EP2690621A1 - Method and Apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side - Google Patents
Method and Apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side Download PDFInfo
- Publication number
- EP2690621A1 EP2690621A1 EP12305914.9A EP12305914A EP2690621A1 EP 2690621 A1 EP2690621 A1 EP 2690621A1 EP 12305914 A EP12305914 A EP 12305914A EP 2690621 A1 EP2690621 A1 EP 2690621A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- data
- matrix
- signals
- downmixing
- audio
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 230000005236 sound signal Effects 0.000 title claims abstract description 36
- 238000009877 rendering Methods 0.000 claims abstract description 20
- 239000011159 matrix material Substances 0.000 claims description 101
- 238000004458 analytical method Methods 0.000 claims description 20
- 230000000875 corresponding effect Effects 0.000 claims description 13
- 230000002123 temporal effect Effects 0.000 claims description 13
- 230000015572 biosynthetic process Effects 0.000 claims description 12
- 238000003786 synthesis reaction Methods 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 6
- 230000002596 correlated effect Effects 0.000 claims description 5
- 238000000926 separation method Methods 0.000 abstract description 4
- 230000002452 interceptive effect Effects 0.000 abstract description 2
- 238000000354 decomposition reaction Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000001276 controlling effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010348 incorporation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/008—Multichannel audio signal coding or decoding using interchannel correlation to reduce redundancy, e.g. joint-stereo, intensity-coding or matrixing
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
- G10L19/0204—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using subband decomposition
Definitions
- the invention relates to a method and to an apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side, wherein the decoder side downmixing is controlled by desired playback configuration data and/or desired object positioning data.
- Audio content providers are facing consumers with increasingly heterogeneous listening situations, e.g. home theatres, mobile audio, car and in-flight entertainment. Audio content cannot be processed by its creator or broadcaster so as to match every possible consumer listening condition, for example audio/video content played back on a mobile phone. Besides different listening conditions, also different listening experiences can be desirable, for instance in a live soccer broadcast a consumer can control his own virtual position within the sound scene of the stadium (pitch or stands), or can control the virtual position and the predominance of the commentator.
- Metadata providers can add guiding metadata to the audio content, such that consumers can control down-mix or dynamic range of selected parts of the audio signal and/or assure high speech intelligibility.
- metadata For incorporation of such metadata into existing broadcasting chains, it is important that the general audio format is not changed (legacy playback) and that only a small amount of extra information (e.g. as ancillary data) is added to the audio bit stream.
- MPEG Spatial Audio Object Coding (SAOC), ISO/IEC 23003-1:2007, MPEG audio technologies - Part 1: MPEG Surround, and ISO/IEC 23003-2:2010, MPEG audio technologies - Part 2: Spatial Audio Object Coding, deals with parametric coding techniques for complex audio scenes at bit rates normally used for mono or stereo sound coding, offering at decoder side an interactive rendering of the audio objects mixed into the audio scene.
- SAOC MPEG Spatial Audio Object Coding
- ISO/IEC 23003-1 2007
- MPEG audio technologies - Part 1 MPEG Surround
- ISO/IEC 23003-2:2010 MPEG audio technologies - Part 2: Spatial Audio Object Coding
- MPEG SAOC was developed starting from Spatial Audio Coding which is based on a 'channel-oriented' approach, by introducing the concept of audio objects, having the purpose to offer even more flexibility at receiver side. Since it is a parametric multiple object coding technique, the additional cost in terms of bit rate is limited to 2-3 kbit/s for each audio object. Although the bit rate increases with the number of audio objects, it still remains at small values in comparison with the actual audio data transmitted as a mono/stereo downmix.
- a standard MPEG SAOC architecture is illustrated in Fig. 1 .
- a variable number N of sound objects Obj.#1, Obj.#2, Obj.#3, Obj.#4 is input to an SAOC encoder 11 that provides an encoded background compatible mono or stereo downmix signal or signals with SAOC parameters and side information.
- the downmix data signal or signals and the side information data are sent as dedicated bit streams to an SAOC decoder 12.
- the receiver side processing is carried out basically in two steps: first, using the side information, SAOC decoder 12 reconstructs the original sound objects Obj.#1, Obj.#2, Obj.#3, Obj.#4. Controlled by interaction/control information, these reconstructed sound objects are re-mixed and rendered in an MPEG surround renderer 13. For example, the reconstructed sound objects are output as a stereo signal with channels #1 and #2.
- SAOC parameters are computed for every time/frequency tile and are transmitted to the decoder side:
- a Distortion Control Unit DCU can be used.
- the final rendering matrix coefficients are computed as a linear combination of user-specified coefficients and the target coefficients which are assumed to be distortion-free.
- the main drawback of the solution offered by MPEG SAOC is the limitation to a maximum of two down-mix channels. Further, the MPEG SAOC standard is not designed for 'Solo/Karaoke' type of applications, which involve the complete suppression of one or more audio objects. In the MPEG SAOC standard this problem is tackled by using residual coding for specific audio objects, thereby increasing the bit rate.
- a problem to be solved by the invention is to overcome these limitations of MPEG SAOC and to allow for adding of side information for a legacy multi-channel audio broadcasting like 5.1.
- This problem is solved by the methods disclosed in claims 1 and 2. Apparatuses that utilise these methods are disclosed in claims 3 and 4, respectively.
- the invention describes how, by adding only a small amount of extra bit rate, at decoder or receiver side a re-mixing of a broadcast audio signal is achieved using information about the actual mix of the audio signal, audio signal characteristics like correlation, level differences, and the desired audio scene rendering.
- a second embodiment shows how to determine already at encoder side the suitability of the actual multi-channel audio signal for a remix at decoder side.
- This feature allows for countermeasures (e.g. changing the mixing matrix used, i.e. how the sound objects are mixed into the different channels) if a decoder side re-mixing without perceivable artefacts, or without problems like the necessary additional transmission of the audio objects itself for a short time, is not possible.
- the invention could be used to amend the MPEG SAOC standard correspondingly, based on the same building blocks.
- the inventive encoding method is suited for downmixing spatial audio signals that can be downmixed at receiver side in a manner different from the manner of downmixing at encoder side, wherein said encoding is based on MPEG SAOC and said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said method including the steps:
- the inventive decoding method is suited downmixing spatial audio signals processed according to the encoding method in a manner different from the manner of downmixing at encoder side, wherein said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said method including the steps:
- the inventive encoding apparatus is suited for downmixing spatial audio signals that can be downmixed at receiver side in a manner different from the manner of downmixing at encoder side, wherein said encoding is based on MPEG SAOC and said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said apparatus including:
- the inventive decoding apparatus is suited for downmixing spatial audio signals processed according to the encoding method in a manner different from the manner of downmixing at encoder side, wherein said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said apparatus including:
- the inventive spatial audio object coding system with five down-mix channels facilitates a backward compatible transmission at bit rates only slighter higher (due to the extended content of the side information: OLD, IOC, DMG and DCLD ) than the bit rates for known 5.1 channel transmission.
