations by ensuring minimal perceptible error, making it suitable for streaming applications.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

Higher Order Ambisonics represents three-dimensional sound independent of a specific loudspeaker set-up. However, transmission of an HOA representation results in a very high bit rate. Therefore compression with a fixed number of channels is used, in which directional and ambient signal components are processed differently. The ambient HOA component is represented by a minimum number of HOA coefficient sequences. The remaining channels contain either directional signals or additional coefficient sequences of the ambient HOA component, depending on what will result in optimum perceptual quality. This processing can change on a frame-by-frame basis.
Need to check novelty before this filing date? Find Prior Art

Description

METHOD AND APPARATUS FOR COMPRESSING AND DECOMPRESSING A HIGHER ORDER AMBISONICS REPRESENTATION Technical field The invention relates to a method and to an apparatus for compressing and decompressing a Higher Order Ambisonics rep- resentation by processing directional and ambient signal components differently. Background Higher Order Ambisonics (HOA) offers one possibility to rep- resent three-dimensional sound among other techniques like wave field synthesis (WFS) or channel based approaches like 22.2. In contrast to channel based methods, however, the HOA representation offers the advantage of being independent of a specific loudspeaker set-up. This flexibility, however, is at the expense of a decoding process which is required for the playback of the HOA representation on a particular loud- speaker set-up. Compared to the WFS approach, where the num- ber of required loudspeakers is usually very large, HOA may also be rendered to set-ups consisting of only few loud- speakers. A further advantage of HOA is that the same repre- sentation can also be employed without any modification for binaural rendering to head-phones. HOA is based on the representation of the spatial density of complex harmonic plane wave amplitudes by a truncated Spher- ical Harmonics (SH) expansion. Each expansion coefficient is a function of angular frequency, which can be equivalently represented by a time domain function. Hence, without loss of generality, the complete HOA sound field representation actually can be assumed to consist of 0 time domain func- tions, where <com:Math><semantics>𝟎<annotation encoding="application / x-tex">\boldsymbol{0}< / annotation>< / semantics>< / com:Math> denotes the number of expansion coefficients. These time domain functions will be equivalently referred to as HOA coefficient sequences or as HOA channels. The spatial resolution of the HOA representation improves with a growing maximum order N of the expansion. Unfortu- nately, the number of expansion coefficients <com:Math><semantics>𝟎<annotation encoding="application / x-tex">\boldsymbol{0}< / annotation>< / semantics>< / com:Math> grows quad- ratically with the order <com:Math><semantics>N<annotation encoding="application / x-tex">N< / annotation>< / semantics>< / com:Math>, in particular <com:Math><semantics>O=(N+1)2<annotation encoding="application / x-tex">O = (N+1)^2< / annotation>< / semantics>< / com:Math>. For example, typical HOA representations using order <com:Math><semantics>N=4<annotation encoding="application / x-tex">N=4< / annotation>< / semantics>< / com:Math> re- quire <com:Math><semantics>O=25<annotation encoding="application / x-tex">O = 25< / annotation>< / semantics>< / com:Math> HOA (expansion) coefficients. According to the previously made considerations, the total bit rate for the transmission of HOA representation, given a desired single- channel sampling rate <com:Math><semantics>fS<annotation encoding="application / x-tex">f_{S}< / annotation>< / semantics>< / com:Math> and the number of bits <com:Math><semantics>Nb<annotation encoding="application / x-tex">N_{b}< / annotation>< / semantics>< / com:Math> per sam- ple, is determined by <com:Math><semantics>0⋅fS⋅Nb<annotation encoding="application / x-tex">0 \cdot f_{S} \cdot N_{b}< / annotation>< / semantics>< / com:Math>. Consequently, transmitting an HOA representation of order <com:Math><semantics>N=4<annotation encoding="application / x-tex">N=4< / annotation>< / semantics>< / com:Math> with a sampling rate of <com:Math><semantics>fS=48kHz<annotation encoding="application / x-tex">f_{S} = 48 \, \text{kHz}< / annotation>< / semantics>< / com:Math> employing <com:Math><semantics>Nb=16<annotation encoding="application / x-tex">N_{b} = 16< / annotation>< / semantics>< / com:Math> bits per sample results in a bit rate of 19.2 MBits / s, which is very high for many practical applications, e.g. for streaming. Compression of HOA sound field representations is proposed in patent applications EP 12306569.0 and EP 12305537.8. In- stead of perceptually coding each one of the HOA coefficient sequences individually, as it is performed e.g. in E. Hellerud, I. Burnett, A. Solvang and U.P. Svensson, "Encoding Higher Order Ambisonics with AAC", 124th AES Convention, Amsterdam, 2008, it is attempted to reduce the number of signals to be perceptually coded, in particular by performing a sound field analysis and decomposing the given HOA representation into a directional and a residual ambient component. The di- rectional component is in general supposed to be represented by a small number of dominant directional signals which can be regarded as general plane wave functions. The order of the residual ambient HOA component is reduced because it is assumed that, after the extraction of the dominant direction- al signals, the lower-order HOA coefficients are carrying the most relevant information. Summary Altogether, by such operation the initial number <com:Math><semantics>(N+1)2<annotation encoding="application / x-tex">(N+1)^2< / annotation>< / semantics>< / com:Math> of HOA coefficient sequences to be perceptually coded is reduced to a fixed number of <com:Math><semantics>D<annotation encoding="application / x-tex">D< / annotation>< / semantics>< / com:Math> dominant directional signals and a number of <com:Math><semantics>(NRED+1)2<annotation encoding="application / x-tex">(N_{RED}+1)^2< / annotation>< / semantics>< / com:Math> HOA coefficient sequences representing the residual ambient HOA component with a truncated order <com:Math><semantics>NRED<N<annotation encoding="application / x-tex">N_{RED} < N< / annotation>< / semantics>< / com:Math>, whereby the number of signals to be coded is fixed, i.e. <com:Math><semantics>D+(NRED+1)2<annotation encoding="application / x-tex">D + (N_{RED} + 1)^2< / annotation>< / semantics>< / com:Math>. In particular, this number is independent of the actually detected number <com:Math><semantics>DACT(k)≤D<annotation encoding="application / x-tex">D_{ACT}(k) \leq D< / annotation>< / semantics>< / com:Math> of active dominant directional sound sources in a time frame <com:Math><semantics>k<annotation encoding="application / x-tex">k< / annotation>< / semantics>< / com:Math>. This means that in time frames <com:Math><semantics>k<annotation encoding="application / x-tex">k< / annotation>< / semantics>< / com:Math>, where the actually detected number <com:Math><semantics>DACT(k)<annotation encoding="application / x-tex">D_{ACT}(k)< / annotation>< / semantics>< / com:Math> of active dominant directional sound sources is smaller than the maximum allowed number <com:Math><semantics>D<annotation encoding="application / x-tex">D< / annotation>< / semantics>< / com:Math> of directional signals, some or even all of the dominant directional signals to be perceptu- ally coded are zero. Ultimately, this means that these chan- nels are not used at all for capturing the relevant infor- mation of the sound field. In this context, a further possibly weak point in the EP 12306569.0 and EP 12305537.8 processings is the criterion for the determination of the amount of active dominant direction- al signals in each time frame, because it is not attempted to determine an optimal amount of active dominant directional signals with respect to the successive perceptual coding of the sound field. For instance, in EP 12305537.8 the amount of dominant sound sources is estimated using a simple power cri- terion, namely by determining the dimension of the subspace of the inter-coefficients correlation matrix belonging to the greatest eigenvalues. In EP 12306569.0 an incremental detec- tion of dominant directional sound sources is proposed, where a directional sound source is considered to be dominant if the power of the plane wave function from the respective di- rection is high enough with respect to the first directional signal. Using power based criteria like in EP 12306569.0 and EP 12305537.8 may lead to a directional-ambient decomposition which is suboptimal with respect to perceptual coding of the sound field. A problem to be solved by the invention is to improve HOA compression by determining for a current HOA audio signal content how to assign to a predetermined reduced number of channels, directional signals and coefficients for the ambi- ent HOA component. The invention improves the compression processing proposed in EP 12306569.0 in two aspects. First, the bandwidth provided by the given number of channels to be perceptually coded is better exploited. In time frames where no dominant sound source signals are detected, the channels originally reserved for the dominant directional signals are used for capturing additional information about the ambient component, in the form of additional HOA coefficient sequences of the residual ambient HOA component. Second, having in mind the goal to ex- ploit a given number of channels to perceptually code a given HOA sound field representation, the criterion for the deter- mination of the amount of directional signals to be extracted from the HOA representation is adapted with respect to that purpose. The number of directional signals is determined such that the decoded and reconstructed HOA representation pro- vides the lowest perceptible error. That criterion compares the modelling errors arising either from extracting a direc- tional signal and using a HOA coefficient sequence less for describing the residual ambient HOA component, or arising from not extracting a directional signal and instead using an additional HOA coefficient sequence for describing the resid- ual ambient HOA component. That criterion further considers for both cases the spatial power distribution of the quanti- sation noise introduced by the perceptual coding of the di- rectional signals and the HOA coefficient sequences of the residual ambient HOA component. In order to implement the above-described processing, before starting the HOA compression, a total number <com:Math><semantics>I<annotation encoding="application / x-tex">I< / annotation>< / semantics>< / com:Math> of signals (channels) is specified compared to which the original number of <com:Math><semantics>O<annotation encoding="application / x-tex">O< / annotation>< / semantics>< / com:Math> HOA coefficient sequences is reduced. The ambient HOA component is assumed to be represented by a minimum number <com:Math><semantics>OmRED<annotation encoding="application / x-tex">O_{m RED}< / annotation>< / semantics>< / com:Math> of HOA coefficient sequences. In some cases, that mini- mum number can be zero. The remaining <com:Math><semantics>D=I−ORED<annotation encoding="application / x-tex">D = I - O_{RED}< / annotation>< / semantics>< / com:Math> channels are supposed to contain either directional signals or additional coefficient sequences of the ambient HOA component, depending on what the directional signal extraction processing decides to be perceptually more meaningful. It is assumed that the assigning of either directional signals or ambient HOA compo- nent coefficient sequences to the remaining <com:Math><semantics>D<annotation encoding="application / x-tex">D< / annotation>< / semantics>< / com:Math> channels can change on frame-by-frame basis. For reconstruction of the sound field at receiver side, information about the assign- ment is transmitted as extra side information. In accordance with a general aspect, a method for decompressing a compressed Higher Order Ambisonics (HOA) representation is provided, the method comprising: decoding the compressed HOA representation to provide a decoded frame of signals; determining a first set of indices indicating active di- rectional signals of the decoded frame of signals, and re- spective directions of the active directional signals; re-distributing the decoded frame of signals based on the first set of indices and the respective directions, wherein the re-distribution determines a frame of HOA direc- tional signals; and re-composing a current decompressed frame of the HOA representation from the frame of HOA directional signals. In accordance with another general aspect, an apparatus for decompressing a Higher Order Ambisonics (HOA) representation is provided, the apparatus comprising: a decoder for decoding the compressed HOA representation to provide a decoded frame of signals; a first processor for determining a first set of indices indicating active directional signals of the decoded frame of signals, and respective directions of the active directional signals; a second processor for re-distributing the decoded frame of channels based on the first set of indices and the respective directions, wherein the re-distribution determines a frame of HOA directional signals; and a third processor for re-composing a current decom- pressed frame of the HOA representation from the frame of HOA directional signals. In principle, the inventive compression method is suited for compressing using a fixed number of perceptual encodings a Higher Order Ambisonics representation of a sound field, de- noted HOA, with input time frames of HOA coefficient se- quences, said method including the following steps which are carried out on a frame-by-frame basis: - for a current frame, estimating a set of dominant direc- tions and a corresponding data set of indices of detected directional signals; - decomposing the HOA coefficient sequences of said current frame into a non-fixed number of directional signals with respective directions contained in said set of dominant di- rection estimates and with a respective data set of indices of said directional signals, wherein said non-fixed number is smaller than said fixed number, and into a residual ambient HOA component that is represent- ed by a reduced number of HOA coefficient sequences and a corresponding data set of indices of said reduced number of residual ambient HOA coefficient sequences, which reduced number corresponds to the difference between said fixed num- ber and said non-fixed number; - assigning said directional signals and the HOA coeffi- cient sequences of said residual ambient HOA component to channels the number of which corresponds to said fixed num- ber, wherein for said assigning said data set of indices of said directional signals and said data set of indices of said reduced number of residual ambient HOA coefficient se- quences are used; - perceptually encoding said channels of the related frame so as to provide an encoded compressed frame. In principle the inventive compression apparatus is suited for compressing using a fixed number of perceptual encodings a Higher Order Ambisonics representation of a sound field, denoted HOA, with input time frames of HOA coefficient se- quences, said apparatus carrying out a frame-by-frame based processing and including: - means being adapted for estimating for a current frame a set of dominant directions and a corresponding data set of indices of detected directional signals; - means being adapted for decomposing the HOA coefficient sequences of said current frame into a non-fixed number of directional signals with respective directions contained in said set of dominant direction estimates and with a respec- tive data set of indices of said directional signals, where- in said non-fixed number is smaller than said fixed number, and into a residual ambient HOA component that is represent- ed by a reduced number of HOA coefficient sequences and a corresponding data set of indices of said reduced number of residual ambient HOA coefficient sequences, which reduced number corresponds to the difference between said fixed num- ber and said non-fixed number; - means being adapted for assigning said directional sig- nals and the HOA coefficient sequences of said residual am- bient HOA component to channels the number of which corre- sponds to said fixed number, wherein for said assigning said data set of indices of said directional signals and said da- ta set of indices of said reduced number of residual ambient HOA coefficient sequences are used; means being adapted for perceptually encoding said chan- nels of the related frame so as to provide an encoded com- pressed frame. In principle, the inventive decompression method is suited for decompressing a Higher Order Ambisonics representation compressed according to the above compression method, said decompressing including the steps: - perceptually decoding a current encoded compressed frame so as to provide a perceptually decoded frame of channels; - re-distributing said perceptually decoded frame of chan- nels, using said data set of indices of detected directional signals and said data set of indices of the chosen ambient 8 HOA coefficient sequences, so as to recreate the correspond- ing frame of directional signals and the corresponding frame of the residual ambient HOA component; - re-composing a current decompressed frame of the HOA rep- resentation from said frame of directional signals and from said frame of the residual ambient HOA component, using said data set of indices of detected directional signals and said set of dominant direction estimates, wherein directional signals with respect to uniformly dis- tributed directions are predicted from said directional sig- nals, and thereafter said current decompressed frame is re- composed from said frame of directional signals, said pre- dicted signals and said residual ambient HOA component. In principle the inventive decompression apparatus is suited for decompressing a Higher Order Ambisonics representation compressed according to the above compression method, said apparatus including: - means being adapted for perceptually decoding a current en- coded compressed frame so as to provide a perceptually de- coded frame of channels; - means being adapted for re-distributing said perceptually decoded frame of channels, using said data set of indices of detected directional signals and said data set of indices of the chosen ambient HOA coefficient sequences, so as to rec- reate the corresponding frame of directional signals and the corresponding frame of the residual ambient HOA component; - means being adapted for re-composing a current decom- pressed frame of the HOA representation from said frame of directional signals, said frame of the residual ambient HOA component, said data set of indices of detected directional signals, and said set of dominant direction estimates, wherein directional signals with respect to uniformly dis- tributed directions are predicted from said directional sig- . 9 nals, and thereafter said current decompressed frame is re- composed from said frame of directional signals, said pre- dicted signals and said residual ambient HOA component. 5 Advantageous additional embodiments of the invention are dis- closed herein. Brief description of drawings Exemplary embodiments of the invention are described with reference to the accompanying drawings, which show in: Fig. 1 block diagram for the HOA compression; Fig. 2 estimation of dominant sound source directions; Fig. 3 block diagram for the HOA decompression; Fig. 4 spherical coordinate system; Fig. 5 normalised dispersion function <com:Math><semantics>vN(θ)<annotation encoding="application / x-tex">v_N(\theta)< / annotation>< / semantics>< / com:Math> for different Am- bisonics orders N and for angles <com:Math><semantics>θ∈[0,π]<annotation encoding="application / x-tex">\theta \in [0,\pi]< / annotation>< / semantics>< / com:Math>. Description of embodiments A. Improved HOA compression The compression processing according to the invention, which is based on EP 12306569.0, is illustrated in Fig. 1 where the signal processing blocks that have been modified or newly in- troduced compared to EP 12306569.0 are presented with a bold box, and where '<com:Math><semantics>𝒢<annotation encoding="application / x-tex">\mathcal{G}< / annotation>< / semantics>< / com:Math>' (direction estimates as such) and '<com:Math><semantics>𝒞<annotation encoding="application / x-tex">\boldsymbol{\mathcal{C}}< / annotation>< / semantics>< / com:Math>' in this application correspond to <com:Math><semantics>′A′<annotation encoding="application / x-tex">'A'< / annotation>< / semantics>< / com:Math> (matrix of direction esti- mates) and <com:Math><semantics>′D′<annotation encoding="application / x-tex">'D'< / annotation>< / semantics>< / com:Math> in EP 12306569.0, respectively. For the HOA compression a frame-wise processing with non- overlapping input frames <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\boldsymbol{\mathcal{C}}(k)$ of HOA coefficient sequences of length <com:Math><semantics>L<annotation encoding="application / x-tex">L< / annotation>< / semantics>< / com:Math> is used, where <com:Math><semantics>k<annotation encoding="application / x-tex">k< / annotation>< / semantics>< / com:Math> denotes the frame index. The frames are defined with respect to the HOA coefficient sequences specified in equation (45) as [Image disponible dans le document PDF, Image available in the PDF document] (1) where <com:Math><semantics>TS<annotation encoding="application / x-tex">T_{S}< / annotation>< / semantics>< / com:Math> indicates the sampling period. The first step or stage 11 / 12 in Fig. 1 is optional and con- sists of concatenating the non-overlapping <com:Math><semantics>k<annotation encoding="application / x-tex">k< / annotation>< / semantics>< / com:Math>-th and the <com:Math><semantics>(k−1)<annotation encoding="application / x-tex">(k-1)< / annotation>< / semantics>< / com:Math>-th frames of HOA coefficient sequences into a long frame <com:Math><semantics>C̃(k)<annotation encoding="application / x-tex">\widetilde{C}(k)< / annotation>< / semantics>< / com:Math> as [Image disponible dans le document PDF, Image available in the PDF document] (2) which long frame is 50% overlapped with an adjacent long frame and which long frame is successively used for the es- timation of dominant sound source directions. Similar to the notation for <com:Math><semantics>𝑪̃(k)<annotation encoding="application / x-tex">\tilde{\boldsymbol{C}}(k)< / annotation>< / semantics>< / com:Math>, the tilde symbol is used in the following description for indicating that the respective quantity re- fers to long overlapping frames. If step / stage 11 / 12 is not present, the tilde symbol has no specific meaning. In principle, the estimation step or stage 13 of dominant sound sources is carried out as proposed in EP 13305156.5, but with an important modification. The modification is re- lated to the determination of the amount of directions to be detected, i.e. how many directional signals are supposed to be extracted from the HOA representation. This is accom- plished with the motivation to extract directional signals only if it is perceptually more relevant than using instead additional HOA coefficient sequences for better approxima- tion of the ambient HOA component. A detailed description of this technique is given in section A.2. The estimation provides a data set <com:Math><semantics>ĨDIR,ACT(k)⊆{1,...,D}<annotation encoding="application / x-tex">\tilde{I}_{\text{DIR,ACT}}(k) \subseteq \{1,...,D\}< / annotation>< / semantics>< / com:Math> of indi- ces of directional signals