The present invention relates to the encoding / decoding of spatialized sound data, in particular in a surround sound context (hereinafter also referred to as “ambisonic”). The coders / decoders (hereinafter called “codecs”) which are currently used in mobile telephony are mono (a single signal channel for reproduction on a single loudspeaker). The 3GPP EVS (for “Enhanced Voice Services”) codec makes it possible to offer “Super-HD” quality (also called “High Definition +” or HD + voice) with an audio band in super-wide band (SWB for “super- wideband ”) for signals sampled at 32 or 48 kHz or full band (FB for“ Fullband ”) for signals sampled at 48 kHz; the audio bandwidth is 14.4 to 16 kHz in SWB mode (9.6 to 128 kbit / s) and 20 kHz in FB mode (16.4 to 128 kbit / s). The next quality development in conversational services offered by operators should be immersive services, using terminals such as smartphones, for example, equipped with several microphones or spatialized audio conferencing or tele-presence type videoconferencing equipment. , or even “live” content sharing tools, with spatialized 3D sound rendering, which is far more immersive than a simple 2D stereo rendering. With the increasingly widespread use of listening on mobile phones with an audio headset and the appearance of advanced audio equipment (accessories such as a 3D microphone, voice assistants with acoustic antennas, virtual reality headsets, etc. As such, the future 3GPP standard "IVAS" (for "Immersive Voice And Audio Services") proposes the extension of the EVS codec to immersive by accepting as input format of the codec at least the spatialized sound formats listed below. below (and their combinations): - Multichannel format (channel-based in English) of stereo type, 5.1 where each channel feeds a speaker (for example L and R in stereo, or L, R, Ls, Rs and C in 5.1) - Object-based format where sound objects are described as an audio signal (generally mono) associated with metadata describing the attributes of this object (position in space, spatial width of the source, etc. ), and - Ambisonic format (scene-based in English) which describes the sound field at a given point, generally picked up by a spherical microphone or synthesized in the field of spherical harmonics. Hereinafter, we are typically interested in the coding of a sound in ambisonic format, by way of example of an embodiment (at least certain aspects presented in connection with the invention below can also be applied to other formats. than ambisonics). Ambisonics is a recording method (“encoding” in the acoustic sense) of spatialized sound and a reproduction system (“decoding” in the acoustic sense). An ambisonic microphone (at order 1) comprises at least four capsules (typically of the cardoid or sub-cardoid type) arranged on a spherical grid, for example the vertices of a regular tetrahedron. The audio channels associated with these capsules are called “A-format”. This format is converted into a “B-format”, in which the sound field is broken down into four components (spherical harmonics) denoted W, X, Y, Z, which correspond to four coincident virtual microphones. The W component corresponds to an omnidirectional capture of the sound field while the X, Y and Z components, more directive, can be compared to pressure gradients oriented along the three dimensions of space. An ambisonic system is a flexible system in the sense that recording and playback are separate and decoupled. It allows decoding (in the acoustic sense) on any speaker configuration (for example, binaural, type 5.1 surround sound or periphery (with elevation) type 7.1.4). Of course, the ambisonic approach can be generalized to more than four channels in B-format and this generalized representation is commonly called “HOA” (for “Higher-Order Ambisonics”). The fact of breaking down the sound into more spherical harmonics improves the spatial accuracy of reproduction when rendering on loudspeakers. An ambisonic system is a flexible system in the sense that recording and playback are separate and decoupled. It allows decoding (in the acoustic sense) on any speaker configuration (for example, binaural, type 5.1 surround sound or periphery (with elevation) type 7.1.4). Of course, the ambisonic approach can be generalized to more than four channels in B-format and this generalized representation is commonly called “HOA” (for “Higher-Order Ambisonics”). The fact of breaking down the sound into more spherical harmonics improves the spatial accuracy of reproduction when rendering on loudspeakers. An ambisonic system is a flexible system in the sense that recording and playback are separate and decoupled. It allows decoding (in the acoustic sense) on any speaker configuration (for example, binaural, type 5.1 surround sound or periphery (with elevation) type 7.1.4). Of course, the ambisonic approach can be generalized to more than four channels in B-format and this generalized representation is commonly called “HOA” (for “Higher-Order Ambisonics”). The fact of breaking down the sound into more spherical harmonics improves the spatial accuracy of reproduction when rendering on loudspeakers. 