- a number of M 5 channels containing the ambience signals and a number of L audio objects mixed over the ambience are considered.
- An example is the stadium ambience of a soccer match plus specific sound effects (ball kicks, whistle) and one or more commentators.
- M is at least '2' and L is '1' or greater.
- At decoder side it is not intended to reconstruct the audio objects, but to offer the possibility of re-mixing, attenuating, totally suppressing, and changing the position of the audio objects in the rendered audio scene.
- any time/frequency transform can be used.
- hybrid Quadrature Mirror Filter (hQMF) banks are used for better selectivity in the frequency domain.
- the spatial audio input signals are processed in non-overlapping multiple-sample temporal slots, in particular 64-samples temporal slots. These temporal slots are used for computing the perceptual cues or characteristics for every successive frame, which has a length of a fixed number of temporal slots, in particular 16 temporal slots.
- 71 frequency bands are used according to the sensitivity of the human auditory system, and are grouped into K processing bands, K having a value of '2', '3' or '4', thereby obtaining different levels of accuracy.
- the hQMF filter bank transforms in each case 64 time samples into 71 frequency samples.
- FIG.2 A basic block diagram of the inventive encoder and decoder is shown in Fig.2 and Fig.3 , respectively.
- a hybrid Quadrature Mirror Filter analysis filter bank step or stage 21 is applied to all audio input signals, e.g. ambience channels Ch.#1 to Ch.#5 and sound objects Obj.#1 to Obj.# L (at least one sound object).
- the number of ambience channels is not limited to five.
- the ambience channels are not independent sound objects but are usually correlated sound signals.
- the corresponding filter bank outputs time/frequency domain signals X, which are fed on one hand to a down-mixer step or stage 22 that multiplies them with a downmix matrix D and provides, via a hybrid Quadrature Mirror Filter synthesis filter bank step or stage 24 that has an inverse operation of the analysis filter bank, the audio channels in time domain for transmission, and on the other hand are fed to an enhanced MPEG SAOC parameter calculator step or stage 23.
- the synthesis filter bank inverses the function of the analysis filter bank.
- the enhanced SAOC parameters determine the rendering flexibility in a decoder and include, as mentioned above, Object Level Differences data OLD, Inter-Object Coherence data IOC, Downmix Gains data DMG, Downmix Channel Level Differences data DCLD, and can comprise Object Energy parameter data NRG. Except the DMG and DCLD data, the other parameters correspond to original MPEG SAOC parameters. From these side information data items a rank can be determined as described below in a rank calculator step or stage 25 (i.e. step/stage 25 is optional), and the side information data items and data items regarding any re-mix constraints (described below) are transmitted.
- the output signals of step/stage 24 together with the output signals of step/stage 25 are used to form an enhanced MPEG SAOC bitstream.
- the object parameters are quantised and coded efficiently, and are correspondingly decoded and inversely quantised at receiver side.
- the downmix signal Before transmission, the downmix signal can be compressed, and is correspondingly decompressed at receiver side.
- the SAOC side information is (or can be) encoded according to the MPEG SAOC standard and is transmitted together with DMG and DCLG data e.g. as ancillary data portion of the downmix bitstream.
- a hybrid Quadrature Mirror Filter analysis filter bank step or stage 31 corresponding to filter bank 21 receives and processes the transmitted hQMF synthesised data from the enhanced MPEG SAOC bitstream, and feds them as time/frequency domain data to an enhanced MPEG SAOC decoder or transcoder step or stage 32 that is controlled by an estimation matrix T .
- a rendering matrix step or stage 35 receives user data regarding perceptual cues, e.g. a desired playback configuration and desired object positions, as well as the transmitted re-mix constraint data items, and there from a corresponding rendering matrix A is determined.
- Matrix A has the size of down-mixing matrix D and its coefficients are based on the coefficients of matrix D.
- the last row of rendering matrix A contains zero values only and all other matrix coefficients are identical to the coefficients in matrix D. I.e., each sound object is represented by a different row in matrix A.
- Matrix A together with an estimated covariance matrix C and a reconstructed down-mixing matrix D, are used for determining (as explained below) in an estimation matrix generator step or stage 36 the estimation matrix T.
- the estimated covariance matrix C and the reconstructed down-mixing matrix D are determined from the received side information data in a covariance and down-mixing matrix calculation step or stage 34.
- the estimation matrix T is used for decoding or transcoding the audio signals of the new audio scene.
- the downmixed signals of step/stage 32 are output as channel signals (e.g. Ch.#1 to Ch.#5) via a hybrid Quadrature Mirror Filter synthesis filter bank step or stage 33 (corresponding to synthesis filter bank 24).
- the perceptual cues are used as ancillary data in the main bitstream.
- Minimum bit rate means: such that the resulting audio quality is not affected, i.e. the distortions caused due to slightly less available bit rate for the audio signals are not audible, or at least not annoying.
- the Object Level Differences data OLD and Inter Object Coherence data IOC are used, the values of which are computed in step/stage 23 e.g. according to section Annex D.2 "Calculation of SAOC parameters" of the MPEG SAOC standard, for every frame/frequency processing band tile ( l,m ), i.e. for every non-overlapping 16 temporal slots and every K processing bands.
- nrg l , m i ⁇ j ⁇ t ⁇ l ⁇ k ⁇ m X t , k i * X t , k * j ⁇ t ⁇ l ⁇ k ⁇ m 1 + ⁇ , where the indices i and j stand for the ambience channel number and the audio object number, respectively, m is a current frequency processing band, k is a running frequency sample index within frequency processing band m , l is a current frame, t is a running temporal slot index within frame l , and ⁇ has a small value (e.g. 10- 9 ) and avoids a division by zero in the following computations.
- the DMG and DCLD data or parameters are extended versions of the corresponding MPEG SAOC parameters because they are not limited to two channels like in MPEG SAOC.
- the time/frequency resolution of DMG and DCLD can be modified with the moving speed of the audio objects in the audio scene. This time/frequency resolution change will not affect the performance of the inventive processing, and therefore it is assumed for simplicity that the time/frequency resolution at which these parameters are computed is equal to the time/frequency resolution at which the processing is done.
