JP6652990B2 - Apparatus and method for surround audio signal processing - Google Patents

Description

  The present invention relates to surround audio signal processing systems, and in particular to encoding and decoding of audio signals that can be used in any digitized and compressed audio signal storage or transmission application and in rendering for audio playback applications. About

  It is desirable for the audience to have a high degree of audio envelopment when listening to music or watching video with audio, as it gives a better sense of the audio / video scene. The meaning of audio development includes immersive 3D audio and accurate audio localization. Immersive 3D audio means that the audio system can virtualize the sound source at any location in space. Accurate audio localization means that the audio system can position the sound source precisely in terms of both direction and distance with the original audio scene [1].

  The sensation of audio development can be provided by a 3D audio system, which uses multiple loudspeakers. The loudspeakers surround the audience (audience) and can be placed in high, medium and low vertical positions.

  Three types of input signals and formats are commonly used in 3D audio systems: channel-based inputs, object-based inputs, and higher-order ambisonics.

  Channel-based inputs are commonly used in today's 2D and 3D audio signal generation processing and media (eg, 22.2, 9.1, 8.1, 7.1, 5.1, etc.), and individual The generated audio signal channel is intended to directly drive the loudspeaker at the designated location.

  For object-based inputs, each generated audio signal channel is intended to be rendered at a specified spatial location, regardless of the number or location of the loudspeakers actually available. Represents the source.

  For Higher Order Ambisonics (HOA), each generated audio signal channel is part of the overall description of the entire sound scene, regardless of the number or location of loudspeakers actually available. .

  Among the three formats, the HOA format is a representation of an audio scene that can render an Ambisonic signal into any playback setup, including non-standard speaker layouts.

  In the prior art, such as a model for MPEG-H 3D audio standardization, for the HOA format, on the decoder side, the HOA signal is first reconstructed from the decoded core signal and then rendered into the speaker setup. You.

  FIG. 1 shows the decoder in the model of MPEG-H 3D audio standardization for the HOA format.

  First, the input bitstream is demultiplexed into N bitstreams originally generated by the AAC-family mono encoder and the parameters needed to reassemble the entire HOA representation from these bitstreams. (101).

  In the multi-channel perceptual decoding components (102, 103 and 104), the N bit streams are individually decoded by AAC-family mono decoders to generate N signals.

  In the subsequent spatial decoding component, first the actual value ranges of these signals are reconstructed by an inverse gain control process (105). In the next step, the N signal is redistributed to provide M predominant signals and (NM) HOA coefficient signals representing the more ambient HOA component (105).

  A fixed subset of the (NM) HOA coefficient signal is re-correlated. This is to reverse the decorrelation in the HOA coding stage (107).

  Next, all of the (NM) HOA coefficient signals are used to create an ambient HOA component (107).

  The predominant HOA component is synthesized from the M predominant signals and the corresponding parameters.

  Finally, the predominant and ambient HOA components are assembled into the desired complete HOA representation (108) and further rendered to a given loudspeaker setup (109).

  The detailed process of predominant sound synthesis, ambience synthesis, HOA composition and rendering is described below.

  In the predominant sound synthesis (PSS) block (106), the HOA representation of the predominant component is calculated from one of two methods. These methods are referred to as "direction-based" and "vector-based."

In vector-based PSS, the predominant sound is calculated from the vector-based signal X _VEC (k). The X _VEC (k) signals represent time-domain audio signals decoupled from their spatial characteristics. The reconstructed HOA coefficients are calculated by multiplying the vector-based signal X _VEC (k) by a corresponding plurality of transform vectors (represented by multiple vectors of M _VEC (k)). Thus, M _VEC (k) includes the spatial characteristics (such as directivity and width) of the corresponding X _VEC (k) time-domain audio signal. The calculation is as follows.

ここで、
Ｘ_ＶＥＣ（ｋ）は、デコードされたベクトルベースの、プレドミナントサウンドを示す。
Ｍ_ＶＥＣ（ｋ）は、ベクトルベースのプレドミナントサウンドからＨＯＡ係数を再構築するマトリクスを示す。
Ｃ_ＶＥＣ（ｋ）は、ベクトルベースのプレドミナントサウンドから再構築されたＨＯＡ係数を示す。

here,
X _VEC (k) indicates the decoded vector-based, predominant sound.
M _VEC (k) denotes the matrix that reconstructs the HOA coefficients from the vector-based predominant sound.
C _VEC (k) indicates the HOA coefficient reconstructed from the vector-based predominant sound.

For direction-based PSS, the HOA coefficients are calculated from all direction-based predominant sound signals X _PS (k). Using the tuple set M _DIR (k), the calculation is as follows.

ここで、
Ｘ_ＰＳ（ｋ）は、デコードされた方向ベースの、プレドミナントサウンドを示す。
Ｍ_ＤＩＲ（ｋ）は、方向ベースのプレドミナントサウンドからＨＯＡ係数を再構築するマトリクスを示す。
Ｃ_ＤＩＲ（ｋ）は、方向ベースのプレドミナントサウンドから再構築されたＨＯＡ係数を示す。

here,
_XPS (k) indicates the decoded direction-based, predominant sound.
M _DIR (k) indicates the matrix that reconstructs the HOA coefficients from the direction-based predominant sound.
C _DIR (k) indicates the HOA coefficient reconstructed from the direction-based predominant sound.

In ambience synthesis, the ambient HOA component frame _CAMB (k) is obtained as follows according to reference [2].

ここで、
Ｏ_ＭＩＮは、アンビエントＨＯＡ係数の最小数を示す。
Ψ_ＭＩＮは、ある固定の所定方向に関するモードマトリクスを示す。
ｃ_{Ｉ，ＡＭＢ，ｎ}（ｋ）は、デコードされたアンビエントサウンド信号を示す。 1) The first O _MIN coefficient of the ambient HOA component is obtained below.

here,
O _MIN indicates the minimum number of ambient HOA coefficients.
Ψ _MIN indicates a mode matrix related to a fixed predetermined direction.
c _{I, AMB, n} (k) indicates the decoded ambient sound signal.