1 or periphery (with elevation) type 7.1.4). Of course, the ambisonic approach can be generalized to more than four channels in B-format and this generalized representation is commonly called “HOA” (for “Higher-Order Ambisonics”). The fact of breaking down the sound into more spherical harmonics improves the spatial accuracy of reproduction when rendering on loudspeakers. 1 or periphery (with elevation) type 7.1.4). Of course, the ambisonic approach can be generalized to more than four channels in B-format and this generalized representation is commonly called “HOA” (for “Higher-Order Ambisonics”). The fact of breaking down the sound into more spherical harmonics improves the spatial accuracy of reproduction when rendering on loudspeakers. An ambisonic signal at order N comprises (N + 1) ² components and, at order 1 (if N = 1), we find the four components of the original ambisonic which is commonly called FOA (for First-Order Ambisonics). There is also a so-called “planar” variant of ambisonics which breaks down the sound defined in a plane which is generally the horizontal plane. In this case, the number of components is 2N + 1 channels. First-order ambisonics (4 channels: W, X, Y, Z) and 1st-order planar ambisonics (3 channels: W, X, Y) are hereinafter referred to as “ambisonics” indiscriminately to facilitate reading, the treatments presented being applicable independently of the planar type or not. If, however, in some passages it is necessary to make a distinction, the terms “1st-order ambisonics” and “1st-order planar ambisonics” are used. In the following, a predetermined order B-format signal is called “ambisonic sound”. In variations, the ambisonic sound can be defined in another format such as A-format or pre-combined channels by fixed matrixing (keeping the number of channels or reducing it to a 3 or 2 channel case), as will be seen. further away. The signals to be processed by the encoder / decoder are in the form of successions of blocks of sound samples called “frames” or “sub-frames” below. In addition, hereafter, the mathematical notations follow the following convention: - Vector: u (lowercase, bold) - Matrix: A (uppercase, bold) The simplest approach to encoding a stereo or ambisonic signal is to use a mono encoder and apply it in parallel to all the channels with possibly a different bit allocation depending on the channels. This approach is called here “multi-mono” (even if in practice we can generalize the approach to multi-stereo or the use of several parallel instances of the same core codec). Such an embodiment is shown in Figure 1. The input signal is divided into channels (mono) by block 100. These channels are individually coded by blocks 120 to 122 according to a predetermined allocation. Their bit stream is multiplexed (block 130) and after transmission and / or storage it is demultiplexed (block 140) to apply a decoding of each of the channels (blocks 150 to 152) which are recombined (block 160). The associated quality varies according to the mono coding used, and it is generally satisfactory only at very high speed, for example with a data rate of at least 48 kbit / s per mono channel for an EVS coding. Thus at order 1 we obtain a minimum bit rate of 4x48 = 192 kbit / s. The solutions currently proposed for more sophisticated codecs, for ambisonic spatialization in particular, are not satisfactory, in particular in terms of complexity, delay and efficient use of the bit rate, to ensure efficient decorrelation between ambisonic channels. For example, the MPEG-H codec for ambisonic sounds uses an add-overlap operation which adds delay and complexity, as well as linear interpolation on direction vectors which is suboptimal and introduces defects. A basic problem with this codec is that it implements a decomposition into predominant components and ambience because the predominant components are supposed to be perceptually distinct from ambience, but this decomposition is not fully specified. The MPEG-H encoder suffers from a problem of mismatch between the directions of the principal components from one frame to another: the order of the components (signals) can be swapped just like the associated directions. Furthermore, it would be possible to use frequency coding approaches (in the FFT or MDCT domain) rather than time coding as in the MPEG-H codec, but signal processing in the frequency domain (sub-bands) makes it necessary to transmit data by sub-band to a decoder, thus increasing the bit rate necessary for this transmission. The present invention improves this situation. To this end, it proposes a method of coding in compression of sound signals forming a succession in time of frames of samples, in each of N channels in ambisonic representation of order greater than 0, the method comprising: - form, from the channels for a current frame, a covariance matrix between channels and search for eigenvectors of the covariance matrix to obtain a matrix of eigenvectors, - test the eigenvector matrix to check that it represents a rotation in a space of dimension N and otherwise correct the eigenvector matrix until obtaining a rotation matrix, for the