- the perceptual cues are used in step/stage 34 for approximating the covariance matrix C of the original input channels.
- the original rendering or mixing matrix is required at decoder side.
- the downmix Gains DMG values and the Downmix Channel Level Differences DCLD values from the additional side information are used.
- Matrix D is computed differently than according to MPEG SAOC, but its resulting content can assumed to be identical.
- every frame l and processing band m (cf. the above table), mixing matrix D l,m ( i,j ) and the rendering matrix A l,m ( i,j ) used for remixing the L objects have a size 5 by N in this embodiment.
- T l,m argmin ⁇ E l ⁇ Z t , m - Z ⁇ t , m ⁇ Z t , m - Z ⁇ t , m H .
- Z t,m is not available at decoder side but is used for the derivation of the following equations, and it finally turns out that knowledge of Z t,m at decoder side is not required.
- the rank of this matrix G l,m is computed in step/stage 25 before final encoding, and one can proceed with the encoding of the side information if that matrix has a full rank.
- the rank of a matrix is the number of independent columns or rows, and this can be hard to determine.
- the effective rank is used, which is a more stable measure for the rank and is described in section 6.3 "Singular Value Decomposition" of the textbook of Gilbert Strang, "Linear Algebra and its applications", 4th edition, published 19 July 2005 .
- the eigenvalues of matrix G l,m are computed.
- the rank value can be used for controlling the number K of frequency bands applied in the inventive processing, and thereby the accuracy of the side information parameters.
- the rank value can also be used for switching on or off a residual coding like in the MPEG SAOC standard.
- G l , m Q l , m ⁇ U l , m ⁇ Q l , m - 1
- Q l,m is a unitary matrix
- Q l , m - 1 Q l , m H
- a singular matrix is not very common and it can become non-singular with a small weighting of one coefficient of the singular matrix.
- matrix G l,m is invertible
- these eigenvalues are modified by adding a weight of ⁇ to each one of them.
- the error introduced by this procedure is of order ⁇ and will not affect the remixing processing in step/stage 32.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Content providers can add metadata to audio content such that consumers can control down-mix or dynamic range of selected parts of the audio signal. MPEG Spatial Audio Object Coding (SAOC) deals with parametric coding techniques for complex audio scenes at bit rates normally used for mono or stereo sound coding, offering at decoder side an interactive rendering of the audio objects mixed into the audio scene, whereby only a small amount of extra information is added to the audio bit stream. In parametric coding the biggest issue is that a perfect separation between objects is usually not possible. This issue is treated in the MPEG SAOC standard by using residual coding techniques and ensuring better separation only for a small set of objects before encoding. The invention describes how, by adding only a small amount of extra information to SAOC parameters, at receiver side a remixing of a broadcast audio signal is achieved, by using information about the actual mix of the audio signal, audio signal characteristics like correlation, and the desired audio scene rendering.
Description
- The invention relates to a method and to an apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side, wherein the decoder side downmixing is controlled by desired playback configuration data and/or desired object positioning data.
- Content providers are facing consumers with increasingly heterogeneous listening situations, e.g. home theatres, mobile audio, car and in-flight entertainment. Audio content cannot be processed by its creator or broadcaster so as to match every possible consumer listening condition, for example audio/video content played back on a mobile phone. Besides different listening conditions, also different listening experiences can be desirable, for instance in a live soccer broadcast a consumer can control his own virtual position within the sound scene of the stadium (pitch or stands), or can control the virtual position and the predominance of the commentator.
- Content providers can add guiding metadata to the audio content, such that consumers can control down-mix or dynamic range of selected parts of the audio signal and/or assure high speech intelligibility. For incorporation of such metadata into existing broadcasting chains, it is important that the general audio format is not changed (legacy playback) and that only a small amount of extra information (e.g. as ancillary data) is added to the audio bit stream. MPEG Spatial Audio Object Coding (SAOC), ISO/IEC 23003-1:2007, MPEG audio technologies - Part 1: MPEG Surround, and ISO/IEC 23003-2:2010, MPEG audio technologies - Part 2: Spatial Audio Object Coding, deals with parametric coding techniques for complex audio scenes at bit rates normally used for mono or stereo sound coding, offering at decoder side an interactive rendering of the audio objects mixed into the audio scene.
- MPEG SAOC was developed starting from Spatial Audio Coding which is based on a 'channel-oriented' approach, by introducing the concept of audio objects, having the purpose to offer even more flexibility at receiver side. Since it is a parametric multiple object coding technique, the additional cost in terms of bit rate is limited to 2-3 kbit/s for each audio object. Although the bit rate increases with the number of audio objects, it still remains at small values in comparison with the actual audio data transmitted as a mono/stereo downmix.
- In parametric coding the biggest issue is that a perfect separation between objects is usually not possible, especially in an extreme mixing scenario, for example when all sound objects are mixed in all channels, or in case of 'applause-like' sound. This issue is treated in the MPEG SAOC standard by using residual coding techniques and ensuring better separation only for a small set of objects before encoding.
- A standard MPEG SAOC architecture is illustrated in
Fig. 1 . A variable number N of sound objects Obj.#1, Obj.#2, Obj.#3, Obj.#4 is input to anSAOC encoder 11 that provides an encoded background compatible mono or stereo downmix signal or signals with SAOC parameters and side information. The downmix data signal or signals and the side information data are sent as dedicated bit streams to anSAOC decoder 12. The receiver side processing is carried out basically in two steps: first, using the side information,SAOC decoder 12 reconstructs the original soundobjects Obj.# 1,Obj.# 2,Obj.# 3,Obj.# 4. Controlled by interaction/control information, these reconstructed sound objects are re-mixed and rendered in anMPEG surround renderer 13. For example, the reconstructed sound objects are output as a stereo signal withchannels # 1 and #2. - In practice, in order to avoid changing twice the temporal domain and the frequency domain, the steps are merged, thereby substantially reducing the computational complexity. The processing in
SAOC encoder 11,SAOC decoder 12 andrenderer 13 is performed in the frequency domain, using a nonuniform scale in order to model most efficiently the human auditory system. To ensure compatibility with MPEG Surround, a hybrid Quadrature Mirror Filter hQMF 21 shown inFig. 2 was chosen, which filter offers a better resolution for low frequencies. After applying the hQMF filter to all input signals, a time/frequency grid is obtained that maps every time slot or frame of the input signals to a number of processing bands obtained by merging the frequency bands. - The following SAOC parameters are computed for every time/frequency tile and are transmitted to the decoder side:
- Object Level Differences data OLD describe the amount of energy contained by each sound object with respect to the sound object having the highest energy. The energy level value of the loudest object is described by the Object Energy parameter data NRG, which can be transmitted to the decoder;
- Inter-Object Coherence data IOC are used to describe the amount of similarity between the sound objects and are computed for every pair of two input audio objects;
- Downmix Gains data DMG and Downmix Channel Level Differences data DCLD are used for describing the gains applied to each sound object in the mixing process, and at decoder side they are used for reconstructing the downmixing matrix.