2) Sample values of the remaining coefficients of the ambient HOA component are calculated according to:

（方向ベースの合成に対するもの）

（ベクトルベースの合成に対するもの）
ここで、
Ｃ_ＶＥＣ（ｋ）は、ベクトルベースのプレドミナントサウンドから再構築されたＨＯＡ係数を示す。
Ｃ_ＤＩＲ（ｋ）は、方向ベースのプレドミナントサウンドから再構築されたＨＯＡ係数を示す。
Ｃ_ＡＭＢ（ｋ）は、アンビエント信号から再構築されたＨＯＡ係数を示す。
Ｃ（ｋ）は、最終的な再構築されたＨＯＡ係数を示す。 Finally, within the HOA composition, the ambient HOA component and the predominant HOA component are superimposed to provide a decoded HOA frame. If prediction is not working for direction-based predominant synthesis, the decoded HOA frame C (k) is calculated by:

(For direction-based compositing)

(For vector-based compositing)
here,
C _VEC (k) indicates the HOA coefficient reconstructed from the vector-based predominant sound.
C _DIR (k) indicates the HOA coefficient reconstructed from the direction-based predominant sound.
_CAMB (k) indicates the HOA coefficient reconstructed from the ambient signal.
C (k) indicates the final reconstructed HOA coefficient.

ここで、
Ｃ（ｋ）は、最終的な再構築されたＨＯＡ係数を示す。
Ｗ（ｋ）は、ラウドスピーカ信号を示す。
Ｄｈａ、レンダリングマトリクスを示す。 If no short-range compensation is applied, the decoded HOA coefficients C (k) are converted to a representation of the loudspeaker signal W (k) by multiplication by a rendering matrix D.

here,
C (k) indicates the final reconstructed HOA coefficient.
W (k) indicates a loudspeaker signal.
Dha shows a rendering matrix.

The following notes are provided to calculate the complexity of the above process.
1) The order of the HOA signal is _OHOA , and the number of HOA coefficients is ( _OHOA + 1) ² .
2) The number of playback speakers is L.
3) The total number of core signal channels is N.
4) The number of predominant sound channels is M.
5) The number of ambient sound channels is NM.

ここで、
ＣＯＭ_ＰＳＳは、プレドミナントサウンド合成のための演算量を示す。
Ｍは、プレドミナントサウンドチャネルの数を示す。
Ｏ_ＨＯＡは、ＨＯＡのオーダを示す。
Ｆ_ｓは、サンプリング周波数を示す。 Complexity for predominant sound synthesis is

here,
COM _PSS indicates the amount of calculation for predominant sound synthesis.
M indicates the number of predominant sound channels.
O _HOA indicates the order of HOA.
F _s represents the sampling frequency.

ここで、
ＣＯＭ_{ＲＥＮＤＥＲ}は、レンダリングのための演算量を示す。
Ｌは、再生スピーカの数を示す。
Ｏ_ＨＯＡは、ＨＯＡのオーダを示す。
Ｆ_ｓは、サンプリング周波数を示す。 The amount of computation for rendering is

here,
COM _RENDER indicates the amount of calculation for rendering.
L indicates the number of reproduction speakers.
O _HOA indicates the order of HOA.
F _s represents the sampling frequency.

The number of HOA coefficients is very large in a normal HOA format, for example, if _OHOA = 4, the number of HOA coefficients is (4 + 1) ² = 25.

  Also, to have a better feeling of 3D audio, the number of playback channels is also very large, for example, a 22.2 setup has a total of 24 speakers.

  The sampling frequency for audio signals is typically 44.1 kHz or 48 kHz.

As an example, for M = 4, _OHOA = 4, L = 24 and Fs = 48 kHz, the computational complexity for predominant sound synthesis and rendering is estimated as

  The examples show that both the compositing and rendering processes are very complex, and it is therefore desirable to reduce complexity (computation).

（ベクトルベースの合成に対するもの）

（方向ベースの合成に対するもの） As shown in the HOA composition process (Equations (1) and (2)), predominant sound synthesis is performed according to the following.

(For vector-based compositing)

(For direction-based compositing)

Ambient sound synthesis is performed according to the following.

The rendering is performed according to (Equation (7)).

（ベクトルベースの合成に対するもの） The HOA composition and rendering processes are combined into one process of channel conversion ^* .

(For vector-based compositing)

（方向ベースの合成に対するもの）

(For direction-based compositing)

As an example, for O _HOA = 4, M = 4, N = 8, L = 24 and Fs = 48 kHz, estimating the amount of computation for predominant sound synthesis and rendering:

  From the above example, by implementing the idea of the present invention, the amount of computation can be greatly reduced.

  In the MPEG-H 3D audio model, there is a prediction component for part of the input sequence, and near-range compensation before rendering for some conditions. The invention is not adapted to the conditions when a prediction component is present or when near field compensation is performed.

  In the MPEG-H 3D audio model, the calculation of the HAO representation from the directional signal is based on the concept of overlap-add to avoid artifacts due to directional changes between successive frames (for direction-based synthesis).

Thus, the HOA representation C _DIR (k) of the active direction signal is calculated as the sum of the fade-out and fade-in components.

Which poses challenges for the method of the invention when fading in and out in the HOA domain. In order to solve this problem, the following ideas are conceived.
_{1) X 'PS (k-} 1) = X PS (k-1) w out; X' to define the _{PS (k) = X PS (} k) w in.
2) Modify equation (11) as follows:

  The above principles can be applied to vector-based compositing if fade-in and fade-out are made in the HOA domain for vector-based compositing.

If fade-in and fade-out are done in the vector domain for vector-based compositing, then:
_{1) X 'VEC (k)} = w out X VEC (k-1) + w in defining _{X VEC} a (k).
2) Modify equation (10) as follows:

FIG. 1 is a decoder diagram of the HOA input MPEG-H 3D audio standard. FIG. 2 is a decoder diagram according to the first embodiment of the present invention. FIG. 3 is a decoder diagram according to the second embodiment of the present invention. FIG. 4 is a decoder diagram according to the third embodiment of the present invention. FIG. 5 is a decoder diagram according to the fourth embodiment of the present invention. FIG. 6A is a decoder diagram according to the fifth embodiment of the present invention. FIG. 6B is another decoder diagram according to the fifth embodiment of the present invention. FIG. 7A is a decoder diagram according to Embodiment 6 of the present invention. FIG. 7B is another decoder diagram according to Embodiment 6 of the present invention. FIG. 8 shows an example of a bit stream according to the seventh embodiment of the present invention. FIG. 9 is a decoder diagram according to the seventh embodiment of the present invention. FIG. 10 is an encoder diagram according to the eighth embodiment of the present invention. FIG. 11 is an encoder diagram according to Embodiment 9 of the present invention. FIG. 12 is an encoder diagram according to the tenth embodiment of the present invention.