that have been detected as well as the set <com:Math><semantics>𝒢̂Ω,ACT(k)<annotation encoding="application / x-tex">\hat{\mathcal{G}}_{\Omega,\mathrm{ACT}}(k)< / annotation>< / semantics>< / com:Math> of corresponding direction estimates. <com:Math><semantics>D<annotation encoding="application / x-tex">D< / annotation>< / semantics>< / com:Math> denotes the maximum number of directional signals that has to be set before starting the HOA compression. In step or stage 14, the current (long) frame <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\widetilde{\bm{\mathcal{C}}}(k)$ of HOA co- efficient sequences is decomposed (as proposed in EP 11 13305156.5) into a number of directional signals <com:Math><semantics>XDIR(k−2)<annotation encoding="application / x-tex">X_{\text{DIR}}(k-2)< / annotation>< / semantics>< / com:Math> belonging to the directions contained in the set <com:Math><semantics>𝒢𝛀,ACT(k)<annotation encoding="application / x-tex">\mathcal{G}_{\mathbf{\Omega},\mathrm{ACT}}(k)< / annotation>< / semantics>< / com:Math>, and a residual ambient HOA component <com:Math><semantics>CAMB(k−2)<annotation encoding="application / x-tex">C_{AMB}(k-2)< / annotation>< / semantics>< / com:Math>. The delay of two frames is introduced as a result of overlap-add pro- cessing in order to obtain smooth signals. It is assumed that <com:Math><semantics>XDIR(k−2)<annotation encoding="application / x-tex">X_{\mathrm{DIR}}(k-2)< / annotation>< / semantics>< / com:Math> is containing a total of <com:Math><semantics>D<annotation encoding="application / x-tex">D< / annotation>< / semantics>< / com:Math> channels, of which however only those corresponding to the active directional signals are non-zero. The indices specifying these channels are assumed to be output in the data set <com:Math><semantics>ℐDIR,ACT(k−2)<annotation encoding="application / x-tex">\mathcal{I}_{\mathrm{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math>. Addi- tionally, the decomposition in step / stage 14 provides some parameters <com:Math><semantics>ζ(k−2)<annotation encoding="application / x-tex">\zeta(k-2)< / annotation>< / semantics>< / com:Math> which are used at decompression side for predicting portions of the original HOA representation from the directional signals (see EP 13305156.5 for more details). In step or stage 15, the number of coefficients of the ambi- ent HOA component <com:Math><semantics>CAMB(k−2)<annotation encoding="application / x-tex">C_{AMB}(k-2)< / annotation>< / semantics>< / com:Math> is intelligently reduced to con- tain only <com:Math><semantics>ORED+D−NDIR,ACT(k−2)<annotation encoding="application / x-tex">O_{\text{RED}} + D - N_{\text{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math> non-zero HOA coefficient se- quences, where <com:Math><semantics>NDIR,ACT(k−2)=|ℐDIR,ACT(k−2)|<annotation encoding="application / x-tex">N_{\mathrm{DIR,ACT}}(k-2) = \left| \mathcal{I}_{\mathrm{DIR,ACT}}(k-2) \right|< / annotation>< / semantics>< / com:Math> indicates the car- dinality of the data set <com:Math><semantics>ℐDIR,ACT(k−2)<annotation encoding="application / x-tex">\mathcal{I}_{\text{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math>, i.e. the number of ac- tive directional signals in frame <com:Math><semantics>k−2<annotation encoding="application / x-tex">k-2< / annotation>< / semantics>< / com:Math>. Since the ambient HOA component is assumed to be always represented by a mini- mum number <com:Math><semantics>O𝖱𝖤𝖣<annotation encoding="application / x-tex">O_{\mathsf{RED}}< / annotation>< / semantics>< / com:Math> of HOA coefficient sequences, this problem can be actually reduced to the selection of the remaining <com:Math><semantics>D−NDIR,ACT(k−2)<annotation encoding="application / x-tex">D-N_{\mathrm{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math> HOA coefficient sequences out of the possible <com:Math><semantics>O−OmRED<annotation encoding="application / x-tex">O-O_{m RED}< / annotation>< / semantics>< / com:Math> ones. In order to obtain a smooth reduced ambient HOA representation, this choice is accomplished such that, compared to the choice taken at the previous frame <com:Math><semantics>k−3<annotation encoding="application / x-tex">k-3< / annotation>< / semantics>< / com:Math>, as few changes as possible will occur. In particular, the three following cases are to be differen- tiated: a) <com:Math><semantics>NDIR,ACT(k−2)=NDIR,ACT(k−3)<annotation encoding="application / x-tex">N_{\text{DIR,ACT}}(k-2) = N_{\text{DIR,ACT}}(k-3)< / annotation>< / semantics>< / com:Math>: In this case the same HOA coef- ficient sequences are assumed to be selected as in frame <com:Math><semantics>k−3<annotation encoding="application / x-tex">k-3< / annotation>< / semantics>< / com:Math>. b) <com:Math><semantics>NDIR,ACT(k−2)<NDIR,ACT(k−3)<annotation encoding="application / x-tex">N_{\mathrm{DIR,ACT}}(k-2) < N_{\mathrm{DIR,ACT}}(k-3)< / annotation>< / semantics>< / com:Math>: In this case, more HOA coeffi- 12 cient sequences than in the last frame <com:Math><semantics>k−3<annotation encoding="application / x-tex">k-3< / annotation>< / semantics>< / com:Math> can be used for representing the ambient HOA component in the current frame. Those HOA coefficient sequences that were selected in <com:Math><semantics>k−3<annotation encoding="application / x-tex">k-3< / annotation>< / semantics>< / com:Math> are assumed to be also selected in the current frame. The additional HOA coefficient sequences can be selected according to different criteria. For instance, selecting those HOA coefficient sequences in <com:Math><semantics>CAMR(k−2)<annotation encoding="application / x-tex">C_{AMR}(k-2)< / annotation>< / semantics>< / com:Math> with the highest average power, or selecting the HOA co- efficients sequences with respect to their perceptual significance. c) <com:Math><semantics>NDIR,ACT(k−2)>NDIR,ACT(k−3)<annotation encoding="application / x-tex">N_{\text{DIR,ACT}}(k-2) > N_{\text{DIR,ACT}}(k-3)< / annotation>< / semantics>< / com:Math>: In this case, less HOA coeffi- cient sequences than in the last frame <com:Math><semantics>k−3<annotation encoding="application / x-tex">k-3< / annotation>< / semantics>< / com:Math> can be used for representing the ambient HOA component in the current frame. The question to be answered here is which of the previously selected HOA coefficient sequences have to be deactivated. A reasonable solution is to deactivate those sequences which were assigned to the channels <com:Math><semantics>i∈ℐDIR,ACT(k−2)<annotation encoding="application / x-tex">i \in \mathcal{I}_{DIR,ACT}(k-2)< / annotation>< / semantics>< / com:Math> at the signal assigning step or stage 16 at frame <com:Math><semantics>k−3<annotation encoding="application / x-tex">k-3< / annotation>< / semantics>< / com:Math>. For avoiding discontinuities at frame borders when addition- al HOA coefficient sequences are activated or deactivated, it is advantageous to smoothly fade in or out the respective signals. The final ambient HOA representation with the reduced number of <com:Math><semantics>ORED+NDIRACT(k−2)<annotation encoding="application / x-tex">O_{\text{RED}} + N_{\text{DIRACT}}(k-2)< / annotation>< / semantics>< / com:Math> non-zero coefficient sequences is de- noted by <com:Math><semantics>CAMB,RED(k−2)<annotation encoding="application / x-tex">C_{\mathrm{AMB,RED}}(k-2)< / annotation>< / semantics>< / com:Math>. The indices of the chosen ambient HOA coefficient sequences are output in the data set <com:Math><semantics>ℐAMB.ACT(k−<annotation encoding="application / x-tex">\mathcal{I}_{AMB.ACT}(k-< / annotation>< / semantics>< / com:Math> 2). In step / stage 16, the active directional signals contained in <com:Math><semantics>XDIR(k−2)<annotation encoding="application / x-tex">X_{DIR}(k-2)< / annotation>< / semantics>< / com:Math> and the HOA coefficient sequences contained in <com:Math><semantics>CAMB,RED(k−2)<annotation encoding="application / x-tex">C_{\mathrm{AMB,RED}}(k-2)< / annotation>< / semantics>< / com:Math> are assigned to the frame <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math> of <com:Math><semantics>I<annotation encoding="application / x-tex">I< / annotation>< / semantics>< / com:Math> channels for individual perceptual encoding. To describe the signal assignment in more detail, the frames <com:Math><semantics>XDIR(k−2)<annotation encoding="application / x-tex">X_{\mathrm{DIR}}(k-2)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>CAMB,RED(k−2)<annotation encoding="application / x-tex">C_{\mathrm{AMB,RED}}(k-2)< / annotation>< / semantics>< / com:Math> are assumed to consist of the individual sig- nals <com:Math><semantics>xDIR,d(k−2)<annotation encoding="application / x-tex">x_{\text{DIR},d}(k-2)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>d∈{1,...,D}<annotation encoding="application / x-tex">d \in \{1,...,D\}< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>yi(k−2)<annotation encoding="application / x-tex">y_i(k-2)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>i∈{1,...,I}<annotation encoding="application / x-tex">i \in \{1,...,I\}< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>cAMB.RED,o(k−1)<annotation encoding="application / x-tex">c_{\text{AMB.RED},o}(k-1)< / annotation>< / semantics>< / com:Math> 2), <com:Math><semantics>o∈{1,...,0}<annotation encoding="application / x-tex">o \in \{1, ..., 0\}< / annotation>< / semantics>< / com:Math> as follows: [Image disponible dans le document PDF, Image available in the PDF document] [Image disponible dans le document PDF, Image available in the PDF document] (3) The active directional signals are assigned such that they keep their channel indices in order to obtain continuous signals for the successive perceptual coding. This can be expressed by [Image disponible dans le document PDF, Image available in the PDF document] (4) The HOA coefficient sequences of the ambient component are assigned such the minimum number of <com:Math><semantics>O𝖱𝖤𝖣<annotation encoding="application / x-tex">O_{\mathsf{RED}}< / annotation>< / semantics>< / com:Math> coefficient sequenc- es is always contained in the last <com:Math><semantics>ORED<annotation encoding="application / x-tex">O_{\text{RED}}< / annotation>< / semantics>< / com:Math> signals of <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math>, i.e. [Image disponible dans le document PDF, Image available in the PDF document] (5) For the additional <com:Math><semantics>D−NDIR,ACT(k−2)<annotation encoding="application / x-tex">D-N_{\mathrm{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math> HOA coefficient sequences of the ambient component it is to be differentiated whether or not they were also selected in the previous frame: a) If they were also selected to be transmitted in the pre- vious frame, i.e. if the respective indices are also con- tained in data set <com:Math><semantics>IAMB,ACT(k−3)<annotation encoding="application / x-tex">I_{AMB,ACT}(k-3)< / annotation>< / semantics>< / com:Math>, the assignment of these coefficient sequences to the signals in <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math> is the same as for the previous frame. This operation assures smooth signals <com:Math><semantics>yi(k−2)<annotation encoding="application / x-tex">y_i(k-2)< / annotation>< / semantics>< / com:Math>, which is favourable for the suc- cessive perceptual coding in step or stage 17. b) Otherwise, if some coefficient sequences are newly se- 14 lected, i.e. if their indices are contained in data set <com:Math><semantics>JAMB,ACT(k−2)<annotation encoding="application / x-tex">J_{AMB,ACT}(k-2)< / annotation>< / semantics>< / com:Math> but not in data set <com:Math><semantics>JAMB,ACT(k−3)<annotation encoding="application / x-tex">J_{AMB,ACT}(k-3)< / annotation>< / semantics>< / com:Math>, they are first arranged with respect to their indices in an as- cending order and are in this order assigned to channels <com:Math><semantics>i∉ℐDIR.ACT(k−2)<annotation encoding="application / x-tex">i \notin \mathcal{I}_{\text{DIR.ACT}}(k-2)< / annotation>< / semantics>< / com:Math> of <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math> which are not yet occupied by di- rectional signals. This specific assignment offers the advantage that, dur- ing a HOA decompression process, the signal re-distri- bution and composition can be performed without the knowledge about which ambient HOA coefficient sequence is contained in which channel of <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math>. Instead, the as- signment can be reconstructed during HOA decompression with the mere knowledge of the data sets <com:Math><semantics>IAMB,ACT(k−2)<annotation encoding="application / x-tex">I_{AMB,ACT}(k-2)< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>J̃DIR.ACT(k)<annotation encoding="application / x-tex">\tilde{J}_{\text{DIR.ACT}}(k)< / annotation>< / semantics>< / com:Math>. Advantageously, this assigning operation also provides the assignment vector <com:Math><semantics>γ(k)∈ℝD−NDIR,ACT(k−2)<annotation encoding="application / x-tex">\gamma(k) \in \mathbb{R}^{D-N_{\text{DIR,ACT}}(k-2)}< / annotation>< / semantics>< / com:Math>, whose elements <com:Math><semantics>γo(k)<annotation encoding="application / x-tex">\gamma_o(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>o=1,...,D−NDIR,ACT(k−2)<annotation encoding="application / x-tex">o=1,...,D-N_{\mathrm{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math>, denote the indices of each one of the additional <com:Math><semantics>D−NDIR,ACT(k−2)<annotation encoding="application / x-tex">D-N_{\mathrm{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math> HOA coefficient sequences of the ambient component. To say it differently, the elements of the assignment vector <com:Math><semantics>γ(k)<annotation encoding="application / x-tex">\gamma(k)< / annotation>< / semantics>< / com:Math> provide information about which of the additional <com:Math><semantics>O−OmRED<annotation encoding="application / x-tex">O-O_{m RED}< / annotation>< / semantics>< / com:Math> HOA coefficient sequences of the am- bient HOA component are assigned into the <com:Math><semantics>D−NDIR,ACT(k−2)<annotation encoding="application / x-tex">D-N_{\mathrm{DIR,ACT}}(k-2)< / annotation>< / semantics>< / com:Math> channels with inactive directional signals. This vector can be transmitted additionally, but less frequently than by the frame rate, in order to allow for an initialisation of the re-distribution procedure performed for the HOA decompres- sion (see section B). Perceptual coding step / stage 17 en- codes the I channels of frame <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math> and outputs an encoded frame <com:Math><semantics>Y¯(k−2)<annotation encoding="application / x-tex">\overline{Y}(k-2)< / annotation>< / semantics>< / com:Math>. For frames for which vector <com:Math><semantics>γ(k)<annotation encoding="application / x-tex">\gamma(k)< / annotation>< / semantics>< / com:Math> is not transmitted from step / stage 16, at decompression side the data parameter sets <com:Math><semantics>ℐDIR,ACT(k)<annotation encoding="application / x-tex">\mathcal{I}_{\text{DIR,ACT}}(k)< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>ℐAMB,ACT(k−2)<annotation encoding="application / x-tex">\mathcal{I}_{\text{AMB,ACT}}(k-2)< / annotation>< / semantics>< / com:Math> instead of vector <com:Math><semantics>γ(k)<annotation encoding="application / x-tex">\gamma(k)< / annotation>< / semantics>< / com:Math> are used for 15 the performing the re-distribution. A.1 Estimation of the dominant sound source directions The estimation step / stage 13 for dominant sound source di- rections of Fig. 1 is depicted in Fig. 2 in more detail. It is essentially performed according to that of EP 13305156.5, but with a decisive difference, which is the way of deter- mining the amount of dominant sound sources, corresponding to the number of directional signals to be extracted from the given HOA representation. This number is significant because it is used for controlling whether the given HOA representa- tion is better represented either by using more directional signals or instead by using more HOA coefficient sequences to better model the ambient HOA component. The dominant sound source directions estimation starts in step or stage 21 with a preliminary search for the dominant sound source directions, using the long frame <com:Math><semantics>𝒄̃(k)<annotation encoding="application / x-tex">\tilde{\boldsymbol{c}}(k)< / annotation>< / semantics>< / com:Math> of input HOA coefficient sequences. Along with the preliminary direc- tion estimates <com:Math><semantics>𝛀̃DOM(d)(k)<annotation encoding="application / x-tex">\tilde{\mathbf{\Omega}}_{\mathrm{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>1≤d≤D<annotation encoding="application / x-tex">1 \leq d \leq D< / annotation>< / semantics>< / com:Math>, the corresponding direc- tional signals <com:Math><semantics>x̃DOM(d)(k)<annotation encoding="application / x-tex">\widetilde{x}_{\mathrm{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math> and the HOA sound field components <span class="math inline">$\widetilde{\bm{C}}_{ ext{DOM.CORR}}^{(d)}(k)$, which are supposed to be created by the individ- ual sound sources, are computed as described in EP 13305156.5. In step or stage 22, these quantities are used together with the frame <com:Math><semantics>C̃(k)<annotation encoding="application / x-tex">\tilde{C}(k)< / annotation>< / semantics>< / com:Math> of input HOA coefficient sequences for deter- mining the number <com:Math><semantics>D̃(k)<annotation encoding="application / x-tex">\widetilde{D}(k)< / annotation>< / semantics>< / com:Math> of directional signals to be extract- ed. Consequently, the direction estimates <com:Math><semantics>Ω̃DOM(d)(k)<annotation encoding="application / x-tex">\widetilde{\Omega}_{\mathrm{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>D̃(k)<d≤D<annotation encoding="application / x-tex">\widetilde{D}(k) < d \leq D< / annotation>< / semantics>< / com:Math>, the corresponding directional signals <com:Math><semantics>x̃DOM(d)(k)<annotation encoding="application / x-tex">\widetilde{x}_{\text{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, and HOA sound field components <span class="math inline">$\widetilde{\bm{\mathcal{C}}}_{ ext{DOM.CORR}}^{(d)}(k)$ are discarded. Instead, only the direction estimates <com:Math><semantics>𝛺̃DOM(d)(k)<annotation encoding="application / x-tex">\widetilde{\pmb{\varOmega}}_{\text{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>1≤d≤D̃(k)<annotation encoding="application / x-tex">1 \leq d \leq \widetilde{D}(k)< / annotation>< / semantics>< / com:Math> are then assigned to previously found sound sources. In step or stage 23, the resulting direction trajectories are smoothed according to a sound source movement model and it is determined which ones of the sound sources are sup- posed to be active (see EP 13305156.5). The last operation provides the set <com:Math><semantics>ℐ̃DIR,𝑨𝑪𝑻(k)<annotation encoding="application / x-tex">\tilde{\mathcal{I}}_{\text{DIR},\mathbf{ACT}}(k)< / annotation>< / semantics>< / com:Math> of indices of active directional sound sources and the set <com:Math><semantics>𝒢̃Ω,𝖠𝖢𝖳(k)<annotation encoding="application / x-tex">\tilde{\mathcal{G}}_{\Omega,\mathsf{ACT}}(k)< / annotation>< / semantics>< / com:Math> of the corresponding di- rection estimates. A.2 Determination of number of extracted directional signals For determining the number of directional signals in step / stage 22, the situation is assumed that there is a giv- en total amount of l channels which are to be exploited for capturing the perceptually most relevant sound field infor- mation. Therefore the number of directional signals to be extracted is determined, motivated by the question whether for the overall HOA compression / decompression quality the current HOA representation is represented better by using either more directional signals, or more HOA coefficient se- quences for a better modelling of the ambient HOA component. To derive in step / stage 22 a criterion for the determination of the number of directional sound sources to be extracted, which criterion is related to the human perception, it is taken into consideration that HOA compression is achieved in particular by the following two operations: - reduction of HOA coefficient sequences for representing the ambient HOA component (which means reduction of the number of related channels); perceptual encoding of the directional signals and of the HOA coefficient sequences for representing the ambient HOA component. Depending on the number <com:Math><semantics>M<annotation encoding="application / x-tex">M< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>0≤M≤D<annotation encoding="application / x-tex">0 \le M \le D< / annotation>< / semantics>< / com:Math>, of extracted direction- al signals, the first operation results in the approximation [Image disponible dans le document PDF, Image available in the PDF document] (6) [Image disponible dans le document PDF, Image available in the PDF document] (7) where <com:Math><semantics>C̃DIR(M)(k):=∑d=1MC̃DOM,CORR(d)(k)<annotation encoding="application / x-tex">\widetilde{C}_{\text{DIR}}^{(M)}(k) := \sum_{d=1}^{M} \widetilde{C}_{\text{DOM,CORR}}^{(d)}(k)< / annotation>< / semantics>< / com:Math> (8) denotes the HOA representation of the directional component consisting of the HOA sound field components <com:Math><semantics>C̃DOM,CORR(d)(k)<annotation encoding="application / x-tex">\widetilde{C}_{\mathrm{DOM,CORR}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>1≤d≤M<annotation encoding="application / x-tex">1 \le d \le M< / annotation>< / semantics>< / com:Math>, supposed to be created by the <com:Math><semantics>M<annotation encoding="application / x-tex">M< / annotation>< / semantics>< / com:Math> individually con- sidered sound sources, and <com:Math><semantics>𝑪̃AMB,RED(M)(k)<annotation encoding="application / x-tex">\widetilde{\pmb{C}}_{\mathrm{AMB,RED}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> denotes the HOA repre- sentation of the ambient component with only <com:Math><semantics>I−M<annotation encoding="application / x-tex">I-M< / annotation>< / semantics>< / com:Math> non-zero HOA coefficient sequences. The approximation from the second operation can be expressed by [Image disponible dans le document PDF, Image available in the PDF document] (9) [Image disponible dans le document PDF, Image available in the PDF document] (10) where <span class="math inline">$\widehat{\bm{c}}_{\mathrm{DIR}}^{(M)}(k)$ and <span class="math inline">$\widehat{\bm{c}}_{\mathrm{AMB.RED}}^{(M)}(k)$ denote the composed directional and ambient HOA components after perceptual decoding, re- spectively. Formulation of criterion The number <com:Math><semantics>D(k)<annotation encoding="application / x-tex">D(k)< / annotation>< / semantics>< / com:Math> of directional signals to be extracted is chosen such that the total approximation error [Image disponible dans le document PDF, Image available in the PDF document] (11) with <com:Math><semantics>M=D̃(k)<annotation encoding="application / x-tex">M = \widetilde{D}(k)< / annotation>< / semantics>< / com:Math> is as less significant as possible with respect to the human perception. To assure this, the directional power distribution of the total error for individual Bark scale critical bands is considered at a predefined number <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\boldsymbol{Q}$ of test directions <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_q< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, which are nearly uniformly distributed on the unit sphere. To be more specific, the di- rectional power distribution for the <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>-th critical band, <com:Math><semantics>b=1,...,B<annotation encoding="application / x-tex">b = 1, ..., B< / annotation>< / semantics>< / com:Math>, is represented by the vector [Image