current frame, and - Applying said rotation matrix to the signals of the N channels before encoding by separate channels of said signals. Thus, the present invention makes it possible to improve a decorrelation between the N channels to be encoded separately subsequently. This separate encoding is hereinafter also referred to as "multi-mono encoding". In one embodiment, the method may further include: - coding parameters taken from the rotation matrix with a view to transmission via a network. These parameters can typically be quaternion and / or rotation angle and / or Euler angle values ​​as will be seen below, or even simply elements of this matrix for example. In one embodiment, the method may further include: - compare the eigenvector matrix obtained for the current frame with a rotation matrix obtained for a frame preceding the current frame, and - swap columns of the eigenvector matrix of the current frame to ensure consistency with the rotation matrix of the previous frame. Such an embodiment makes it possible to maintain overall homogeneity and in particular to avoid audible clicks from one frame to another, during sound reproduction. However, certain transformations implemented to obtain the eigenvectors from the covariance matrix (like the “PCA / KLT” seen later) are likely to reverse the meaning of some of the eigenvectors and it is then necessary to check at the same time a coherence of axis, then of direction on this axis, of each eigenvector of the matrix of the current frame. To this end, in one embodiment, the aforementioned permutation of the columns already making it possible to ensure a coherence of the axes of the vectors, the method further comprises: - check, for each eigenvector of the current frame, a coherence of direction with a column vector of corresponding position of the rotation matrix of the previous frame, and - in the event of inconsistency, invert the sign of the elements of this eigenvector in the matrix of eigenvectors of the current frame. Typically, a permutation between columns of the eigenvector matrix inverting the sign of a determinant of the eigenvector matrix and the determinant of a rotation matrix being equal to 1, we can estimate the determinant of the eigenvector matrix, and if the latter is equal to -1, we can then invert the signs of the elements of a chosen column of the eigenvector matrix, so that the determinant is equal to 1 and thereby form a rotation matrix. In one embodiment, the method may further comprise: - an estimate of the difference between the rotation matrix obtained for the current frame and a rotation matrix obtained for a frame preceding the current frame, - as a function of the estimated difference, determining whether at least one interpolation is to be made between the rotation matrix of the current frame and the rotation matrix of the previous frame. Such an interpolation then makes it possible to smooth (“progressively average”) the rotation matrices applied respectively to the previous frame and the current frame and thus to attenuate an audible click effect from one frame to another on playback. In such a realization: - as a function of the estimated difference, a number of interpolations to be made between the rotation matrix of the current frame and the rotation matrix of the previous frame is determined, - the current frame is divided into a number of sub- frames corresponding to the number of interpolations to operate, and - At least this number of interpolations can be coded with a view to transmission via the aforementioned network. In one embodiment, the ambisonic representation is of order 1 and the number N of channels is four, and the rotation matrix of the current frame is represented by two quaternions. In this embodiment and in the case of an interpolation, each interpolation for a current subframe is a linear spherical interpolation (or “SLERP”), carried out as a function of the interpolation of the subframe preceding the subframe. current frame and from the quaternions of the previous subframe. For example, the linear spherical interpolation of the current subframe can be carried out to obtain the quaternions of the current subframe as follows: Or : Q L, t─1 is one of the quaternions of the previous subframe t-1, Q R, t─1 is the other of the quaternions of the previous subframe t-1, Q L, t is one of the quaternions of the current subframe t, Q R, t is the other of the quaternions of the current sub-frame t, and a corresponds to an interpolation factor. In one embodiment, the search for the eigenvectors is performed by principal component analysis (or "PCA") or by Karhunen Loeve transform (or "KLT"), in the time domain. Of course, other embodiments can be envisaged (decomposition into singular values, or others). In one embodiment, the method comprises a preliminary step of forecasting the budget for allocation of bits per ambisonic channel, comprising: - for each ambisonic channel, an estimate of current acoustic energy in the channel, - the selection in a memory of a predetermined score, of quality, according to this ambisonic channel and of a current flow in the network, the estimation of