  The following embodiments are merely examples for various inventive steps principles. Of course, variations on the detailed description herein will be apparent to those skilled in the art. Those skilled in the art can modify and apply the present invention without departing from the spirit of the present invention.

1. Embodiment 1
As a first embodiment of the present invention, a surround sound decoder according to the present invention includes: a bit stream demultiplexer that decompresses a bit stream into spatial parameters and core parameters; and a set of core decoders that decodes core parameters into a set of core signals. A matrix deriving unit that derives a rendering matrix from the spatial parameters and the layout of the playback speakers; and a renderer that renders the decoded core signal into a playback signal using the rendering matrix.

  FIG. 2 shows the above-described decoder according to the first embodiment.

  The bitstream demultiplexer (200) decompresses the bitstream into spatial and core parameters.

  The set of core decoders (201, 202, 203) decodes the core parameters into a set of core signals, and the decoders are MPEG-1 Audio Layer III, AAC, HE-AAC, Dolby AC-3, MPEG USAC standard, etc. Any existing or new codec.

The matrix deriving unit (204) calculates a rendering matrix from the spatial parameters and the layout of the reproduction speakers. Rendering may be derived using some or all of the following parameters.
The number of target speakers (5.1, 7.1, 10.1, or 22.2 ...),
Speaker position (distance from sweet spot, horizontal angle and elevation angle),
Position of spherical modeling (horizontal and elevation),
HOA order (first order (HOA coefficient of 4), second order (HOA coefficient of 9) or third order (16 HOA coefficients ...)), and
HOA decomposition parameters (direction-based decomposition or PCA or SVD).

  Such as VBAP (vector-based amplitude panning) [3], or DBAP (direction-based amplitude panning) [4], or a method described in the published reference model for MPEG-H 3D for HOA format [2]. There are techniques available to derive a rendering matrix from the reconstructed input signal to the desired speaker layout.

  As an example, if the input signal is a fourth-order HOA, then the playback speaker setup is a standard 22.2 channel setup, with 25 HOA coefficients to cover 25 directions of spherical space. The rendering matrix maps 25 HOA coefficients to 24 speaker channels.

ここで、
ｐは、ＨＯＡ球面方向を示す。
ｌ_ｎは、ラウドスピーカベクトルを示す。
ｇ_ｎは、ｌ_ｎに適用される倍率を示す。
｛ｎ_１，ｎ_２，ｎ_３｝は、アクティブのラウドスピーカの三重項を示す。 When VBAP is used to derive the rendering matrix, VBAP is composed of 24 unit vectors l,. . . , L ₂₄ , a triangular mesh is formed between the loudspeakers. For each of the 25 HOA spherical directions p, it is in one of the triangles formed by the speakers. The three loudspeakers forming the triangle are selected to be active loudspeakers, and the spherical direction p can be calculated by a linear combination of the loudspeakers.

here,
p indicates the HOA spherical surface direction.
l _n denotes a loudspeaker vector.
g _n denotes a magnification to be applied to _{l n.}
{N ₁ , n ₂ , n ₃ } denotes the triplet of the active loudspeaker.

ここで、
ｐは、ＨＯＡ球面方向を示す。
ｌ_ｎは、ラウドスピーカベクトルを示す。
ｇ_ｎは、ｌ_ｎに適用される倍率を示す。
｛ｎ_１，ｎ_２，ｎ_３｝は、アクティブのラウドスピーカの三重項を示す。 In R _3, the vector space is formed by the third vector base. This leads to the following solution:

here,
p indicates the HOA spherical surface direction.
l _n denotes a loudspeaker vector.
g _n denotes a magnification to be applied to _{l n.}
{N ₁ , n ₂ , n ₃ } denotes the triplet of the active loudspeaker.

  The above procedure is repeated for all 25 HOA spherical directions, and all gain parameters for each spherical direction can be derived to form a rendering matrix D.

ここで、
Ｃ’（ｋ）は、完全再構築されたオーディオ信号を示す。
Ｗ（ｋ）は、ラウドスピーカ信号を示す。
Ｄは、レンダリングマトリクスを示す。 The rendering from HOA coefficients to loudspeaker outputs can be described by the following equation:

here,
C ′ (k) indicates a completely reconstructed audio signal.
W (k) indicates a loudspeaker signal.
D indicates a rendering matrix.

ここで、
Ｃ’（ｋ）は、完全再構築されたオーディオ信号を示す。
Ｓ’（ｋ）は、デコードされた信号を示す。
Ｍは、変換マトリクスを示す。 However, in the present invention, a completely reconstructed audio signal is not available. Assume that the reconstructed audio signal can be derived according to the following equation:

here,
C ′ (k) indicates a completely reconstructed audio signal.
S ′ (k) indicates a decoded signal.
M indicates a conversion matrix.

ここで、
Ｃ’（ｋ）は、完全再構築されたオーディオ信号を示す。
Ｗ（ｋ）は、ラウドスピーカ信号を示す。
Ｄは、レンダリングマトリクスを示す。
Ｍは、変換マトリクスを示す。
Ｄ’は、新しいレンダリングマトリクスを示す。 The following is obtained by combining the equations (17) and (18).

here,
C ′ (k) indicates a completely reconstructed audio signal.
W (k) indicates a loudspeaker signal.
D indicates a rendering matrix.
M indicates a conversion matrix.
D 'indicates a new rendering matrix.

  In addition to the approaches described above, it is possible to derive a rendering matrix directly using the decoded core signal and speaker layout information.

  The above procedures and formulas are provided as examples of how to implement the present invention, and those skilled in the art will be able to modify and apply the present invention without departing from the spirit of the invention. .

  Finally, the rendering unit (205) renders the decoded core signal into a reproduction signal using a rendering matrix.

  Effect: In this embodiment, the surround sound signal is reconstructed and rendered into a desired speaker layout in a single step, which improves efficiency and greatly reduces the amount of computation.

2. Embodiment 2
A surround sound decoder according to the present invention comprises: a bitstream demultiplexer for decompressing a bitstream into predominant sound parameters, ambience parameters, channel allocation parameters, and core parameters; and a set of core decoders for decoding the core parameters into a set of core signals. A predominant sound ambience switch for assigning the decoded core signal to the predominant sound and ambience according to the channel assignment parameter; and a matrix deriving unit for deriving a predominant sound rendering matrix from the predominant sound parameters and the layout of the playback speakers And ambience rendering matrices based on the ambience parameters and the layout of the playback speakers A pre-dominant sound renderer that renders predominant sound into a playback signal using a rendering matrix; an ambience renderer that renders ambience into a playback signal using a rendering matrix; An output signal configuration unit that configures a reproduction signal using the obtained predominant sound and ambient sound.