disponible dans le document PDF, Image available in the PDF document] (12) whose components <com:Math><semantics>ildeP̂a(M)(k,b)<annotation encoding="application / x-tex">\hat{ ilde{P}}_{a}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> denote the power of the total error <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\widehat{\bm{E}}^{(M)}(k)$ related to the direction <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\bm{\Omega}_q$, the <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>-th Bark scale crit- ical band and the <com:Math><semantics>k<annotation encoding="application / x-tex">k< / annotation>< / semantics>< / com:Math>-th frame. The directional power distri- bution <span class="math inline">$\widehat{\widetilde{\bm{P}}}^{(M)}(k,b)$ of the total error <span class="math inline">$\widehat{\widetilde{\bm{E}}}^{(M)}(k)$ is compared with the directional perceptual masking power distribution [Image disponible dans le document PDF, Image available in the PDF document] (13) due to the original HOA representation <com:Math><semantics>C̃(k)<annotation encoding="application / x-tex">\widetilde{C}(k)< / annotation>< / semantics>< / com:Math>. Next, for each test direction <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_q< / annotation>< / semantics>< / com:Math> and critical band <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math> the level of percep- tion <com:Math><semantics>ℒ̃0(M)(k,b)<annotation encoding="application / x-tex">\tilde{\mathcal{L}}_{0}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> of the total error is computed. It is here es- sentially defined as the ratio of the directional power of the total error <com:Math><semantics>Ẽ̂(M)(k)<annotation encoding="application / x-tex">\widehat{\tilde{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> and the directional masking power ac- cording to [Image disponible dans le document PDF, Image available in the PDF document] (14) The subtraction of '1' and the successive maximum operation is performed to ensure that the perception level is zero, as long as the error power is below the masking threshold. Finally, the number <com:Math><semantics>D̃(k)<annotation encoding="application / x-tex">\widetilde{D}(k)< / annotation>< / semantics>< / com:Math> of directionals signals to be ex- tracted can be chosen to minimise the average over all test directions of the maximum of the error perception level over all critical bands, i.e., [Image disponible dans le document PDF, Image available in the PDF document] (15) It is noted that, alternatively, it is possible to replace the maximum by an averaging operation in equation (15). Computation of the directional perceptual masking power dis- tribution For the computation of the directional perceptual masking power distribution <span class="math inline">$\widetilde{\bm{\mathcal{P}}}_{\mathsf{MASK}}(k,b)$ due to the original HOA repre- sentation <com:Math><semantics>C̃(k)<annotation encoding="application / x-tex">\tilde{C}(k)< / annotation>< / semantics>< / com:Math>, the latter is transformed to the spatial do- main in order to be represented by general plane waves <com:Math><semantics>ṽq(k)<annotation encoding="application / x-tex">\widetilde{v}_{q}(k)< / annotation>< / semantics>< / com:Math> impinging from the test directions <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_q< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>. When ar- ranging the general plane wave signals <com:Math><semantics>ṽq(k)<annotation encoding="application / x-tex">\widetilde{v}_q(k)< / annotation>< / semantics>< / com:Math> in the matrix <com:Math><semantics>Ṽ(k)<annotation encoding="application / x-tex">\widetilde{V}(k)< / annotation>< / semantics>< / com:Math> as [Image disponible dans le document PDF, Image available in the PDF document] (16) the transformation to the spatial domain is expressed by the operation <com:Math><semantics>Ṽ(k)=ℰTC̃(k)<annotation encoding="application / x-tex">\widetilde{V}(k) = \mathbf{\mathcal{E}}^T \widetilde{C}(k)< / annotation>< / semantics>< / com:Math>, (17) where 3 denotes the mode matrix with respect to the test di- rection <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_q< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, defined by [Image disponible dans le document PDF, Image available in the PDF document] (18) with <com:Math><semantics>Sq≔<annotation encoding="application / x-tex">S_q \coloneqq< / annotation>< / semantics>< / com:Math> [Image disponible dans le document PDF, Image available in the PDF document] . (19) The elements <com:Math><semantics>𝒫̂MASK(k,b)<annotation encoding="application / x-tex">\hat{\mathcal{P}}_{\mathrm{MASK}}(k,b)< / annotation>< / semantics>< / com:Math> of the directional perceptual masking power distribution <com:Math><semantics>𝒫̃𝖬𝖠𝖲𝖪(k,b)<annotation encoding="application / x-tex">\widetilde{\mathcal{P}}_{\mathsf{MASK}}(k,b)< / annotation>< / semantics>< / com:Math>, due to the original HOA repre- sentation <com:Math><semantics>C(k)<annotation encoding="application / x-tex">C(k)< / annotation>< / semantics>< / com:Math>, are corresponding to the masking powers of the general plane wave functions <com:Math><semantics>𝒗̃q(k)<annotation encoding="application / x-tex">\widetilde{\boldsymbol{v}}_{q}(k)< / annotation>< / semantics>< / com:Math> for individual criti- cal bands <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>. Computation of directional power distribution In the following two alternatives for the computation of the directional power distribution <com:Math><semantics>𝒫̃̂(M)(k,b)<annotation encoding="application / x-tex">\widehat{\tilde{\mathcal{P}}}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> are presented: a. One possibility is to actually compute the approximation <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\widehat{\bm{C}}^{(M)}(k)$ of the desired HOA representation <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\widehat{\bm{C}}(k)$ by perform- ing the two operations mentioned at the beginning of sec- tion A.2. Then the total approximation error <com:Math><semantics>Ẽ̂(M)(k)<annotation encoding="application / x-tex">\hat{\tilde{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> is computed according to equation (11). Next, the total ap- proximation error <com:Math><semantics>Ẽ̂(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> is transformed to the spatial do- main in order to be represented by general plane waves <com:Math><semantics>w̃̂a(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{w}}_{a}^{(M)}(k)< / annotation>< / semantics>< / com:Math> impinging from the test directions <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_{q}< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>. Arranging the general plane wave signals in the matrix <com:Math><semantics>W̃̂(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{W}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> as [Image disponible dans le document PDF, Image available in the PDF document] (20) the transformation to the spatial domain is expressed by the operation <com:Math><semantics>W̃̂(M)(k)=ℰTẼ̂(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{W}}^{(M)}(k) = \mathbf{\mathcal{E}}^T \widehat{\widetilde{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> . (21) The elements <span class="math inline">$\hat{\bar{\mathcal{P}}}_{\bm{q}}^{(M)}(k,b)$ of the directional power distribu- tion <span class="math inline">$\widehat{\bm{\mathcal{P}}}^{(M)}(k,b)$ of the total approximation error <span class="math inline">$\widehat{\bm{\mathcal{E}}}^{(M)}(k)$ are obtained by computing the powers of the general plane wave functions <com:Math><semantics>w̃̂q(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{w}}_q^{(M)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, within individual criti- cal bands <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>. b. The alternative solution is to compute only the approxi- mation <com:Math><semantics>𝑪̃(M)(k)<annotation encoding="application / x-tex">\widetilde{\pmb{C}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> instead of <com:Math><semantics>𝑪̃̂(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{\pmb{C}}}^{(M)}(k)< / annotation>< / semantics>< / com:Math>. This method offers the advantage that the complicated perceptual coding of the individual signals needs not be carried out directly. In- stead, it is sufficient to know the powers of the percep- tual quantisation error within individual Bark scale critical bands. For this purpose, the total approximation error defined in equation (11) can be written as a sum of the three following approximation errors: [Image disponible dans le document PDF, Image available in the PDF document] (22) [Image disponible dans le document PDF, Image available in the PDF document] (23) [Image disponible dans le document PDF, Image available in the PDF document] (24) which can be assumed to be independent of each other. Due to this independence, the directional power distribution of the total error <com:Math><semantics>Ẽ̂(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> can be expressed as the sum of the directional power distributions of the three individ- ual errors <com:Math><semantics>Ẽ(M)(k)<annotation encoding="application / x-tex">\widetilde{E}^{(M)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>Ẽ̂DIR(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}_{DIR}^{(M)}(k)< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>Ẽ̂AMBRED(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}_{AMBRED}^{(M)}(k)< / annotation>< / semantics>< / com:Math>. The following describes how to compute the directional power distributions of the three errors for individual Bark scale critical bands: a. To compute the directional power distribution of the er- ror <com:Math><semantics>Ẽ(M)(k)<annotation encoding="application / x-tex">\widetilde{E}^{(M)}(k)< / annotation>< / semantics>< / com:Math>, it is first transformed to the spatial domain by <com:Math><semantics>𝑾̃(M)(k)=𝒵T𝑬̃(M)(k)<annotation encoding="application / x-tex">\widetilde{\boldsymbol{W}}^{(M)}(k) = \boldsymbol{\mathcal{Z}}^T \widetilde{\boldsymbol{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math>, (25) wherein the approximation error <com:Math><semantics>Ẽ(M)(k)<annotation encoding="application / x-tex">\widetilde{E}^{(M)}(k)< / annotation>< / semantics>< / com:Math> is hence represent- ed by general plane waves <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\widetilde{\bm{w}}_q^{(M)}(k)$ impinging from the test directions <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_q< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, which are arranged in the matrix <com:Math><semantics>W̃(M)(k)<annotation encoding="application / x-tex">\widetilde{W}^{(M)}(k)< / annotation>< / semantics>< / com:Math> according to [Image disponible dans le document PDF, Image available in the PDF document] (26) Consequently, the elements <com:Math><semantics>ilde𝒫q(M)(k,b)<annotation encoding="application / x-tex">ilde{\mathcal{P}}_q^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> of the directional power distribution <com:Math><semantics>𝒫̃(M)(k,b)<annotation encoding="application / x-tex">\widetilde{\mathcal{P}}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> of the approximation error <com:Math><semantics>ℰ̃(M)(k)<annotation encoding="application / x-tex">\widetilde{\mathcal{E}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> are obtained by computing the powers of the general plane wave functions <com:Math><semantics>𝒘̃q(M)(k)<annotation encoding="application / x-tex">\widetilde{\boldsymbol{w}}_{q}^{(M)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, within individual criti- cal bands <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>. b. For computing the directional power distribution <com:Math><semantics>𝒫̃DIR(M)(k,b)<annotation encoding="application / x-tex">\widetilde{\mathcal{P}}_{\mathrm{DIR}}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> of