- At decoder side, in order to avoid generation of bad quality audio content due to extreme rendering, a Distortion Control Unit DCU can be used. The final rendering matrix coefficients are computed as a linear combination of user-specified coefficients and the target coefficients which are assumed to be distortion-free.
- The main drawback of the solution offered by MPEG SAOC is the limitation to a maximum of two down-mix channels. Further, the MPEG SAOC standard is not designed for 'Solo/Karaoke' type of applications, which involve the complete suppression of one or more audio objects. In the MPEG SAOC standard this problem is tackled by using residual coding for specific audio objects, thereby increasing the bit rate.
- A problem to be solved by the invention is to overcome these limitations of MPEG SAOC and to allow for adding of side information for a legacy multi-channel audio broadcasting like 5.1. This problem is solved by the methods disclosed in
claims claims - The invention describes how, by adding only a small amount of extra bit rate, at decoder or receiver side a re-mixing of a broadcast audio signal is achieved using information about the actual mix of the audio signal, audio signal characteristics like correlation, level differences, and the desired audio scene rendering.
- A second embodiment shows how to determine already at encoder side the suitability of the actual multi-channel audio signal for a remix at decoder side. This feature allows for countermeasures (e.g. changing the mixing matrix used, i.e. how the sound objects are mixed into the different channels) if a decoder side re-mixing without perceivable artefacts, or without problems like the necessary additional transmission of the audio objects itself for a short time, is not possible.
- Advantageously, due to processing the same type of side information parameters, the invention could be used to amend the MPEG SAOC standard correspondingly, based on the same building blocks.
- In principle, the inventive encoding method is suited for downmixing spatial audio signals that can be downmixed at receiver side in a manner different from the manner of downmixing at encoder side, wherein said encoding is based on MPEG SAOC and said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said method including the steps:
- processing M correlated sound signals, M being greater than '2', and L independent sound signals, L being '1' or greater, in an analysis filter bank providing corresponding time/frequency domain signals;
- multiplying said time/frequency domain signals with a downmix matrix Dl,m, followed by processing the resulting signals in a synthesis filter bank that has an inverse operation of said analysis filter bank and that provides M time domain output signals;
- determining from said time/frequency domain signals MPEG SAOC side information data including Object Level Differences data OLD and Inter-Object Coherence data IOC, as well as enhanced Downmix Gains data DMG and Downmix Channel Level Differences data DCLD, wherein said DMG and DCLD data are related to M channels.
- In principle, the inventive decoding method is suited downmixing spatial audio signals processed according to the encoding method in a manner different from the manner of downmixing at encoder side, wherein said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said method including the steps:
- receiving said processed spatial audio signals and processing them in an analysis filter bank, providing corresponding time/frequency domain signals;
- determining from said desired playback configuration data and/or said desired object positioning data a rendering matrix Al,m ;
- determining from the received OLD, IOC, DMG and DCLD data an estimated covariance matrix Cl,m and a reconstructed down-mixing matrix Dl,m ;
- calculating an estimation matrix
- multiplying said time/frequency domain signals (Y) with said estimation matrix Tl,m so as to get desired-remix signals, followed by processing said desired-remix signals in a synthesis filter bank that has an inverse operation of said analysis filter bank.
- In principle, the inventive encoding apparatus is suited for downmixing spatial audio signals that can be downmixed at receiver side in a manner different from the manner of downmixing at encoder side, wherein said encoding is based on MPEG SAOC and said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said apparatus including:
- an analysis filter bank for processing M correlated sound signals, M being greater than '2', and L independent sound signals, L being '1' or greater, providing corresponding time/frequency domain signals;
- means being adapted for multiplying said time/frequency domain signals with a downmix matrix Dl,m ;
- a synthesis filter bank for said multiplied time/frequency domain signals that has an inverse operation of said analysis filter bank and that provides M time domain output signals;
- means being adapted for determining from said time/frequency domain signals MPEG SAOC side information data including Object Level Differences data OLD and Inter-Object Coherence data IOC, as well as enhanced Downmix Gains data DMG and Downmix Channel Level Differences data DCLD, wherein said DMG and DCLD data are related to M channels.
- In principle, the inventive decoding apparatus is suited for downmixing spatial audio signals processed according to the encoding method in a manner different from the manner of downmixing at encoder side, wherein said downmixing at receiver side can be controlled by desired playback configuration data and/or desired object positioning data, said apparatus including:
- means being adapted for receiving said processed spatial audio signals and for processing them in an analysis filter bank, providing corresponding time/frequency domain signals;
- means being adapted for determining from said desired playback configuration data and/or said desired object positioning data a rendering matrix Al,m ;
- means being adapted for determining from the received OLD, IOC, DMG and DCLD data an estimated covariance matrix Cl,m and a reconstructed down-mixing matrix Dl,m;
- means (36) being adapted for calculating an estimation matrix
- means being adapted for multiplying said time/frequency domain signals with said estimation matrix Tl,m so as to get desired-remix signals, followed by processing said desired-remix signals in a synthesis filter bank that has an inverse operation of said analysis filter bank.
- Advantageous additional embodiments of the invention are disclosed in the respective dependent claims.
- Exemplary embodiments of the invention are described with reference to the accompanying drawings, which show in:
- Fig. 1
- standard MPEG SAOC system;
- Fig. 2
- enhanced MPEG SAOC encoder;
- Fig. 3
- enhanced MPEG SAOC decoder.
- The inventive spatial audio object coding system with five down-mix channels facilitates a backward compatible transmission at bit rates only slighter higher (due to the extended content of the side information: OLD, IOC, DMG and DCLD) than the bit rates for known 5.1 channel transmission. By using the side information, a guided spatial remix is achieved, as well as the adaption of the audio mix to different listening situations, including guided downmix (or even upmix) of multi-channel audio.
- In the following embodiment, a number of M=5 channels containing the ambience signals and a number of L audio objects mixed over the ambiance are considered. An example is the stadium ambiance of a soccer match plus specific sound effects (ball kicks, whistle) and one or more commentators. Thus, the encoder input gets N = M+L audio channels.