  FIG. 3 shows the above-described decoder according to the second embodiment.

  The bitstream demultiplexer (300) decompresses the bitstream into predominant sound parameters, ambience parameters, channel assignment parameters, and core parameters.

  The set of core decoders (301, 302, 303) decodes the core parameters into a set of core signals, and the decoders may be MPEG-1 Audio Layer III, AAC, HE-AAC, Dolby AC-3, MPEG USAC standard, etc. Any existing or new codec.

  A predominant sound / ambience switch (304) assigns the decoded core signal to predominant sound or ambience according to channel assignment parameters.

  A rendering matrix calculation unit (305) calculates a rendering matrix from the predominant sound parameters and the layout of the playback speakers. In the present embodiment, detailed derivation is omitted, and it is assumed that the rendering matrix derived from the predominant sound is D '.

ただし、
Ｗ_ｐｓ（ｋ）は、プレドミナントサウンドから導出された再生信号を示す。
Ｃ_ｐｓ（ｋ）は、デコードされたプレドミナントサウンド信号を示す。
Ｄ’は、ＰＳレンダリングマトリクスを示す。 The predominant sound renderer (306) converts the decoded predominant sound into a reproduction signal using the PS rendering matrix.

However,
_Wps (k) indicates a reproduced signal derived from the predominant sound.
C _ps (k) indicates the decoded predominant sound signal.
D 'indicates a PS rendering matrix.

A rendering matrix calculation unit (307) calculates a rendering matrix from the ambience parameters and the layout of the playback speakers. In the present embodiment, detailed derivation is omitted, and it is assumed that the rendering matrix derived from the ambient sound is _DAMB .

  If the ambient sound was converted to some other format or otherwise processed before encoding, the signal may be post-processed to reconstruct the original ambient sound before rendering. .

ただし、
Ｗ_AMB（ｋ）は、アンビエントサウンドから導出された再生信号を示す。
Ｃ_ＡＭＢ（ｋ）は、デコードされたアンビエントサウンド信号を示す。
Ｄ_ＡＭＢは、アンビエンスレンダリングマトリクスを示す。 The ambience rendering device (308) converts the decoded ambient sound into a reproduction signal using the ambience rendering matrix.

However,
W _AMB (k) indicates a reproduced signal derived from the ambient sound.
_CAMB (k) indicates the decoded ambient sound signal.
_DAMB indicates an ambience rendering matrix.

ただし、
Ｗ_AMB（ｋ）は、アンビエントサウンドから導出された再生信号を示す。
Ｗ_ｐｓ（ｋ）は、プレドミナントサウンドから導出された再生信号を示す。
Ｗ（ｋ）は、最終的な再生信号を示す。 The output signal composition unit composes a reproduced signal using the rendered predominant sound and ambient sound.

However,
W _AMB (k) indicates a reproduced signal derived from the ambient sound.
_Wps (k) indicates a reproduced signal derived from the predominant sound.
W (k) indicates a final reproduced signal.

  Effect: In this embodiment, the predominant sound signal is reconstructed and rendered to the desired speaker layout in one single step, which improves efficiency and greatly reduces the amount of computation. .

3. Embodiment 3
A surround sound decoder according to the present invention comprises: a bitstream demultiplexer for decompressing a bitstream into spatial parameters and core parameters; a set of core decoders for decoding core parameters into a set of core signals; a layout of spatial parameters and playback speakers. And a windowing unit for performing windowing on the decoded core signals of the previous frame and the current frame; and a decoded core signal of the windowed previous frame and the windowed current. A summation unit for summing the decoded core signal of the frame with the derived smoothed core signal; and rendering the smoothed core signal into a reproduced signal using a rendering matrix. And; including.

  It is common to apply windowing in audio signal processing to avoid artificial sounds across frame boundaries.

ここで、
Ｃ’（ｋ）は、完全再構築されたオーディオ信号を示す。
Ｓ’（ｋ）は、現フレームに対するデコードされた信号を示す。
Ｓ’（ｋ−１）は、前フレームに対するデコードされた信号を示す。
ｗｉｎ_ｃｕｒは、現フレームに対するウインドウイング関数を示す。
ｗｉｎ_ｐｒｅは、前フレームに対するウインドウイング関数を示す。
Ｍは、変換マトリクスを示す。 As shown in FIG. 4, windowing is applied to the decoded core signal (404), and equations (17) and (18) are modified as follows.

here,
C ′ (k) indicates a completely reconstructed audio signal.
S ′ (k) indicates a decoded signal for the current frame.
S ′ (k−1) indicates a decoded signal for the previous frame.
win _cur indicates a windowing function for the current frame.
win _pre indicates a windowing function for the previous frame.
M indicates a conversion matrix.

ここで、
Ｓ’（ｋ）は、現フレームに対するデコードされた信号を示す。
Ｓ’（ｋ−１）は、前フレームに対するデコードされた信号を示す。
ｗｉｎ_ｃｕｒは、現フレームに対するウインドウイング関数を示す。
ｗｉｎ_ｐｒｅは、前フレームに対するウインドウイング関数を示す。
Ｗ（ｋ）は、ラウドスピーカ信号を示す。
Ｄ’は、レンダリングマトリクスを示す。

here,
S ′ (k) indicates a decoded signal for the current frame.
S ′ (k−1) indicates a decoded signal for the previous frame.
win _cur indicates a windowing function for the current frame.
win _pre indicates a windowing function for the previous frame.
W (k) indicates a loudspeaker signal.
D 'indicates a rendering matrix.

  Effect: In this embodiment, windowing is applied to avoid artificial sounds across frame boundaries.

4. Embodiment 4
A surround sound decoder according to the present invention comprises: a bitstream demultiplexer for decompressing a bitstream into predominant sound parameters, ambience parameters, channel allocation parameters, and core parameters; and a set of core decoders for decoding the core parameters into a set of core signals. A predominant sound ambience switch for assigning the decoded core signal to the predominant sound and ambience according to the channel assignment parameter; and a matrix deriving unit for deriving a predominant sound rendering matrix from the predominant sound parameters and the layout of the playback speakers And ambience rendering matrices based on the ambience parameters and the layout of the playback speakers A window deriving unit for deriving a pre-dominant sound signal of a previous frame and a current frame; and a windowing unit for performing a windowing process on a pre-dominant sound signal of a previous frame and a current frame; A dominant sound renderer; and an output signal composing unit for composing a reproduction signal using the rendered predominant sound and ambience sound.