the error <com:Math><semantics>ẼDIR(M)(k)<annotation encoding="application / x-tex">\widetilde{E}_{\mathrm{DIR}}^{(M)}(k)< / annotation>< / semantics>< / com:Math>, it is to be borne in mind that this error is introduced into the directional HOA component <com:Math><semantics>C̃DIR(M)(k)<annotation encoding="application / x-tex">\widetilde{C}_{\mathrm{DIR}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> by perceptually coding the directional signals <com:Math><semantics>x̃DOM(a)(k)<annotation encoding="application / x-tex">\widetilde{x}_{\mathrm{DOM}}^{(a)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>1≤d≤M<annotation encoding="application / x-tex">1 \leq d \leq M< / annotation>< / semantics>< / com:Math>. Further, it is to be considered that the directional HOA component is given by equation (8). Then for simplicity it is assumed that the HOA component <span class="math inline">$\widetilde{\bm{C}}_{ exttt{DOM.CORR}}^{(d)}(k)$ is equivalently represented in the spatial do- main by <com:Math><semantics>O<annotation encoding="application / x-tex">O< / annotation>< / semantics>< / com:Math> general plane wave functions <span class="math inline">$\widetilde{\bm{v}}_{\mathrm{GRID},o}^{(d)}(k)$, which are created from the directional signal <com:Math><semantics>x̃𝙳𝙾𝙼(d)(k)<annotation encoding="application / x-tex">\widetilde{x}_{\mathtt{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math> by a mere scaling, i.e. <com:Math><semantics>𝒗̃GRID,o(d)(k)=αo(d)(k)𝒙̃DOM(d)(k)<annotation encoding="application / x-tex">\widetilde{\boldsymbol{v}}_{\text{GRID},o}^{(d)}(k) = \alpha_o^{(d)}(k)\widetilde{\boldsymbol{x}}_{\text{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math> , (27) where <com:Math><semantics>αo(d)(k)<annotation encoding="application / x-tex">\alpha_o^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>o=1,...,0<annotation encoding="application / x-tex">o=1,...,0< / annotation>< / semantics>< / com:Math>, denote the scaling parameters. The respective plane wave directions <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\widetilde{\it \Omega}_{{m ROT},o}^{(d)}(k)$, <com:Math><semantics>o=1,...,0<annotation encoding="application / x-tex">o=1,...,0< / annotation>< / semantics>< / com:Math>, are assumed to be uniformly distributed on the unit sphere and rotated such that <span class="math inline">$\widetilde{\bm{\varOmega}}_{\mathrm{ROT},1}^{(d)}(k)$ corresponds to the direc- tion estimate <span class="math inline">$\widetilde{\bm{\Omega}}_{\mathrm{DOM}}^{(d)}(k)$ . Hence, the scaling parameter <com:Math><semantics>α1(d)(k)<annotation encoding="application / x-tex">\alpha_1^{(d)}(k)< / annotation>< / semantics>< / com:Math> is equal to '1'. When defining <com:Math><semantics>𝒵GRID(d)(k)<annotation encoding="application / x-tex">\mathcal{Z}_{GRID}^{(d)}(k)< / annotation>< / semantics>< / com:Math> to be the mode matrix with respect to the rotated directions <com:Math><semantics>𝛺̃ROT,o(d)(k)<annotation encoding="application / x-tex">\tilde{\boldsymbol{\varOmega}}_{\mathrm{ROT},o}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>o=1,…,0<annotation encoding="application / x-tex">o=1,\ldots,0< / annotation>< / semantics>< / com:Math>, and arrang- ing all scaling parameters <com:Math><semantics>αo(d)(k)<annotation encoding="application / x-tex">\alpha_o^{(d)}(k)< / annotation>< / semantics>< / com:Math> in a vector according to [Image disponible dans le document PDF, Image available in the PDF document] (28) the HOA component <span class="math inline">$\widetilde{\bm{c}}_{\mathtt{DOM.CORR}}^{(d)}(k)$ can be written as [Image disponible dans le document PDF, Image available in the PDF document] (29) Consequently, the error <com:Math><semantics>Ẽ̂DIR(M)(k)<annotation encoding="application / x-tex">\hat{\tilde{E}}_{DIR}^{(M)}(k)< / annotation>< / semantics>< / com:Math> (see equation (23)) between the true directional HOA component [Image disponible dans le document PDF, Image available in the PDF document] (30) and that composed from the perceptually decoded direc- tional signals <com:Math><semantics>x̃̂DOM(d)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{x}}_{\mathrm{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>d=1,...,M<annotation encoding="application / x-tex">d=1,...,M< / annotation>< / semantics>< / com:Math>, by [Image disponible dans le document PDF, Image available in the PDF document] (31) [Image disponible dans le document PDF, Image available in the PDF document] (32) can be expressed in terms of the perceptual coding errors [Image disponible dans le document PDF, Image available in the PDF document] (33) in the individual directional signals by [Image disponible dans le document PDF, Image available in the PDF document] (34) The representation of the error <com:Math><semantics>Ẽ̂DIR(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}_{\mathrm{DIR}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> in the spatial do- main with respect to the test directions <com:Math><semantics>Ωq<annotation encoding="application / x-tex">\Omega_q< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, is given by [Image disponible dans le document PDF, Image available in the PDF document] (35) Denoting the elements of the vector <com:Math><semantics>β(d)(k)<annotation encoding="application / x-tex">\beta^{(d)}(k)< / annotation>< / semantics>< / com:Math> by <com:Math><semantics>βq(d)(k)<annotation encoding="application / x-tex">\beta_q^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>q=1,...,Q<annotation encoding="application / x-tex">q=1,...,Q< / annotation>< / semantics>< / com:Math>, and assuming the individual perceptual coding errors <com:Math><semantics>ẽ̂DOM(d)(k)<annotation encoding="application / x-tex">\hat{\tilde{e}}_{\text{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>d=1,...,M<annotation encoding="application / x-tex">d=1,...,M< / annotation>< / semantics>< / com:Math>, to be independent of each other, it follows from equation (35) that the elements <com:Math><semantics>𝒫̂DIR,a(M)(k,b)<annotation encoding="application / x-tex">\hat{\mathcal{P}}_{\mathrm{DIR},a}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> of the directional power distribution <span class="math inline">$\widehat{\bm{\mathcal{P}}}_{ extsf{DIR}}^{(M)}(k,b)$ of the per- ceptual coding error <com:Math><semantics>Ẽ̂DIR(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}_{\mathrm{DIR}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> can be computed by [Image disponible dans le document PDF, Image available in the PDF document] (36) <com:Math><semantics>σ̃DIR,d2(k,b)<annotation encoding="application / x-tex">\widetilde{\sigma}_{\mathrm{DIR},d}^2(k,b)< / annotation>< / semantics>< / com:Math> is supposed to represent the power of the per- ceptual quantisation error within the <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>-th critical band in the directional signal <com:Math><semantics>𝒙̂̂DOM(d)(k)<annotation encoding="application / x-tex">\widehat{\widehat{\pmb{x}}}_{\text{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>. This power can be as- sumed to correspond to the perceptual masking power of the directional signal <com:Math><semantics>𝒙̃DOM(d)(k)<annotation encoding="application / x-tex">\widetilde{\mathbf{x}}_{\mathrm{DOM}}^{(d)}(k)< / annotation>< / semantics>< / com:Math>. c. For computing the directional power distribution <com:Math><semantics>𝒫̂AMB,RED(M)(k,b)<annotation encoding="application / x-tex">\widehat{\boldsymbol{\mathcal{P}}}_{\mathrm{AMB,RED}}^{(M)}(k,b)< / annotation>< / semantics>< / com:Math> of the error <span class="math inline">$\widehat{\bm{ ilde{E}}}_{AMB.RED}^{(M)}(k)$ resulting from the perceptual cod- ing of the HOA coefficient sequences of the ambient HOA component, each HOA coefficient sequence is assumed to be coded independently. Hence, the errors introduced into the individual HOA coefficient sequences within each Bark scale critical band can be assumed to be uncorrelated. This means that the inter-coefficient correlation matrix of the error <com:Math><semantics>Ẽ̂AMB.RED(M)(k)<annotation encoding="application / x-tex">\widehat{\widetilde{E}}_{\mathrm{AMB.RED}}^{(M)}(k)< / annotation>< / semantics>< / com:Math> with respect to each Bark scale critical band is diagonal, i.e. <com:Math><semantics>Σ̃AMB,RED(M)(k,b)=<annotation encoding="application / x-tex">\widetilde{\Sigma}^{(M)}_{AMB,RED}(k,b) =< / annotation>< / semantics>< / com:Math> [Image disponible dans le document PDF, Image available in the PDF document] (37) The elements <com:Math><semantics>σ̃AMB,RED,o2(M)(k,b)<annotation encoding="application / x-tex">\tilde{\sigma}_{AMB,RED,o}^{2(M)}(k,b)< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>o=1,...,O<annotation encoding="application / x-tex">o=1,...,O< / annotation>< / semantics>< / com:Math>, are supposed to repre- sent the power of the perceptual quantisation error with- in the <com:Math><semantics>b<annotation encoding="application / x-tex">b< / annotation>< / semantics>< / com:Math>-th critical band in the <com:Math><semantics>o<annotation encoding="application / x-tex">o< / annotation>< / semantics>< / com:Math>-th coded HOA coeffi- cient sequence in <span class="math inline">$\widehat{\bm{c}}_{\mathrm{AMB.RED}}^{(M)}(k)$. They can be assumed to cor- respond to the perceptual masking power of the <com:Math><semantics>o<annotation encoding="application / x-tex">o< / annotation>< / semantics>< / com:Math>-th HOA coefficient sequence <com:Math><semantics>𝑪̃AMB,RED(M)(k)<annotation encoding="application / x-tex">\tilde{\pmb{C}}_{\text{AMB,RED}}^{(M)}(k)< / annotation>< / semantics>< / com:Math>. The directional power distribution of the perceptual coding error <span class="math inline">$\widehat{\bm{\bar{E}}}_{\mathrm{AMB},\mathrm{RED}}^{(M)}(k)$ is thus computed by [Image disponible dans le document PDF, Image available in the PDF document] (38) B. Improved HOA decompression The corresponding HOA decompression processing is depicted in Fig. 3 and includes the following steps or stages. In step or stage 31 a perceptual decoding of the <com:Math><semantics>I<annotation encoding="application / x-tex">I< / annotation>< / semantics>< / com:Math> signals contained in <com:Math><semantics>Y(k−2)<annotation encoding="application / x-tex">Y(k-2)< / annotation>< / semantics>< / com:Math> is performed in order to obtain the <com:Math><semantics>I<annotation encoding="application / x-tex">I< / annotation>< / semantics>< / com:Math> decoded signals in <com:Math><semantics>Ŷ(k−2)<annotation encoding="application / x-tex">\widehat{Y}(k-2)< / annotation>< / semantics>< / com:Math>. In signal re-distributing step or stage 32, the perceptually decoded signals in <com:Math><semantics>Ŷ(k−2)<annotation encoding="application / x-tex">\widehat{Y}(k-2)< / annotation>< / semantics>< / com:Math> are re-distributed in order to recreate the frame <com:Math><semantics>X̂DIR(k−2)<annotation encoding="application / x-tex">\widehat{X}_{\mathrm{DIR}}(k-2)< / annotation>< / semantics>< / com:Math> of directional signals and the frame <com:Math><semantics>𝒞¯AMB,RED(k−2)<annotation encoding="application / x-tex">\overline{\mathcal{C}}_{\mathrm{AMB},\mathrm{RED}}(k-2)< / annotation>< / semantics>< / com:Math> of the ambient HOA component. The infor- mation about how to re-distribute the signals is obtained by reproducing the assigning operation performed for the HOA compression, using the index data sets <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\bar{J}_{ ext{DIR,ACT}}(k)$ and <com:Math><semantics>JAMB,ACT(k−2)<annotation encoding="application / x-tex">J_{AMB,ACT}(k-2)< / annotation>< / semantics>< / com:Math>. Since this is a recursive procedure (see sec- tion A), the additionally transmitted assignment vector <com:Math><semantics>γ(k)<annotation encoding="application / x-tex">\gamma(k)< / annotation>< / semantics>< / com:Math> can be used in order to allow for an initialisation of the re-distribution procedure, e.g. in case the transmission is breaking down. In composition step or stage 33, a current frame <com:Math><semantics>Ĉ(k−3)<annotation encoding="application / x-tex">\widehat{C}(k-3)< / annotation>< / semantics>< / com:Math> of the desired total HOA representation is re-composed (accord- ing to the processing described in connection with Fig. 2b and Fig. 4 of EP 12306569.0 using the frame <com:Math><semantics>X̂DIR(k−2)<annotation encoding="application / x-tex">\widehat{X}_{DIR}(k-2)< / annotation>< / semantics>< / com:Math> of the directional signals, the set <com:Math><semantics>J̃DIR,ACT(k)<annotation encoding="application / x-tex">\tilde{J}_{\mathrm{DIR,ACT}}(k)< / annotation>< / semantics>< / com:Math> of the active direc- tional signal indices together with the set <com:Math><semantics>𝒢̃Ω,ACT(k)<annotation encoding="application / x-tex">\mathcal{\tilde{G}}_{\Omega,\mathrm{ACT}}(k)< / annotation>< / semantics>< / com:Math> of the corresponding directions, the parameters <com:Math><semantics>ζ(k−2)<annotation encoding="application / x-tex">\zeta(k-2)< / annotation>< / semantics>< / com:Math> for predict- ing portions of the HOA representation from the directional signals, and the frame <span class="math inline">$\widehat{\bm{C}}_{\mathsf{AMB},\mathsf{RED}}(k-2)$ of HOA coefficient se- quences of the reduced ambient HOA component. <com:Math><semantics>𝒄̂AMB,RED(k−2)<annotation encoding="application / x-tex">\widehat{\boldsymbol{c}}_{\text{AMB,RED}}(k-2)< / annotation>< / semantics>< / com:Math> corresponds to component <com:Math><semantics>𝑫̂A(k−2)<annotation encoding="application / x-tex">\widehat{\boldsymbol{D}}_{A}(k-2)< / annotation>< / semantics>< / com:Math> in EP 12306569.0, and <com:Math><semantics>𝒢̃Ω.ACT(k)<annotation encoding="application / x-tex">\tilde{\mathcal{G}}_{\Omega.