a weighting to be operated for the allocation of bits to this channel, by multiplication of the score selected at the estimated energy. This embodiment then makes it possible to manage an optimal allocation of bits to be allocated for each channel to be coded. It is advantageous as such and could possibly be the subject of separate protection. The present invention also relates to a method of decoding sound signals forming a succession in time of frames of samples, in each of N channels in ambisonic representation of order greater than 0, the method comprising: - receive, for a current frame, in addition to the signals of the N channels of this current frame, parameters of a rotation matrix, - construct an inverse rotation matrix from said parameters, - Applying said reverse rotation matrix to signals from the N received channels, before separate channel decoding of said signals. Such an embodiment also makes it possible to improve on decoding a decorrelation between the N channels. The present invention is also aimed at a coding device comprising a processing circuit for implementing the coding method presented above. It also relates to a decoding device comprising a processing circuit for implementing the above decoding method. It also relates to a computer program comprising instructions for implementing the above method, when these instructions are executed by a processor of a processing circuit. It also relates to a non-transient memory medium storing the instructions of such a computer program. Other advantages and characteristics and characteristics of the invention will become apparent on reading the exemplary embodiments presented in the detailed description below, and on examining the appended drawings in which: - Figure 1 illustrates multi-mono coding (state of the art), - Figure 2 illustrates a succession of main steps of an example process within the meaning of the invention, FIG. 3 shows the general structure of an example of an encoder according to the invention, FIG. 4 shows details of the analysis and the PCA / KLT transformation carried out by the block 310 of the encoder of FIG. 3, - Figure 5 shows an example of a decoder according to the invention, - Figure 6 shows the decoding and reverse PCA / KLT synthesis of Figure 4, on decoding, - Figure 7 illustrates structural embodiments of an encoder and a decoder within the meaning of the invention. The invention aims to allow an optimized coding by: - an adaptive temporal matrixing (in particular with an adaptive transformation obtained by PCA / KLT (“PCA” designating a principal component analysis and “KLT” designating a Karhunen Loeve transform), - preferentially followed by multi-mono coding. Adaptive matrixing allows more efficient channelization than fixed matrixing. The matrixing according to the invention advantageously makes it possible to decorrelate the channels before multi-mono coding, so that the coding noise introduced by the coding of each of the channels globally distorts the spatial image as little as possible when the channels are recombined to reconstruct a ambisonic signal on decoding. In addition, the invention makes it possible to ensure a gentle adaptation of the matrixing parameters in order to avoid “click” type artefacts at the edge of the frame or too rapid fluctuations in the spatial image, or even coding artefacts due to too strong variations (for example linked to untimely permutations of sound sources between channels) in the various individual channels resulting from the mastering which are then coded by different instances of a mono codec. A multi-mono coding is presented below with preferentially variable allocation of the bits between channels (after adaptive matrixing), but in variants several instances of a stereo or other core codec can be used. In order to facilitate understanding of the invention, certain explanatory concepts concerning rotations in dimension n, decompositions of the PCA / KLT or SVD type (“SVD” denoting a decomposition into singular values) are recalled below. Rotations and "quaternions" The signals are represented by successive blocks of sound samples, these blocks being called “sub-frames” below. The invention uses a representation of the rotations in dimension with parameters suitable for a quantization per frame and especially an efficient interpolation per sub-frame. We define below the representations of rotations used in dimension 2, 3 and 4. A rotation (around the origin) is a transformation of space into dimension that changes one vector to another vector, such as: - The amplitude of the vector is preserved - The cross product of vectors defining an orthonormal coordinate system before rotation is preserved after rotation (there is no reflection). A matrix M of size nxn is a rotation matrix if and only if M T .M = I n where I n designates the identity matrix of size nxn (i.e. M is a unit matrix, M T designating the transpose of M) and its determinant is +1. Several representations are used in the invention which are equivalent to the representation by rotation matrix: In two dimensions (in a 2D plane) (n = 2): We use as representation the angle of