  As shown in FIG. 5, windowing is applied to the predominant sound (506) to ensure a continuous and flat occurrence of the sound field over the frame boundaries.

ただし、
Ｗ_ｐｓ（ｋ）は、プレドミナントサウンドから導出された再生信号を示す。
Ｃ_ｐｓ（ｋ）は、現フレームに対するデコードされたプレドミナントサウンド信号を示す。
Ｃ_ｐｓ（ｋ−１）は、前フレームに対するデコードされたプレドミナントサウンド信号を示す。
Ｄ’は、ＰＳレンダリングマトリクスを示す。 Since windowing is applied to the predominant sound, equation (20) is modified as follows:

However,
_Wps (k) indicates a reproduced signal derived from the predominant sound.
C _ps (k) indicates the decoded predominant sound signal for the current frame.
_Cps (k-1) indicates the decoded predominant sound signal for the previous frame.
D 'indicates a PS rendering matrix.

  Effect: In this embodiment, windowing is applied to ensure a continuous and flat occurrence of the sound field over the frame boundaries.

5. Embodiment 5
As shown in FIG. 6A, a surround sound decoder according to the present invention includes a bit stream demultiplexer that decompresses a bit stream into spatial parameters and core parameters; and a core decoder (601, 602) that decodes core parameters into a set of core signals. And 603); a matrix deriving unit (604) for deriving a rendering matrix for the decoded signal of the current frame from the spatial parameters and the layout of the reproduction speakers; and a decoding of the current frame using the rendering matrix. A windowing and rendering unit (605) for performing windowing and rendering on the decoded core signal of the previous frame using the rendering matrix. Wing and render the windowing and rendering unit to perform (606); summing unit to form the final reproduced signal by adding the reproduced signal of the reproduced signal and the current frame the previous frame and (607); including.

  It is common to apply windowing in audio signal processing to avoid artificial sounds across frame boundaries.

  In the first embodiment, since the decoded core signals of the previous frame and the current frame have different spatial directions, if the windowing cannot be applied to the decoded core signal, the windowing is performed by using the reconstructed HOA coefficients. Must be applied to

ただし、
Ｓ’（ｋ）は、現フレームに対するデコードされた信号を示す。
Ｓ’（ｋ−１）は、前フレームに対するデコードされた信号を示す。
Ｓ’’（ｋ）は、現フレームに対するウインドウイングされた信号を示す。
Ｓ’’（ｋ−１）は、前フレームに対するウインドウイングされた信号を示す。
ｗｉｎ_ｃｕｒは、現フレームに対するウインドウイング関数を示す。
ｗｉｎ_ｐｒｅは、前フレームに対するウインドウイング関数を示す。
Ｗ（ｋ）は、ラウドスピーカ信号を示す。
Ｄ’_ｃｕｒは、現フレームに対する新しいレンダリングマトリクスを示す。
Ｄ’_ｐｒｅは、前フレームに対する新しいレンダリングマトリクスを示す。
Ｃ’（ｋ）は、現フレームに対する、完全再構築されたオーディオ信号を示す。
Ｃ’（ｋ−１）は、前フレームに対する、完全再構築されたオーディオ信号を示す。
Ｄは、レンダリングマトリクスを示す。
Ｍ_ｃｕｒは、現フレームに対する変換マトリクスを示す。
Ｍ_ｐｒｅは、前フレームに対する変換マトリクスを示す。 Equation (18) is then modified as follows:

However,
S ′ (k) indicates a decoded signal for the current frame.
S ′ (k−1) indicates a decoded signal for the previous frame.
S ″ (k) indicates the windowed signal for the current frame.
S ″ (k−1) indicates a windowed signal for the previous frame.
win _cur indicates a windowing function for the current frame.
win _pre indicates a windowing function for the previous frame.
W (k) indicates a loudspeaker signal.
D' _cur indicates the new rendering matrix for the current frame.
D' _pre indicates a new rendering matrix for the previous frame.
C ′ (k) indicates a completely reconstructed audio signal for the current frame.
C ′ (k−1) indicates a completely reconstructed audio signal for the previous frame.
D indicates a rendering matrix.
M _cur indicates a conversion matrix for the current frame.
M _pre indicates a transformation matrix for the previous frame.

  As shown in FIG. 6A, windowing and rendering are first performed independently (605 and 606) on the decoded core signal of the current frame and the decoded core signal of the previous frame, followed by the previous frame. And the current frame's rendered signal are added together to form the final output (607).

  For windowing & rendering on the decoded core signal of the previous frame, if the rendering matrix of the previous frame is available / stored, it can be picked up from the calculation of the previous frame. If not available / unstored, the rendering matrix can be calculated according to the same manner as (604), but using the spatial parameters and speaker layout information of the previous frame.

  Another method is shown in FIG. 6B. First, rendering is performed on the decoded signal of the current frame (615), followed by windowing on the rendered signal of the previous frame and the rendered signal of the current frame, and finally, The inked previous frame rendered signal and the current frame rendered signal are added together to form the final output (616).

  Effect: In this embodiment, windowing is applied to avoid artificial sounds across frame boundaries.

6. Embodiment 6
As shown in FIG. 7A, a surround sound decoder according to the present invention includes a bitstream demultiplexer (700) for decompressing a bitstream into predominant sound parameters, ambience parameters, channel allocation parameters, and core parameters; A set of core decoders (701, 702 and 703) for decoding into a set of signals; a predominant sound ambience switch (704) for assigning the decoded core signal to predominant sound and ambience according to channel assignment parameters; Deriving a predominant sound rendering matrix for the predominant sound signal of the current frame from the sound parameters and the layout of the playback speakers A window deriving unit (705) that performs windowing and rendering on the predominant sound signal of the current frame; and a window that performs windowing and rendering on the predominant sound signal of the previous frame. An adding unit (708) for adding a rendered predominant sound of the previous frame and a predominant sound of the current frame to form a rendered predominant sound; and an ambience parameter and a reproduction speaker. A matrix deriving unit (709) for deriving an ambience rendering matrix from the layout of Ambience renderer to render Nsu playback signal (710); including; using the rendered pre-dominant sound and ambient sound, an output signal composing units that constitute the reproduced signal (711).