- In other embodiments, M is at least '2' and L is '1' or greater.
- At decoder side it is not intended to reconstruct the audio objects, but to offer the possibility of re-mixing, attenuating, totally suppressing, and changing the position of the audio objects in the rendered audio scene.
- For the processing part of the system any time/frequency transform can be used. In this embodiment, hybrid Quadrature Mirror Filter (hQMF) banks are used for better selectivity in the frequency domain. The spatial audio input signals are processed in non-overlapping multiple-sample temporal slots, in particular 64-samples temporal slots. These temporal slots are used for computing the perceptual cues or characteristics for every successive frame, which has a length of a fixed number of temporal slots, in particular 16 temporal slots.
- In the frequency domain, 71 frequency bands are used according to the sensitivity of the human auditory system, and are grouped into K processing bands, K having a value of '2', '3' or '4', thereby obtaining different levels of accuracy. The hQMF filter bank transforms in each case 64 time samples into 71 frequency samples. The processing band borders are represented in the following table:
Table 2.1: Processing Bands according to different quality levels K=4 K=3 K=2 Processing Band 10-7 0-7 0-20 Processing Band 28-20 8-55 21-70 Processing Band 321-29 56-70 - Processing Band 430-70 - - - A basic block diagram of the inventive encoder and decoder is shown in
Fig.2 andFig.3 , respectively. - In the encoder in
Fig.2 , a hybrid Quadrature Mirror Filter analysis filter bank step orstage 21 is applied to all audio input signals, e.g. ambience channels Ch.#1 toCh.# 5 and soundobjects Obj.# 1 to Obj.#L (at least one sound object). The number of ambience channels is not limited to five. In this invention, contrary to MPEG SAOC, the ambience channels are not independent sound objects but are usually correlated sound signals. The corresponding filter bank outputs time/frequency domain signals X, which are fed on one hand to a down-mixer step orstage 22 that multiplies them with a downmix matrix D and provides, via a hybrid Quadrature Mirror Filter synthesis filter bank step orstage 24 that has an inverse operation of the analysis filter bank, the audio channels in time domain for transmission, and on the other hand are fed to an enhanced MPEG SAOC parameter calculator step orstage 23. The synthesis filter bank inverses the function of the analysis filter bank. The enhanced SAOC parameters determine the rendering flexibility in a decoder and include, as mentioned above, Object Level Differences data OLD, Inter-Object Coherence data IOC, Downmix Gains data DMG, Downmix Channel Level Differences data DCLD, and can comprise Object Energy parameter data NRG. Except the DMG and DCLD data, the other parameters correspond to original MPEG SAOC parameters. From these side information data items a rank can be determined as described below in a rank calculator step or stage 25 (i.e. step/stage 25 is optional), and the side information data items and data items regarding any re-mix constraints (described below) are transmitted. - The output signals of step/
stage 24 together with the output signals of step/stage 25 are used to form an enhanced MPEG SAOC bitstream. - As mentioned in section 5.1 SAOC overview/Introduction of the MPEG SAOC standard, the object parameters are quantised and coded efficiently, and are correspondingly decoded and inversely quantised at receiver side. Before transmission, the downmix signal can be compressed, and is correspondingly decompressed at receiver side. The SAOC side information is (or can be) encoded according to the MPEG SAOC standard and is transmitted together with DMG and DCLG data e.g. as ancillary data portion of the downmix bitstream.
- In the receiver-side decoder in
Fig.3 , a hybrid Quadrature Mirror Filter analysis filter bank step orstage 31 corresponding to filterbank 21 receives and processes the transmitted hQMF synthesised data from the enhanced MPEG SAOC bitstream, and feds them as time/frequency domain data to an enhanced MPEG SAOC decoder or transcoder step orstage 32 that is controlled by an estimation matrix T. A rendering matrix step orstage 35 receives user data regarding perceptual cues, e.g. a desired playback configuration and desired object positions, as well as the transmitted re-mix constraint data items, and there from a corresponding rendering matrix A is determined. Matrix A has the size of down-mixing matrix D and its coefficients are based on the coefficients of matrix D. In case e.g. only the 'last' sound object Obj.#L shall not be present in the decoder output signals, the last row of rendering matrix A contains zero values only and all other matrix coefficients are identical to the coefficients in matrix D. I.e., each sound object is represented by a different row in matrix A. - Matrix A, together with an estimated covariance matrix C and a reconstructed down-mixing matrix D, are used for determining (as explained below) in an estimation matrix generator step or
stage 36 the estimation matrix T. - The estimated covariance matrix C and the reconstructed down-mixing matrix D are determined from the received side information data in a covariance and down-mixing matrix calculation step or
stage 34. The estimation matrix T is used for decoding or transcoding the audio signals of the new audio scene. The downmixed signals of step/stage 32 are output as channel signals (e.g. Ch.#1 to Ch.#5) via a hybrid Quadrature Mirror Filter synthesis filter bank step or stage 33 (corresponding to synthesis filter bank 24). - In order to obtain a minimum bit rate for the side information encoding, the perceptual cues are used as ancillary data in the main bitstream. "Minimum bit rate" means: such that the resulting audio quality is not affected, i.e. the distortions caused due to slightly less available bit rate for the audio signals are not audible, or at least not annoying. For characterising the input audio objects the Object Level Differences data OLD and Inter Object Coherence data IOC are used, the values of which are computed in step/
stage 23 e.g. according to section Annex D.2 "Calculation of SAOC parameters" of the MPEG SAOC standard, for every frame/frequency processing band tile (l,m), i.e. for every non-overlapping 16 temporal slots and every K processing bands. - First the auto-correlation and the cross-correlation of any two objects are computed and saved in a matrix format:
where the indices i and j stand for the ambience channel number and the audio object number, respectively, m is a current frequency processing band, k is a running frequency sample index within frequency processing band m, l is a current frame, t is a running temporal slot index within frame l, and ε has a small value (e.g. 10-9) and avoids a division by zero in the following computations. -
- In order to characterise at decoder side the down-mixing matrix coefficients, in step/
stage 23 the values for Downmix Gains data DMG and Downmix Channel Level Differences data DCLD are calculated, again for every frame l with 16 temporal slots and frequency processing band m:
where r=1:N represents the input signal index and j=1:5 (in other embodiments j=1:M) represents the down-mix channel index. The DMG and DCLD data or parameters are extended versions of the corresponding MPEG SAOC parameters because they are not limited to two channels like in MPEG SAOC. Depending on the mixing procedure, the time/frequency resolution of DMG and DCLD can be modified with the moving speed of the audio objects in the audio scene. This time/frequency resolution change will not affect the performance of the inventive processing, and therefore it is assumed for simplicity that the time/frequency resolution at which these parameters are computed is equal to the time/frequency resolution at which the processing is done. - At decoder side the perceptual cues are used in step/
stage 34 for approximating the covariance matrix C of the original input channels. When considering the definition of the Object Level Differences OLD and Inter Object Coherence IOC, the following form for the estimated covariance matrix C is obtained, computed for every time/frequency (l,m) tile: - In order to remix the audio objects, besides the new rendering matrix also the original rendering or mixing matrix is required at decoder side. For re-constructing the downmixing matrix D in step/
stage 34, the Downmix Gains DMG values and the Downmix Channel Level Differences DCLD values from the additional side information are used. The content of the original down-mix matrix D is computed in step/stage 34 as:
where r=1:N represents the input signal index and i=2:5 (in other embodiments i=2:M) represents the down-mix channel index, j represents here a running input signal index within the sum, and k is a running frequency sample index within frequency processing band m. Matrix D is computed differently than according to MPEG SAOC, but its resulting content can assumed to be identical. - In the transcoding at decoder side, every frame l and processing band m (cf. the above table), mixing matrix Dl,m (i,j) and the rendering matrix Al,m (i,j) used for remixing the L objects have a
size 5 by N in this embodiment. For every time/frequency tile the original down-mixed signals Y at encoder side are, and the desired remixed signals Z at decoder side would be: -
- I.e., the calculation of Y is carried out for each temporal slot t in a current frame l but for all these temporal slots the same downmix matrix D is used.