\mathrm{ACT}}(k)< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>ℐ̃DIR.ACT(k)<annotation encoding="application / x-tex">\tilde{\mathcal{I}}_{\mathrm{DIR.ACT}}(k)< / annotation>< / semantics>< / com:Math> correspond to <com:Math><semantics>AΩ̂(k)<annotation encoding="application / x-tex">A_{\widehat{\Omega}}(k)< / annotation>< / semantics>< / com:Math> in EP 12306569.0, wherein active directional signal indices are marked in the matrix elements of <com:Math><semantics>Aâ(k)<annotation encoding="application / x-tex">A_{\widehat{a}}(k)< / annotation>< / semantics>< / com:Math>. I.e., directional signals with re- spect to uniformly distributed directions are predicted from the directional signals <com:Math><semantics>(X̂DIR(k−2))<annotation encoding="application / x-tex">(\widehat{X}_{\mathrm{DIR}}(k-2))< / annotation>< / semantics>< / com:Math> using the received param- eters <com:Math><semantics>(ζ(k−2))<annotation encoding="application / x-tex">(\zeta(k-2))< / annotation>< / semantics>< / com:Math> for such prediction, and thereafter the cur- rent decompressed frame <com:Math><semantics>(𝒄̂(k−3))<annotation encoding="application / x-tex">(\widehat{\boldsymbol{c}}(k-3))< / annotation>< / semantics>< / com:Math> is re-composed from the frame of directional signals <com:Math><semantics>(X̂DIR(k−2))<annotation encoding="application / x-tex">(\widehat{X}_{\mathrm{DIR}}(k-2))< / annotation>< / semantics>< / com:Math>, the predicted por- 25 tions and the reduced ambient HOA component <com:Math><semantics>(CAMB,RED(k−2))<annotation encoding="application / x-tex">(C_{AMB,RED}(k-2))< / annotation>< / semantics>< / com:Math>. C. Basics of Higher Order Ambisonics Higher Order Ambisonics (HOA) is based on the description of a sound field within a compact area of interest, which is assumed to be free of sound sources. In that case the spati- otemporal behaviour of the sound pressure <com:Math><semantics>p(t,x)<annotation encoding="application / x-tex">p(t,x)< / annotation>< / semantics>< / com:Math> at time <com:Math><semantics>t<annotation encoding="application / x-tex">t< / annotation>< / semantics>< / com:Math> and position x within the area of interest is physically fully determined by the homogeneous wave equation. In the follow- ing a spherical coordinate system as shown in Fig. 4 is as- sumed. In the used coordinate system the <com:Math><semantics>x<annotation encoding="application / x-tex">x< / annotation>< / semantics>< / com:Math> axis points to the frontal position, the <com:Math><semantics>y<annotation encoding="application / x-tex">y< / annotation>< / semantics>< / com:Math> axis points to the left, and the z axis points to the top. A position in space <com:Math><semantics>x=(r,θ,ϕ)T<annotation encoding="application / x-tex">x = (r, \theta, \phi)^T< / annotation>< / semantics>< / com:Math> is represented by a radius <com:Math><semantics>r>0<annotation encoding="application / x-tex">r>0< / annotation>< / semantics>< / com:Math> (i.e. the distance to the coor- dinate origin), an inclination angle <com:Math><semantics>θ∈[0,π]<annotation encoding="application / x-tex">\theta \in [0,\pi]< / annotation>< / semantics>< / com:Math> measured from the polar axis z and an azimuth angle <com:Math><semantics>ϕ∈[0,2π]<annotation encoding="application / x-tex">\phi \in [0,2\pi]< / annotation>< / semantics>< / com:Math> measured coun- ter-clockwise in the <com:Math><semantics>x−y<annotation encoding="application / x-tex">x-y< / annotation>< / semantics>< / com:Math> plane from the <com:Math><semantics>x<annotation encoding="application / x-tex">x< / annotation>< / semantics>< / com:Math> axis. Further, <com:Math><semantics>(⋅)T<annotation encoding="application / x-tex">(\cdot)^T< / annotation>< / semantics>< / com:Math> denotes the transposition. It can be shown (see E.G. Williams, "Fourier Acoustics", volume 93 of Applied Mathematical Sciences, Academic Press, 1999) that the Fourier transform of the sound pressure with respect to time denoted by <com:Math><semantics>ℱt(⋅)<annotation encoding="application / x-tex">\mathcal{F}_t(\cdot)< / annotation>< / semantics>< / com:Math>, i.e. [Image disponible dans le document PDF, Image available in the PDF document] (39) with <com:Math><semantics>ω<annotation encoding="application / x-tex">\omega< / annotation>< / semantics>< / com:Math> denoting the angular frequency and <com:Math><semantics>i<annotation encoding="application / x-tex">i< / annotation>< / semantics>< / com:Math> indicating the imaginary unit, can be expanded into a series of Spherical Harmonics according to [Image disponible dans le document PDF, Image available in the PDF document] (40) In equation (40), <com:Math><semantics>cs<annotation encoding="application / x-tex">c_s< / annotation>< / semantics>< / com:Math> denotes the speed of sound and <com:Math><semantics>k<annotation encoding="application / x-tex">k< / annotation>< / semantics>< / com:Math> denotes the angular wave number, which is related to the angular frequency <com:Math><semantics>ω<annotation encoding="application / x-tex">\omega< / annotation>< / semantics>< / com:Math> by <com:Math><semantics>k=ωcs<annotation encoding="application / x-tex">k=\frac{\omega}{c_s}< / annotation>< / semantics>< / com:Math>. Further, <com:Math><semantics>jn(⋅)<annotation encoding="application / x-tex">j_n(\cdot)< / annotation>< / semantics>< / com:Math> denote the spherical Bes- sel functions of the first kind and <com:Math><semantics>Snm(θ,ϕ)<annotation encoding="application / x-tex">S_n^m(\theta,\phi)< / annotation>< / semantics>< / com:Math> denote the real valued Spherical Harmonics of order <com:Math><semantics>n<annotation encoding="application / x-tex">n< / annotation>< / semantics>< / com:Math> and degree <com:Math><semantics>m<annotation encoding="application / x-tex">m< / annotation>< / semantics>< / com:Math>, which are defined in below section C.1. The expansion coefficients <com:Math><semantics>Anm(k)<annotation encoding="application / x-tex">A_n^m(k)< / annotation>< / semantics>< / com:Math> are depending only on the angular wave number k. In the foregoing it has been implicitly assumed that sound pressure is spatially band-limited. Thus the series of Spherical Har- monics is truncated with respect to the order index <com:Math><semantics>n<annotation encoding="application / x-tex">n< / annotation>< / semantics>< / com:Math> at an upper limit <com:Math><semantics>N<annotation encoding="application / x-tex">N< / annotation>< / semantics>< / com:Math>, which is called the order of the HOA repre- sentation. If the sound field is represented by a superposition of an infinite number of harmonic plane waves of different angular frequencies <com:Math><semantics>ω<annotation encoding="application / x-tex">\omega< / annotation>< / semantics>< / com:Math> arriving from all possible directions speci- fied by the angle tuple <com:Math><semantics>(θ,ϕ)<annotation encoding="application / x-tex">(\theta,\phi)< / annotation>< / semantics>< / com:Math>, it can be shown (see B. Ra- faely, "Plane-wave Decomposition of the Sound Field on a Sphere by Spherical Convolution", Journal of the Acoustical Society of America, vol.4(116), pages 2149-2157, 2004) that the respective plane wave complex amplitude function <com:Math><semantics>C(ω,θ,ϕ)<annotation encoding="application / x-tex">C(\omega,\theta,\phi)< / annotation>< / semantics>< / com:Math> can be expressed by the following Spherical Harmonics expan- sion [Image disponible dans le document PDF, Image available in the PDF document] (41) where the expansion coefficients <com:Math><semantics>Cnm(k)<annotation encoding="application / x-tex">C_n^m(k)< / annotation>< / semantics>< / com:Math> are related to the expansion coefficients <com:Math><semantics>Anm(k)<annotation encoding="application / x-tex">A_n^m(k)< / annotation>< / semantics>< / com:Math> by <com:Math><semantics>Anm(k)=4πinCnm(k)<annotation encoding="application / x-tex">A_n^m(k) = 4\pi i^n C_n^m(k)< / annotation>< / semantics>< / com:Math>. (42) Assuming the individual coefficients <com:Math><semantics>Cnm(ω=kcs)<annotation encoding="application / x-tex">C_n^m(\omega=kc_s)< / annotation>< / semantics>< / com:Math> to be func- tions of the angular frequency <com:Math><semantics>ωℓ<annotation encoding="application / x-tex">\omega_{\ell}< / annotation>< / semantics>< / com:Math> the application of the in- verse Fourier transform (denoted by <com:Math><semantics>ℱ−1(⋅)<annotation encoding="application / x-tex">\mathcal{F}^{-1}(\cdot)< / annotation>< / semantics>< / com:Math>) provides time do- main functions [Image disponible dans le document PDF, Image available in the PDF document] (43) for each order <com:Math><semantics>n<annotation encoding="application / x-tex">n< / annotation>< / semantics>< / com:Math> and degree <com:Math><semantics>m<annotation encoding="application / x-tex">m< / annotation>< / semantics>< / com:Math>, which can be collected in a single vector <com:Math><semantics>c(t)<annotation encoding="application / x-tex">c(t)< / annotation>< / semantics>< / com:Math> by <com:Math><semantics>c(t)=<annotation encoding="application / x-tex">c(t) =< / annotation>< / semantics>< / com:Math> (44) [Image disponible dans le document PDF, Image available in the PDF document] The position index of a time domain function <com:Math><semantics>cnm(t)<annotation encoding="application / x-tex">c_n^m(t)< / annotation>< / semantics>< / com:Math> within the vector <com:Math><semantics>c(t)<annotation encoding="application / x-tex">c(t)< / annotation>< / semantics>< / com:Math> is given by <com:Math><semantics>n(n+1)+1+m<annotation encoding="application / x-tex">n(n+1)+1+m< / annotation>< / semantics>< / com:Math>. The overall number of elements in vector <com:Math><semantics>c(t)<annotation encoding="application / x-tex">c(t)< / annotation>< / semantics>< / com:Math> is given by <com:Math><semantics>0=(N+1)2<annotation encoding="application / x-tex">0 = (N+1)^2< / annotation>< / semantics>< / com:Math>. The final Ambisonics format provides the sampled version of <com:Math><semantics>c(t)<annotation encoding="application / x-tex">c(t)< / annotation>< / semantics>< / com:Math> using a sampling frequency <com:Math><semantics>fS<annotation