  In the second embodiment, since the predominant sound signal of the previous frame and the current frame have different spatial directions, if windowing cannot be applied to the decoded predominant sound signal, the windowing is applied to the reconstructed HOA coefficients. Must be applied.

ただし、
Ｃ’_ＰＳ（ｋ）は、現フレームに対するデコードされたプレドミナントサウンド信号を示す。
Ｃ’_ＰＳ（ｋ−１）は、前フレームに対するデコードされたプレドミナントサウンド信号を示す。
Ｃ’’_ＰＳ（ｋ）は、現フレームに対するウインドウイングされたプレドミナントサウンド信号を示す。
Ｃ’’_ＰＳ（ｋ−１）は、前フレームに対するウインドウイングされたプレドミナントサウンド信号を示す。
ｗｉｎ_ｃｕｒは、現フレームに対するウインドウイング関数を示す。
ｗｉｎ_ｐｒｅは、前フレームに対するウインドウイング関数を示す。
Ｗ_ＰＳ（ｋ）は、プレドミナントサウンドからのラウドスピーカ信号を示す。
Ｄ’_ｃｕｒは、現フレームに対する新しいレンダリングマトリクスを示す。
Ｄ’_ｐｒｅは、前フレームに対する新しいレンダリングマトリクスを示す。
Ｃ’（ｋ）は、現フレームに対する、再構築されたオーディオ信号を示す。
Ｃ’（ｋ−１）は、前フレームに対する、再構築されたオーディオ信号を示す。
Ｄは、レンダリングマトリクスを示す。
Ｍ_ｃｕｒは、現フレームに対する変換マトリクスを示す。
Ｍ_ｐｒｅは、前フレームに対する変換マトリクスを示す。 Equation (19) is then modified as follows:

However,
C ′ _PS (k) indicates a decoded predominant sound signal for the current frame.
C ′ _PS (k−1) indicates a decoded predominant sound signal for the previous frame.
C ″ _PS (k) indicates the windowed predominant sound signal for the current frame.
C ″ _PS (k−1) indicates the windowed predominant sound signal for the previous frame.
win _cur indicates a windowing function for the current frame.
win _pre indicates a windowing function for the previous frame.
W _PS (k) indicates the loudspeaker signal from the predominant sound.
D' _cur indicates the new rendering matrix for the current frame.
D' _pre indicates a new rendering matrix for the previous frame.
C ′ (k) indicates the reconstructed audio signal for the current frame.
C ′ (k−1) indicates the reconstructed audio signal for the previous frame.
D indicates a rendering matrix.
M _cur indicates a conversion matrix for the current frame.
M _pre indicates a transformation matrix for the previous frame.

  As shown in FIG. 7A, windowing and rendering are first performed independently (706 and 707) with respect to the decoded predominant sound signal of the current frame and the decoded predominant sound signal of the previous frame, Subsequently, the rendered signal of the previous frame and the rendered signal of the current frame are added together to form the final predominant sound output (708).

  For windowing & rendering for the predominant sound of the previous frame, it is possible to pick up from the calculation of the previous frame if the PS matrix of the previous frame is available / stored. If not available / not stored, the PS rendering matrix can be calculated according to the same manner as (705), but using the previous previous frame's spatial parameters and speaker layout information.

  Another method is shown in FIG. 7B. First, rendering is performed on the decoded predominant sound signal (716) of the current frame, followed by windowing on the rendered signal of the previous frame and the rendered signal of the current frame. Finally, the rendered signal of the windowed previous frame and the rendered signal of the current frame are added together to form the final predominant sound output (717).

  Effect: In this embodiment, windowing is applied to ensure a continuous and flat occurrence of the sound field over the frame boundaries.

7. Embodiment 7
A surround sound decoder according to the present invention comprises: a bitstream demultiplexer for decompressing a bitstream into rendering flags, predominant sound parameters, ambience parameters, channel allocation parameters, and core parameters; a core for decoding core parameters into a set of core signals. A set of decoders; a predominant sound ambience switch for assigning the decoded core signal to predominant sound and ambience according to channel assignment parameters; and a predominant sound parameter and a playback speaker using a calculation method specified by a rendering flag. A matrix deriving unit for deriving a predominant sound rendering matrix from the layout of the ambience; A matrix deriving unit for deriving an ambience rendering matrix from the parameters and the layout of the reproduction speakers; a predominant sound renderer for rendering a predominant sound into a reproduction signal using the rendering matrix; and reproducing the ambience using the rendering matrix. An ambience renderer for rendering a signal; and an output signal composition unit for composing a reproduced signal using the rendered predominant sound and ambient sound.

  In this embodiment, the bitstream has a rendering flag that indicates whether there is any other data in the bitstream that makes the implementation of the invented idea impractical.

  FIG. 8 shows one bit stream as an example.

  When the bitstream contains only PS parameter data, ambience parameter data, channel allocation parameter data, and core coder data, it is recommended to use the invented idea to achieve low computational complexity and rendering. Therefore, the rendering flag LC_RENDER_FLAG is set to 1.

  When the bitstream contains prediction data and near field compensation data, it is impractical to use the invented idea and it is recommended to use conventional decoding, construction and rendering tools, thus: The rendering flag LC_RENDER_FLAG is set to 0.

  FIG. 9 shows the aforementioned decoder of this embodiment.

  The bitstream demultiplexer (901) decompresses the bitstream into LC_RENDER_FLAG and other parameters.

  If LC_RENDER_FLAG is equal to 1, the decoder (902) of the present invention is selected to perform decoding, configuration and rendering to complete a low complexity solution.

  If LC_RENDER_FLAG is equal to 0, the conventional decoder (903) is selected to perform decoding, composition and rendering.

  Effect: This embodiment solves the problem of bitstream incompatibility.

8. Embodiment 8
In this embodiment, the encoder comprises: a spatial encoder that analyzes the input signal and encodes the input signal into spatial parameters and an N-generated signal; a set of core encoders that encodes the N-generated signal into a set of core parameters; And a bitstream multiplexer that packs core parameters into the bitstream.

  A surround sound decoder according to the present invention comprises: a bitstream demultiplexer for decompressing a bitstream into spatial parameters and core parameters; a set of core decoders for decoding core parameters into a set of core signals; a layout of spatial parameters and playback speakers. And a rendering unit that derives a rendering matrix from the decoded core signal into a reproduction signal using the rendering matrix.