rotation as follows. Given the angle of rotation! one deduces from it the matrix of rotation: " Given a rotation matrix, we can calculate the angle! by observing that the trace of the matrix is ​​2cos !. Note that it is also possible to estimate! directly from a covariance matrix before applying a decomposition into main components (PCA) and decomposition into eigenvalues ​​(EVD) presented later. The interpolation between two rotations of respective angles q 1 and q 2 can be done by linear interpolation between q 1 and q 2 , by taking into account the constraint of shortest path on the unit circle between these two angles. In three-dimensional space (3D) (n = 3): We use as representation the Euler angles and the quaternions. In variants, it is also possible to use a representation by axis-angle which is not recalled here. A rotation matrix of size 3x3 can be decomposed into a product of 3 elementary angle rotations! along the x, y, or z axes. According to the combinations of axes, the angles are said to be Eulerian or Cardanic. Another representation of 3D rotations, however, is given by quaternions. Quaternions are a generalization of complex number representations with four components in the form of a number q = a + bi + cj + dk where i 2 = j 2 = k 2 = ijk = ‒1. The real part / is called a scalar and the three imaginary parts (0, 2, 4) form a 3D vector. The norm of a quaternion is | . | unit quaternions (of norm 1) represent rotations - however this representation is not unique; thus, if 8 represents a rotation, -8 represents the same rotation. Given a unitary quaternion q = a + bi + cj + dk (with a 2 + b 2 + c 2 + d 2 = 1), the associated rotation matrix is: Euler angles do not allow correct interpolation of 3D rotations; to do this, we use quaternions or the axis-angle representation instead. The SLERP interpolation method (for "spherical linear interpolation") consists of interpolating according to the formula: where 0 £ to £ 1 is the interpolation factor to go from q 1 to q 2 and Ω is the angle between the two quaternions: where q 1 . q 2 denotes the dot product between two quaternions (identical to the dot product between two vectors of dimension 4). This amounts to interpolating by following a large circle on a 4D sphere with a constant angular speed as a function of ^. It should be ensured that the shortest path is used for the interpolant by changing the sign of one of the quaternions when q 1 . q 2 <0. Note that other quaternion interpolation methods can be used (normalized linear interpolation or nlerp, splines,…). Note that it is also possible to interpolate 3D rotations through the axis-angle representation; in this case the angle is interpolated as in the 2D case and the axis can be interpolated for example by the SLERP method (in 3D) making sure that the short path pls is taken on a 3D unit sphere and taking into account due to the fact that the representation given by the axis G and the angle! is equivalent to that given by the axis of opposite direction - r and the angle In dimension 4 (n = 4), a rotation can be parameterized by 6 angles (n (n-1) / 2) and we show that the multiplication of two matrices of size 4x4 called quaternion (I ^ ) and antiquaternion (associated with quaternions 1 = a + bi + cj + dk and 2 = w + xi + yj + zk give a rotation matrix of size 4x4. It is possible to find the associated double quaternion (q 1 q 2 ) and associated quaternion and antiquaternion matrices such as: and Their product gives a 4x4 size matrix: and it is possible to check that this matrix satisfies the properties of a rotation matrix (unit matrix and determinant equal to 1). Conversely, given a 4x4 rotation matrix, one can factorize this matrix into a product of matrices in the form for example with the method known as "factorization of Cayley ”. This involves calculating an intermediate matrix called a "tetragonal transform" (or associated matrix) and deducing the quaternions from it up to an indeterminacy on the sign of the two quaternions (which can be removed by an additional constraint of "shortest path" mentioned further away). Singular value decomposition (or "SVD") The singular value decomposition (SVD in English) consists in factoring a real matrix A of size mxn in the form: A = USV T where U is a unit matrix (U T U = I m ) of size m × m, S is a rectangular diagonal matrix of size m × n with real and positive coefficients s i ³ 0 (i = 1… where p = min ( m, n)), V is a unit matrix (V T V = I n ) of size n × n and V T is the transpose of V. The coefficients s i in the diagonal of R are the singular values ​​of the matrix P By convention, they are generally listed in decreasing order, and in this case the diagonal matrix R associated with P is unique. The rank r of A is given by the number of non-zero coefficients s I. We can therefore rewrite the decomposition in singular values ​​as: where r = [u 1 , u 2 ,…, u r ] are the left singular vectors (or output vectors) of A, S r = diag (s 1 ,…, s r ) and V r = [v 1 , v 2 ,…, v r ] are the right singular vectors (or input vectors) of P. This matrix formulation can also be rewritten as: If the sum is limited to an index i