encoding="application / x-tex">f_S< / annotation>< / semantics>< / com:Math> as [Image disponible dans le document PDF, Image available in the PDF document] (45) where <com:Math><semantics>TS=1 / fS<annotation encoding="application / x-tex">T_S = 1 / f_S< / annotation>< / semantics>< / com:Math> denotes the sampling period. The elements of <com:Math><semantics>c(lTS)<annotation encoding="application / x-tex">c(lT_S)< / annotation>< / semantics>< / com:Math> are here referred to as Ambisonics coefficients. The time domain signals <com:Math><semantics>cnm(t)<annotation encoding="application / x-tex">c_n^m(t)< / annotation>< / semantics>< / com:Math> and hence the Ambisonics coeffi- cients are real-valued. C.1 Definition of real-valued Spherical Harmonics The real-valued spherical harmonics <com:Math><semantics>Snm(θ,ϕ)<annotation encoding="application / x-tex">S_n^m(\theta,\phi)< / annotation>< / semantics>< / com:Math> are given by [Image disponible dans le document PDF, Image available in the PDF document] (46) [Image disponible dans le document PDF, Image available in the PDF document] (47) The associated Legendre functions <com:Math><semantics>Pn,m(𝒙)<annotation encoding="application / x-tex">P_{n,m}(\mathbf{x})< / annotation>< / semantics>< / com:Math> are defined as [Image disponible dans le document PDF, Image available in the PDF document] (48) with the Legendre polynomial <com:Math><semantics>Pn(x)<annotation encoding="application / x-tex">P_n(x)< / annotation>< / semantics>< / com:Math> and, unlike in the above- mentioned Williams article, without the Condon-Shortley phase term <com:Math><semantics>(−1)m<annotation encoding="application / x-tex">(-1)^m< / annotation>< / semantics>< / com:Math>. C.2 Spatial resolution of Higher Order Ambisonics A general plane wave function <com:Math><semantics>x(t)<annotation encoding="application / x-tex">x(t)< / annotation>< / semantics>< / com:Math> arriving from a direction <com:Math><semantics>𝛀0=(θ0,ϕ0)T<annotation encoding="application / x-tex">\mathbf{\Omega}_0 = (\theta_0, \phi_0)^T< / annotation>< / semantics>< / com:Math> is represented in HOA by [Image disponible dans le document PDF, Image available in the PDF document] (49) The corresponding spatial density of plane wave amplitudes <com:Math><semantics>c(t,Ω):=ℱt−1(𝒞(ω,Ω))<annotation encoding="application / x-tex">c(t,\Omega):=\mathcal{F}_t^{-1}\big(\mathcal{C}(\omega,\Omega)\big)< / annotation>< / semantics>< / com:Math> is given by [Image disponible dans le document PDF, Image available in the PDF document] (50) [Image disponible dans le document PDF, Image available in the PDF document] (51) It can be seen from equation (51) that it is a product of the general plane wave function <com:Math><semantics>x(t)<annotation encoding="application / x-tex">x(t)< / annotation>< / semantics>< / com:Math> and of a spatial disper- sion function <com:Math><semantics>vN(𝚯)<annotation encoding="application / x-tex">v_N(\boldsymbol{\Theta})< / annotation>< / semantics>< / com:Math>, which can be shown to only depend on the angle <com:Math><semantics>θ<annotation encoding="application / x-tex">\theta< / annotation>< / semantics>< / com:Math> between <com:Math><semantics>Ω<annotation encoding="application / x-tex">\Omega< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>Ω0<annotation encoding="application / x-tex">\Omega_0< / annotation>< / semantics>< / com:Math> having the property [Image disponible dans le document PDF, Image available in the PDF document] (52) As expected, in the limit of an infinite order, i.e., <com:Math><semantics>N→∞<annotation encoding="application / x-tex">N \to \infty< / annotation>< / semantics>< / com:Math>, the spatial dispersion function turns into a Dirac delta [Image disponible dans le document PDF, Image available in the PDF document] (53) However, in the case of a finite order <com:Math><semantics>N<annotation encoding="application / x-tex">N< / annotation>< / semantics>< / com:Math>, the contribution of the general plane wave from direction <com:Math><semantics>Ω0<annotation encoding="application / x-tex">\Omega_0< / annotation>< / semantics>< / com:Math> is smeared to neighbouring directions, where the extent of the blurring decreases with an increasing order. A plot of the normalised function <com:Math><semantics>vN(θ)<annotation encoding="application / x-tex">v_N(\theta)< / annotation>< / semantics>< / com:Math> for different values of N is shown in Fig. 5. It should be pointed out that for any direction <com:Math><semantics>Ω<annotation encoding="application / x-tex">\Omega< / annotation>< / semantics>< / com:Math> the time domain behaviour of the spatial density of plane wave ampli- tudes is a multiple of its behaviour at any other direction. In particular, the functions <com:Math><semantics>c(t,Ω1)<annotation encoding="application / x-tex">c(t,\Omega_1)< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>c(t,Ω2)<annotation encoding="application / x-tex">c(t,\Omega_2)< / annotation>< / semantics>< / com:Math> for some fixed directions <com:Math><semantics>Ω1<annotation encoding="application / x-tex">\Omega_1< / annotation>< / semantics>< / com:Math> and <com:Math><semantics>Ω2<annotation encoding="application / x-tex">\Omega_2< / annotation>< / semantics>< / com:Math> are highly correlated with each other with respect to time <com:Math><semantics>t<annotation encoding="application / x-tex">t< / annotation>< / semantics>< / com:Math>. C.3 Spherical Harmonic Transform If the spatial density of plane wave amplitudes is discre- tised at a number of <com:Math><semantics>O<annotation encoding="application / x-tex">O< / annotation>< / semantics>< / com:Math> spatial directions <com:Math><semantics>Ωo<annotation encoding="application / x-tex">\Omega_{o}< / annotation>< / semantics>< / com:Math>, <com:Math><semantics>1≤o≤O<annotation encoding="application / x-tex">1 \le o \le O< / annotation>< / semantics>< / com:Math>, which are nearly uniformly distributed on the unit sphere, <com:Math><semantics>o<annotation encoding="application / x-tex">o< / annotation>< / semantics>< / com:Math> di- rectional signals <com:Math><semantics>c(t,Ωo)<annotation encoding="application / x-tex">c(t,\Omega_o)< / annotation>< / semantics>< / com:Math> are obtained. Collecting these sig- nals into a vector as <com:Math><semantics>𝒄SPAT(t):=[c(t,𝛀1)...c(t,𝛀O)]T<annotation encoding="application / x-tex">\boldsymbol{c}_{\mathrm{SPAT}}(t) := [c(t, \boldsymbol{\Omega}_1) \ ... \ c(t, \boldsymbol{\Omega}_O)]^T< / annotation>< / semantics>< / com:Math> , (54) by using equation (50) it can be verified that this vector can be computed from the continuous Ambisonics representa- tion <com:Math><semantics>d(t)<annotation encoding="application / x-tex">d(t)< / annotation>< / semantics>< / com:Math> defined in equation (44) by a simple matrix multi- plication as <com:Math><semantics>cSPAT(t)=ΨHc(t)<annotation encoding="application / x-tex">c_{SPAT}(t) = \Psi^H c(t)< / annotation>< / semantics>< / com:Math> , (55) where <com:Math><semantics>(⋅)H<annotation encoding="application / x-tex">(\cdot)^H< / annotation>< / semantics>< / com:Math> indicates the joint transposition and conjugation, and <com:Math><semantics>Ψ<annotation encoding="application / x-tex">\Psi< / annotation>< / semantics>< / com:Math> denotes a mode-matrix defined by <com:Math><semantics>Ψ:=[S1…So]<annotation encoding="application / x-tex">\Psi := [S_1 \dots S_o]< / annotation>< / semantics>< / com:Math> (56) with [Image disponible dans le document PDF, Image available in the PDF document] (57) 29 Because the directions <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\boldsymbol{\varOmega_o}$ are nearly uniformly distributed on the unit sphere, the mode matrix is invertible in gen- eral. Hence, the continuous Ambisonics representation can be computed from the directional signals <com:Math><semantics>c(t,Ωo)<annotation encoding="application / x-tex">c(t,\Omega_o)< / annotation>< / semantics>< / com:Math> by [Image disponible dans le document PDF, Image available in the PDF document] (58) Both equations constitute a transform and an inverse trans- form between the Ambisonics representation and the spatial domain. These transforms are here called the Spherical Har- monic Transform and the inverse Spherical Harmonic Trans- form. It should be noted that since the directions <com:Math><semantics>Ω0<annotation encoding="application / x-tex">\Omega_0< / annotation>< / semantics>< / com:Math> are nearly uniformly distributed on the unit sphere, the approximation [Image disponible dans le document PDF, Image available in the PDF document] (59) is available, which justifies the use of <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\bm{\Psi}^{-1}$ instead of <Formule mathématique disponible dans le document PDF, Math available in the PDF document>$\bm{\Psi}^{H}$ in equation (55). Advantageously, all the mentioned relations are valid for the discrete-time domain, too. The inventive processing can be carried out by a single pro- cessor or electronic circuit, or by several processors or electronic circuits operating in parallel and / or operating on different parts of the inventive processing.

Claims

30 CLAIMS 1. A method for decompressing a compressed Higher Order Ambisonics (HOA) representation, the method comprising: decoding the compressed HOA representation to provide a decoded frame of signals and an assignment vector, the assignment vector conveying indices of inactive directional signals of the decoded frame of signals possibly containing coefficient sequences of relating to non-zero ambient HOA components; determining a first set of indices indicating active directional signals of the decoded frame of signals, and determining respective directions of the active directional signals; re-distributing the decoded frame of signals based on the first set of indices and the respective directions of the active directional signals, wherein the re-distributing determines a frame of HOA directional signals and a frame of ambient HOA components; outputting the frame of HOA directional signals; and outputting the frame of ambient HOA components.

2. The method of claim 1, wherein the frame of HOA directional signals is created based on the first set of indices.

3. The method of claim 1, wherein the first set of indices correspond to a set direction estimates.

4. The method of claim 3, wherein direction estimates in the set of direction estimates correspond to dominant components of the compressed HOA representation.

5. A non-transitory computer readable storage medium containing instructions that, when executed by a processor, perform the method according to any one of claims 1 to 4.

6. An apparatus for decompressing a compressed Higher Order Ambisonics (HOA) representation, the apparatus comprising: a decoder for decoding the compressed HOA representation to provide a decoded frame of signals and an assignment vector, the assignment vector conveying indices of inactive directional signals of the decoded frame of signals possibly containing coefficient sequences of relating to non-zero ambient HOA components; a first processor for determining a first set of indices indicating active directional signals of the decoded frame of signals, and determining respective directions of the active directional signals; a second processor for re-distributing the decoded frame of signals based on the first set of indices and the respective directions of the active directional signals, wherein the re- distributing determines a frame of HOA directional signals and a frame of ambient HOA components; and the second processor being configured for outputting the frame of HOA directional signals and for outputting the frame of ambient HOA components.

7. The apparatus of claim 6, wherein the first set of indices correspond with a set of direction estimates.

8. The apparatus of claim 7, wherein direction estimates in 5 the set of direction estimates correspond to dominant components of the compressed HOA representation.