-
-
-
-
-
-
- Because the expression of G l,m is depending only on parameters known at encoder side and does not depend on the rendering matrix, it can be decided before encoding whether or not remixing is feasible at decoder side. Practice, and the special form of the down-mixing and correlation matrices of the ambiance channels, show that in most cases matrix Gl,m will be invertible.
- In order to ensure the functionality of the decoding processing, the rank of this matrix Gl,m is computed in step/
stage 25 before final encoding, and one can proceed with the encoding of the side information if that matrix has a full rank. The rank of a matrix is the number of independent columns or rows, and this can be hard to determine. Thus, instead of using the rank, the effective rank is used, which is a more stable measure for the rank and is described in section 6.3 "Singular Value Decomposition" of the textbook of Gilbert Strang, "Linear Algebra and its applications", 4th edition, published 19 July 2005. First, the eigenvalues of matrix Gl,m are computed. Next, the number of eigenvalues greater than a tolerance value τ is counted and this number is taken as the effective rank. Practice shows that τ = 10-5 (a preferential range for τ is 10-4...10-8) can be chosen, and for matrices with full effective rank it can be assumed that it is accurate enough for the inverse computation. - The rank value can be used for controlling the number K of frequency bands applied in the inventive processing, and thereby the accuracy of the side information parameters. The rank value can also be used for switching on or off a residual coding like in the MPEG SAOC standard.
- Measuring the sensitivity entirely by the smallest singular value may not be sufficient, because a simple multiplication of the matrix with factor 105 will indicate a much less singular matrix, and an ill-conditioning problem is not solved by a simple re-scaling. Thus, before computing the effective rank as described above, a normalisation of matrix Gl,m is carried out.
- Because Cl,m represents a correlation matrix of the input audio signals, it will be a symmetric matrix and obviously Gl,m =
where Ql,m is a unitary matrix - An interesting property of the Schur decomposition is that the inverse of Gl,m can be easily computed as:
whereU l,m is a diagonal matrix having on the main diagonal the inverse of the eigenvalues of matrix Gl,m diag(U l,m)= {1/λ1,1/λ2,1/λ3,1/λ4,1/λ5}. Proof: matrix Ul,m is a diagonal matrix and, if the values from the main diagonal are different from zero, the matrix is invertible, whereby the inverse is equal toU l,m. Thus the processing starts with the following computation: -
- A singular matrix is not very common and it can become non-singular with a small weighting of one coefficient of the singular matrix. Thus, in order to ensure at decoder side that matrix Gl,m is invertible, after computing the Schur decomposition and finding eigenvalues smaller than the defined tolerance value τ, these eigenvalues are modified by adding a weight of τ to each one of them. In this way, when computing the inverse of Gl,m, using the described property of the Schur decomposition, it is sure that this matrix is well-conditioned. The error introduced by this procedure is of order τ and will not affect the remixing processing in step/
stage 32. - The values used for Cl,m and Dl,m are estimated according to equations (4) and (5).
Claims (12)
- Method for encoding by downmixing (22) spatial audio signals (Ch.#1 - Ch.#5, Obj.#1 - Obj.#L) that can be downmixed at receiver side in a manner different from the manner of downmixing at encoder side, wherein said encoding is based on MPEG SAOC and said downmixing at receiver side can be controlled (35) by desired playback configuration data and/or desired object positioning data, said method including the steps:- processing M correlated sound signals (Ch.#1 - Ch.#5), M being greater than '2', and L independent sound signals (Obj.#1 - Obj.#L), L being '1' or greater, in an analysis filter bank (21) providing corresponding time/frequency domain signals (X);- multiplying (22) said time/frequency domain signals with a downmix matrix Dl,m, followed by processing the resulting signals (Y) in a synthesis filter bank (24) that has an inverse operation of said analysis filter bank (21) and that provides M time domain output signals;- determining (23) from said time/frequency domain signals (X) MPEG SAOC side information data including Object Level Differences data OLD and Inter-Object Coherence data IOC, as well as enhanced Downmix Gains data DMG and Downmix Channel Level Differences data DCLD, wherein said DMG and DCLD data are related to M channels.
- Method for downmixing (32) spatial audio signals processed according to claim 1 in a manner different from the manner of downmixing at encoder side, wherein said downmixing at receiver side can be controlled (35) by desired playback configuration data and/or desired object positioning data, said method including the steps:- receiving said processed spatial audio signals and processing them in an analysis filter bank (31), providing corresponding time/frequency domain signals (Y);- determining (35) from said desired playback configuration data and/or said desired object positioning data a rendering matrix Al,m ;- determining (34) from the received OLD, IOC, DMG and DCLD data an estimated covariance matrix Cl,m and a reconstructed down-mixing matrix Dl,m ;- multiplying (32) said time/frequency domain signals (Y) with said estimation matrix Tl,m so as to get desired-remix signals (Ẑ), followed by processing said desired-remix signals in a synthesis filter bank (33) that has an inverse operation of said analysis filter bank (31).