  FIG. 10 shows the above-described encoder and decoder of this embodiment.

  The spatial encoder (1001) analyzes the input signal and encodes the input signal into a spatial parameter and an N generated signal.

  Spatial encoding determines how many sound sources or audio objects are present in an input audio scene based on an analysis of the audio scene and determines how to extract and encode the sound sources or audio objects. I can do it. As an example, principal component analysis (PCA) may be used to extract sound sources or audio objects, and N sound sources may be extracted and encoded. During this process, PCA parameters and N audio signals are derived. The PCA parameters and the N-generated audio signal are encoded and sent to the decoder side.

ここで、
Ｃ（ｋ）は、インプットオーディオ信号を示す。
Ｓ（ｋ）は、生成されたオーディオ信号を示す。
Ｍは、変換マトリクスを示す。 The generated signal may be derived according to the following equation:

here,
C (k) indicates an input audio signal.
S (k) indicates the generated audio signal.
M indicates a conversion matrix.

  The set of core encoders (1002, 1003, 1004) encode the N generated signal into a set of core parameters, but the encoder uses MPEG-1 Audio Layer III, AAC, HE-AAC, Dolby AC-3, and MPEG USAC standards. And any existing or new codec.

  The bitstream multiplexer (1005) packs the spatial and core parameters into a bitstream.

  The corresponding decoder may be the decoder shown in FIG.

9. Embodiment 9
In a ninth embodiment of the present invention, an encoder analyzes an input signal and encodes the input signal into a plurality of predominant sounds and ambience sounds, and further into corresponding predominant sound parameters and ambience parameters. An audio scene analysis and spatial encoder; a channel allocation unit that allocates a core decoder to encode predominant and ambience sounds; and encodes both predominant and ambience sounds into a set of core parameters, N A set of core encoders that encode channel audio signals; and pre-dominant sound parameters, ambience parameters, channel assignment information, and core parameters. Including; a bit stream multiplexer pack bets stream.

  A surround sound decoder according to the present invention comprises: a bitstream demultiplexer for decompressing a bitstream into predominant sound parameters, ambience parameters, channel allocation parameters, and core parameters; and a set of core decoders for decoding the core parameters into a set of core signals. A predominant sound ambience switch for assigning the decoded core signal to the predominant sound and ambience; a matrix deriving unit for deriving a rendering matrix of the predominant sound from the predominant sound parameters and the layout of the playback speakers; and an ambience parameter Matrix derivation unit that derives an ambience rendering matrix from the A predominant sound renderer that renders predominant sound into a playback signal using a rendering matrix; an ambience renderer that renders ambience into a playback signal using a rendering matrix; rendered predominant sound and ambience An output signal forming unit for forming a reproduced signal using sound.

  FIG. 11 shows the above-described encoder according to the second embodiment.

  An audio scene analysis and spatial encoder that analyzes the input signal and encodes the input signal into a plurality of predominant sounds and ambience sounds, and further into corresponding predominant sound and ambience parameters; A channel allocation unit for allocating a decoder to encode predominant and ambience sounds; and a set of core encoders for encoding N-channel audio signals, including encoding both the predominant and ambience sounds into a set of core parameters. And a video for packing predominant sound parameters, ambience parameters, channel assignment information, and core parameters into a bitstream. Including; a preparative stream multiplexer.

  The audio scene analysis and spatial encoder (1101) analyzes the input signal and encodes the input signal into a plurality of predominant sounds and ambience sounds, and further into corresponding predominant sound parameters and ambience parameters.

  Audio Scene Analysis and Spatial Encoding analyzes the audio scene, determines how many sound sources or audio objects are present in the input audio scene, and extracts and encodes the sound sources or audio objects. Is to be determined. As an example, Principal Component Analysis (PCA) may be used to extract sound sources or audio objects, and M sound sources may be extracted and encoded. During this process, PCA parameters and M predominant sound signals are derived. The PCA parameters and the M predominant audio signal are encoded and sent to the decoder side.

ここで、
Ｃ（ｋ）は、インプットオーディオ信号を示す。
Ｃ_ＰＳ（ｋ）は、生成されたオーディオ信号を示す。
Ｍは、変換マトリクスを示す。 The generated signal may be derived according to the following equation:

here,
C (k) indicates an input audio signal.
C _PS (k) indicates the generated audio signal.
M indicates a conversion matrix.

ここで、
Ｃ’（ｋ）は、プレドミナントサウンドから、再構築されるオーディオ信号を示す。
Ｃ_ＰＳ（ｋ）は、デコードされたプレドミナントサウンド信号を示す。
Ｍは、変換マトリクスを示す。 The audio scene analysis and spatial encoder may extract and encode the residue between the input signal and the composite signal from the predominant sound signal, which may be termed the ambient signal. Spatial encoding extracts an ambient signal from the difference between the input signal and the composite signal from the predominant sound signal. The synthesis of the predominant sound can be made according to the following equation:

here,
C ′ (k) indicates an audio signal reconstructed from the predominant sound.
C _PS (k) indicates the decoded predominant sound signal.
M indicates a conversion matrix.

ここで、
Ｃ’（ｋ）は、プレドミナントサウンドから、再構築されるオーディオ信号を示す。
Ｃ（ｋ）は、インプットオーディオ信号を示す。
Ｃ_ＡＭＢ（ｋ）は、アンビエンス信号を示す。 The ambient signal can be derived according to the following equation:

here,
C ′ (k) indicates an audio signal reconstructed from the predominant sound.
C (k) indicates an input audio signal.
_CAMB (k) indicates an ambience signal.

  Of all the ambient signals, it was determined which of the ambient signals was to be encoded. Ambient signals may be processed or converted to other formats so that they can be more efficiently encoded.

  The channel allocation unit (1101) allocates a core encoder and encodes predominant sound and ambience sound. Information about the selection of the sequence of ambient HOA coefficients to be transmitted, their assignment, and the assignment of the predominant sound signal to a given N channel is sent to the decoder side.

  The set of core encoders (1102, 1103, 1104) encodes the M predominant sound signal and the (NM) ambient signal into a set of core parameters, but the encoder uses MPEG-1 Audio Layer III, AAC or HE- Any existing or new codec, such as AAC, Dolby AC-3, or MPEG USAC standard may be used.