- Apparatus for encoding by downmixing (22) spatial audio signals (Ch.#1 - Ch.#5, Obj.#1 - Obj.#L) that can be downmixed at receiver side in a manner different from the manner of downmixing at encoder side, wherein said encoding is based on MPEG SAOC and said downmixing at receiver side can be controlled (35) by desired playback configuration data and/or desired object positioning data, said apparatus including:- an analysis filter bank (21) for processing M correlated sound signals (Ch.#1 - Ch.#5), M being greater than '2', and L independent sound signals (Obj.#1 - Obj.#L), L being '1' or greater, providing corresponding time/frequency domain signals (X);- means (22) being adapted for multiplying said time/frequency domain signals (X) with a downmix matrix Dl,m ;- a synthesis filter bank (24) for said multiplied time/frequency domain signals (Y) that has an inverse operation of said analysis filter bank (21) and that provides M time domain output signals;- means (23) being adapted for determining from said time/frequency domain signals (X) MPEG SAOC side information data including Object Level Differences data OLD and Inter-Object Coherence data IOC, as well as enhanced Downmix Gains data DMG and Downmix Channel Level Differences data DCLD, wherein said DMG and DCLD data are related to M channels.
- Apparatus for downmixing (32) spatial audio signals processed according to claim 1 in a manner different from the manner of downmixing at encoder side, wherein said downmixing at receiver side can be controlled (35) by desired playback configuration data and/or desired object positioning data, said apparatus including:- means (31) being adapted for receiving said processed spatial audio signals and for processing them in an analysis filter bank, providing corresponding time/frequency domain signals (Y);- means (35) being adapted for determining from said desired playback configuration data and/or said desired object positioning data a rendering matrix Al,m ;- means (34) being adapted for determining from the received OLD, IOC, DMG and DCLD data an estimated covariance matrix Cl,m and a reconstructed down-mixing matrix Dl,m ;- means (32) being adapted for multiplying said time/frequency domain signals (Y) with said estimation matrix Tl,m so as to get desired-remix signals (Ẑ), followed by processing said desired-remix signals in a synthesis filter bank (33) that has an inverse operation of said analysis filter bank (31).
- Method according to the method of claim 1 or 2, or apparatus according to the apparatus of claim 3 or 4, wherein said spatial input signals are processed in non-overlapping multiple-sample temporal slots, a fixed number of such temporal slots representing a frame l, and in K frequency processing bands into which the total frequency range is divided, K having a value of '2', '3' or '4'.
- Method according to the method of claim 5, or apparatus according to the apparatus of claim 5, wherein said Downmix Gains data DMG and Downmix Channel Level Differences data DCLD are calculated for every input signal frame l and processing band m according to:
where r=1:(M+L) represents an spatial audio input signal index, j=1:5 represents a down-mix channel index and value ε is used for avoiding a division by zero in other related computations. - Method according to the method of claim 5 or 6, or apparatus according to the apparatus of claim 5 or 6, wherein said rendering matrix Al,m has a size of said downmix matrix Dl,m and its coefficients are based on the coefficients of matrix Dl,m, wherein each sound object is represented by a different row in matrix A l,m .
- Method according to the method of one of claims 5 to 7, or apparatus according to the apparatus of one of claims 5 to 7, wherein said estimated covariance matrix C l,m is calculated according to
where r=1:(M+L) represents a spatial audio input signal index, i=2:5 represents a down-mix channel index, and k is a running frequency sample index within a current frequency processing band. - Method according to the method of one of claims 5 to 8, or apparatus according to the apparatus of one of claims 5 to 8, wherein from said OLD, IOC, DMG and DCLD data a matrix
- Method according to the method of claim 9, or apparatus according to the apparatus of claim 9, wherein said rank value is used to control the number K of frequency bands used in the processing.
- Method according to the method of claim 9 or 10, or apparatus according to the apparatus of claim 9 or 10, wherein said rank value is used to switch on or off a residual coding.
- Digital audio signal that is encoded according to the method of one of claims 1, 5 and 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP12305914.9A EP2690621A1 (en) | 2012-07-26 | 2012-07-26 | Method and Apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP12305914.9A EP2690621A1 (en) | 2012-07-26 | 2012-07-26 | Method and Apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side |
Publications (1)
Publication Number | Publication Date |
---|---|
EP2690621A1 true EP2690621A1 (en) | 2014-01-29 |
Family
ID=47002786
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP12305914.9A Withdrawn EP2690621A1 (en) | 2012-07-26 | 2012-07-26 | Method and Apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side |
Country Status (1)
Country | Link |
---|---|
EP (1) | EP2690621A1 (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015056383A1 (en) * | 2013-10-17 | 2015-04-23 | パナソニック株式会社 | Audio encoding device and audio decoding device |
CN106796804A (en) * | 2014-10-02 | 2017-05-31 | 杜比国际公司 | For talking with enhanced coding/decoding method and decoder |
US9805727B2 (en) | 2013-04-03 | 2017-10-31 | Dolby Laboratories Licensing Corporation | Methods and systems for generating and interactively rendering object based audio |
CN108632048A (en) * | 2017-03-22 | 2018-10-09 | 展讯通信(上海)有限公司 | Conference telephone control method, device and mostly logical terminal |
US10136240B2 (en) | 2015-04-20 | 2018-11-20 | Dolby Laboratories Licensing Corporation | Processing audio data to compensate for partial hearing loss or an adverse hearing environment |
CN109036441A (en) * | 2014-03-24 | 2018-12-18 | 杜比国际公司 | To the method and apparatus of high-order clear stereo signal application dynamic range compression |
WO2019227991A1 (en) * | 2018-05-31 | 2019-12-05 | 华为技术有限公司 | Method and apparatus for encoding stereophonic signal |
CN110739000A (en) * | 2019-10-14 | 2020-01-31 | 武汉大学 | Audio object coding method suitable for personalized interactive system |
CN113678199A (en) * | 2019-03-28 | 2021-11-19 | 诺基亚技术有限公司 | Determination of the importance of spatial audio parameters and associated coding |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080049943A1 (en) * | 2006-05-04 | 2008-02-28 | Lg Electronics, Inc. | Enhancing Audio with Remix Capability |
US20110166867A1 (en) * | 2008-07-16 | 2011-07-07 | Electronics And Telecommunications Research Institute | Multi-object audio encoding and decoding apparatus supporting post down-mix signal |
US20120078642A1 (en) * | 2009-06-10 | 2012-03-29 | Jeong Il Seo | Encoding method and encoding device, decoding method and decoding device and transcoding method and transcoder for multi-object audio signals |
US20120177204A1 (en) * | 2009-06-24 | 2012-07-12 | Oliver Hellmuth | Audio Signal Decoder, Method for Decoding an Audio Signal and Computer Program Using Cascaded Audio Object Processing Stages |
-
2012
- 2012-07-26 EP EP12305914.9A patent/EP2690621A1/en not_active Withdrawn
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080049943A1 (en) * | 2006-05-04 | 2008-02-28 | Lg Electronics, Inc. | Enhancing Audio with Remix Capability |
US20110166867A1 (en) * | 2008-07-16 | 2011-07-07 | Electronics And Telecommunications Research Institute | Multi-object audio encoding and decoding apparatus supporting post down-mix signal |
US20120078642A1 (en) * | 2009-06-10 | 2012-03-29 | Jeong Il Seo | Encoding method and encoding device, decoding method and decoding device and transcoding method and transcoder for multi-object audio signals |
US20120177204A1 (en) * | 2009-06-24 | 2012-07-12 | Oliver Hellmuth | Audio Signal Decoder, Method for Decoding an Audio Signal and Computer Program Using Cascaded Audio Object Processing Stages |
Non-Patent Citations (1)
Title |
---|
GILBERT STRANG: "Singular Value Decomposition", 19 July 2005, article "Linear Algebra and its applications" |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11769514B2 (en) | 2013-04-03 | 2023-09-26 | Dolby Laboratories Licensing Corporation | Methods and systems for rendering object based audio |
US10276172B2 (en) | 2013-04-03 | 2019-04-30 | Dolby Laboratories Licensing Corporation | Methods and systems for generating and interactively rendering object based audio |
US11270713B2 (en) | 2013-04-03 | 2022-03-08 | Dolby Laboratories Licensing Corporation | Methods and systems for rendering object based audio |
US9805727B2 (en) | 2013-04-03 | 2017-10-31 | Dolby Laboratories Licensing Corporation | Methods and systems for generating and interactively rendering object based audio |
US10832690B2 (en) | 2013-04-03 | 2020-11-10 | Dolby Laboratories Licensing Corporation | Methods and systems for rendering object based audio |
US10553225B2 (en) | 2013-04-03 | 2020-02-04 | Dolby Laboratories Licensing Corporation | Methods and systems for rendering object based audio |
US10002616B2 (en) | 2013-10-17 | 2018-06-19 | Socionext Inc. | Audio decoding device |
WO2015056383A1 (en) * | 2013-10-17 | 2015-04-23 | パナソニック株式会社 | Audio encoding device and audio decoding device |
US9779740B2 (en) | 2013-10-17 | 2017-10-03 | Socionext Inc. | Audio encoding device and audio decoding device |
CN109036441B (en) * | 2014-03-24 | 2023-06-06 | 杜比国际公司 | Method and apparatus for applying dynamic range compression to high order ambisonics signals |
US11838738B2 (en) | 2014-03-24 | 2023-12-05 | Dolby Laboratories Licensing Corporation | Method and device for applying Dynamic Range Compression to a Higher Order Ambisonics signal |
CN109036441A (en) * | 2014-03-24 | 2018-12-18 | 杜比国际公司 | To the method and apparatus of high-order clear stereo signal application dynamic range compression |
CN106796804A (en) * | 2014-10-02 | 2017-05-31 | 杜比国际公司 | For talking with enhanced coding/decoding method and decoder |
CN106796804B (en) * | 2014-10-02 | 2020-09-18 | 杜比国际公司 | Decoding method and decoder for dialog enhancement |
US10136240B2 (en) | 2015-04-20 | 2018-11-20 | Dolby Laboratories Licensing Corporation | Processing audio data to compensate for partial hearing loss or an adverse hearing environment |
CN108632048A (en) * | 2017-03-22 | 2018-10-09 | 展讯通信(上海)有限公司 | Conference telephone control method, device and mostly logical terminal |
CN108632048B (en) * | 2017-03-22 | 2020-12-22 | 展讯通信(上海)有限公司 | Conference call control method and device and multi-pass terminal |
US11462224B2 (en) | 2018-05-31 | 2022-10-04 | Huawei Technologies Co., Ltd. | Stereo signal encoding method and apparatus using a residual signal encoding parameter |
US11978463B2 (en) | 2018-05-31 | 2024-05-07 | Huawei Technologies Co., Ltd. | Stereo signal encoding method and apparatus using a residual signal encoding parameter |
WO2019227991A1 (en) * | 2018-05-31 | 2019-12-05 | 华为技术有限公司 | Method and apparatus for encoding stereophonic signal |
CN113678199A (en) * | 2019-03-28 | 2021-11-19 | 诺基亚技术有限公司 | Determination of the importance of spatial audio parameters and associated coding |
CN110739000A (en) * | 2019-10-14 | 2020-01-31 | 武汉大学 | Audio object coding method suitable for personalized interactive system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP2690621A1 (en) | Method and Apparatus for downmixing MPEG SAOC-like encoded audio signals at receiver side in a manner different from the manner of downmixing at encoder side | |
US9257128B2 (en) | Apparatus and method for coding and decoding multi object audio signal with multi channel | |
US9578435B2 (en) | Apparatus and method for enhanced spatial audio object coding | |
JP5189979B2 (en) | Control of spatial audio coding parameters as a function of auditory events | |
JP5592974B2 (en) | Enhanced coding and parameter representation in multi-channel downmixed object coding | |
EP2941771B1 (en) | Decoder, encoder and method for informed loudness estimation employing by-pass audio object signals in object-based audio coding systems | |
US8867753B2 (en) | Apparatus, method and computer program for upmixing a downmix audio signal | |
US8515759B2 (en) | Apparatus and method for synthesizing an output signal | |
CN105518775B (en) | Artifact cancellation for multi-channel downmix comb filters using adaptive phase alignment | |
KR101798117B1 (en) | Encoder, decoder and methods for backward compatible multi-resolution spatial-audio-object-coding | |
US11501785B2 (en) | Method and apparatus for adaptive control of decorrelation filters | |
MX2012005781A (en) | Apparatus for providing an upmix signal represen. | |
EP2439736A1 (en) | Down-mixing device, encoder, and method therefor | |
WO2023172865A1 (en) | Methods, apparatus and systems for directional audio coding-spatial reconstruction audio processing | |
RU2485605C2 (en) | Improved method for coding and parametric presentation of coding multichannel object after downmixing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
AX | Request for extension of the european patent |
Extension state: BA ME |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN |
|
18D | Application deemed to be withdrawn |
Effective date: 20140730 |