  The bitstream multiplexer (1105) packs predominant sound parameters, ambience parameters, channel assignment information, and core parameters into a bitstream.

  The corresponding decoder may be the decoder shown in FIG.

10. Embodiment 10
FIG. 12 shows the above-described encoder of this embodiment.

  The audio scene analysis and spatial encoder (1201) analyzes the input signal and encodes the input signal.

  Audio scene analysis and spatial encoding perform an analysis of the audio scene, determine whether the generated parameters are compatible with the invented idea, and reflect the determination by sending LC_RENDER_FLAG.

  Achieve low computational complexity configuration and rendering if all generated parameters, such as PS parameter data, ambience parameter data, channel assignment parameter data, and core coder data are compatible with the invented idea In order to do so, it is recommended to use the invented idea in the decoder side, so that the rendering flag LC_RENDER_FLAG is set to one.

  Unless all generated parameters are compatible with the invented idea, it is impractical to use the invented idea and use conventional decoding, construction and rendering tools to It is recommended to be used internally, so the rendering flag LC_RENDER_FLAG is set to zero.

  Effect: In this embodiment, the problem of bitstream incompatibility is solved.

Reference [1] ISO / IEC JTC1 / SC29 / WG11 / N13411 "Call for Proposals for 3D Audio"
[2] ISO / IEC JTC1 / SC29 / WG11 / N14264 “WD1-HOA Text of MPEG-H 3D Audio”
[3] V. Pulki, "Virtual Sound Source Positioning Using Vector Base Amplitude Panning," Audio Eng. Soc. , Vol. 45, 1997
[4] T.I. Lossius, P .; Baltazar, and T.W. d. l. Hogue, "DBAP-Distance-based amplitude panning," in International Computer Music Conference (ICMC). Montreal, 2009.

Claims (14)

A device for decoding a surround audio signal,
A bitstream demultiplexer that decompresses the bitstream into predominant sound parameters, ambience parameters, channel assignment parameters, core parameters, and rendering flags;
A set of core decoders for decoding core parameters into a set of core signals;
A predominant sound ambience switch for assigning the decoded core signal to predominant sound and ambience according to the channel assignment parameter;
Deriving a predominant sound rendering matrix using the predominant sound parameters and the layout information of the playback speakers,
A matrix deriving unit that derives an ambience rendering matrix using the ambience parameter and the layout information of the reproduction speaker;
A predominant sound rendering device that renders a predominant sound into a playback signal using a predominant sound rendering matrix;
An ambience rendering device that renders ambience into a playback signal using an ambience rendering matrix;
An output signal composition unit for composing a playback signal using the rendered predominant sound and the rendered ambient sound;
Including
In the matrix deriving unit, based on a value of the rendering flag,
Decide not to use the rendering matrix derivation process,
apparatus.   The apparatus according to claim 1, wherein the core decoder corresponds to an MPEG-1 Audio Layer III, AAC, HE-AAC, Dolby AC-3, or MPEG USAC standard.   The apparatus according to claim 1, wherein the surround audio signal is a high-order ambisonic signal.   The matrix derivation is performed using some or all of a parameter group consisting of the number of target speakers, speaker positions, spherical modeling positions (horizontal and elevation), HOA order, and HOA decomposition parameters. The apparatus of claim 1. The value of the rendering flag indicates that, when the bit stream includes at least short-range compensation data, the rendering process of the rendering matrix in the matrix deriving unit is not used.
The device according to claim 1. The value of the rendering flag indicates that a rendering process of a rendering matrix in the matrix deriving unit is used when the bit stream does not include at least the short-range compensation data.
The device according to claim 1. A device for encoding a surround audio signal,
A spatial encoder that encodes the input signal into a plurality of predominant sounds and corresponding predominant sound parameters, and a plurality of ambience sounds and corresponding ambience parameters based on an audio scene analysis result of the input signal;
A channel allocation unit that allocates a core decoder to encode predominant and ambience sounds;
A rendering flag determination unit that determines a rendering flag used in the decoder side;
A set of core encoders that encode the generated audio signal, including encoding both the predominant sound and the ambience sound into a set of core parameters, and rendering flags, predominant sound parameters, ambience parameters, channel assignment information, And a bitstream multiplexer that packs core parameters into a bitstream;
An apparatus, including: A method of decoding a surround audio signal,
Decompressing the bitstream into predominant sound parameters, ambience parameters, channel allocation parameters, core parameters, and rendering flags,
Decode the core parameters into a set of core signals,
Assigning the decoded core signal to predominant sound and ambience according to the channel assignment parameters,
Deriving a predominant sound rendering matrix using the predominant sound parameters and the layout information of the playback speakers,
Deriving an ambience rendering matrix using the ambience parameters and the layout information of the playback speakers,
Using the predominant sound rendering matrix, render the predominant sound into a playback signal,
Render the ambience into a playback signal using the ambience rendering matrix,
Using the rendered pre-dominant sound and the rendered ambient sound, output comprising a playback signal;
Upon deriving the matrix, based on the value of the rendering flag, determine not to use the rendering matrix derivation process,
Method. 9. The method of claim 8 , wherein the core decoder corresponds to an MPEG-1 Audio Layer III, AAC, HE-AAC, Dolby AC-3, or MPEG USAC standard. The method according to claim 8 , wherein the surround audio signal is a high-order ambisonic signal. The matrix derivation is performed using some or all of a parameter group consisting of the number of target speakers, speaker positions, spherical modeling positions (horizontal and elevation), HOA order, and HOA decomposition parameters. The method of claim 8 . The value of the rendering flag indicates that when at least the short-range compensation data is included in the bit stream, the rendering process of the rendering matrix in the matrix derivation is not used.
The method according to claim 8 . The value of the rendering flag indicates that when at least the short-range compensation data is not included in the bitstream, a rendering matrix derivation process in the matrix derivation is used.
The method according to claim 8 . A method of encoding a surround audio signal,
Encoding the input signal into a plurality of predominant sounds and corresponding predominant sound parameters, and a plurality of ambience sounds and corresponding ambience parameters based on an audio scene analysis result of the input signal;
Assign a core decoder and assign a channel to encode predominant sound and ambience sound,
Determine the rendering flag used in the decoder side,
A set of core encoders that encode the generated audio signal, including encoding both the predominant sound and the ambience sound into a set of core parameters, and rendering flags, predominant sound parameters, ambience parameters, channel assignment information, And packing core parameters into a bitstream,
Method.

Images