1330992 九、發明說明： 【發明所屬之技術領域】 、本發明係關於一種編碼方法及其裳置及-種解碼方 ^及其襄置’更係關於—種可利用可補償降混信號之額外 貝訊對夕通道音訊信號加以加編碼或解碼的編碼方法 及其裝置與解碼方法及其裝置。 【先前技術】1330992 IX. Description of the invention: [Technical field to which the invention pertains] The present invention relates to an encoding method, an arranging method thereof, and a decoding method, and a device thereof, which are more related to the use of a compensable downmix signal. An encoding method, a device and a decoding method thereof and a device thereof for encoding or decoding an audio signal of a Bayer channel. [Prior Art]
在對一多通道音訊信號加以編碼的典型方法中，一 多通道音訊信號被降混為—單聲道或立體聲信號且該單 聲或立體聲信號被與空間資訊—塊編碼，而非對該多通道 音訊信號之每-通道加以編碼。在該方法中，該空間資訊 被用以回设為其原始多通道音訊信號。 第一圖為-對多通道音訊信號加以編瑪/解碼之血型 糸統的方塊圖。請參閱第—圖，—音訊信號編碼器包含一 降混模組及_ ”參數估計触，其巾卿混模組用以產 ^降混信號，藉由降混—多通道音訊信號成—身歷聲或 =聲道信號的方式為之，該空間估計模組則用以產生空間 資訊。該系統可接收—在外部受處理的人項混信號，而 非產生-降混信號。-音訊信號解碼輯該”參數估計 模組產生之空間資訊加以解譯，並根據該解譯結果回復該 原始多通道音訊信號ϋ在音訊信號編碼器產生一降 昆信號或-人卫降混信號產生的期間’信號位準衰減 形可能會在加上不同通道信號的過程中產生1加上分別 ^30992 一有位準L1及L2的二通道為例，若該二通道並不重疊， i破此間有-偏移存在，此時該二信號加成所得之一通道 位準DL12低於L1與L2的和。 在一解碼處理中，一降混信號之位準的衰減可能會造 、4遽的失真。舉例而言，通道之位準間的關係可根據通 道位準差（⑽)資訊加以料，其中通道位準差資訊是-種空間|訊，其指出通道位準⑽差異。不過，當通道相 加所得之降混信號的料衰減時，解碼所狀降混信號的 位準會低於肩始降混信號的位準。 由於刚述現象存在之故，一由解石馬所得之多通道音訊 信號在-預定頻率上可能會被提升或下抑，故會造成聲音 ^的破壞。此外，由於-信號因另一信號造成該信號本 身的部份偏差而產生之位準的衰減程度會隨頻域的不同 而不同’故-信號在通過一音訊編碼器及一音訊解碼器之 後的失真程度亦會隨頻率的不同而不同，且此問題無法藉 由在-預定頻域内改變-降混信號之能量的方式而獲完 全解決。 【發明内容】 [技術問題] 本發明提供一種利用可補僧_隊、θ 殉彳貝降混信號之額外資訊 對一多通道音訊信號加以編碼的編碼方法及其裝置。 本發明同時還提供-種利用可補償一降混信號之額 外資訊對-多通道音訊信號加以解碼的解碼方法及其裝 置。 1330992 [技術手段] 本發明之一態樣為一種解碼方法，該解碼方法包含下 列步驟：自一輸入信號取出一降混信號及額外資訊；自該額 、 外資訊取出空間資訊及一補償參數;根據該降混信號及該 空間資訊產生一多通道音訊信號；以及根據該補償參數補 償該多通道音訊信號。 本發明之另一態樣為一種解碼裝置，該解碼裝置包含 一解多工器、一核心解碼器、一參數編碼器以及一多通道 φ 合成單元，該解多工器自一輸入信號取出一經編碼之降混 信號及額外資訊；該核心解碼器藉解碼該經編碼之降混信 號的方式產生一降混信號；該參數編碼器自該額外資訊取 出空間資訊及一補償參數；該多通道合成單元用以根據該 降混信號及該空間資訊產生一多通道音訊信號，並利用該 補償參數補償該多通道音訊信號。 本發明之又一態樣為一種編碼方法，該解碼方法包含 下列步驟：根據一多通道音訊信號及一降混信號計算空 • 間資訊；以及根據該多通道音訊信號及該降混信號計算一 補償參數。 本發明之再一態樣為一種編碼裝置，該編碼裝置包含 一空間資訊計算單元、一補償參數計算單元以及一位元流 產生單元，該空間資訊計算單元用以根據一多通道音訊信 號及一降混信號計算空間資訊，該補償參數計算單元用以 根據該多通道音訊信號及該降混信號計算一補償參數;該 位元流產生單元用以藉編碼該空間資訊、該補償參數及該 1330992 降混信號與結合該編碼之結果的方式產生一位元流。 本發明之又再一態樣為一種電腦可讀記錄媒體，具有 一用以執行上述之解碼方法的程式記錄於上。 _ 本發明之又再一態樣為一種電腦可讀記錄媒體，具有 一用以執行上述編碼方法的程式記錄於上。 [功效] 在該編碼方法中，根據一多通道音訊信號及一降混信 號計算空間資訊，且根據該多通道音訊信號及該降混信號 參計算一補償參數。其後，一位元流被藉由編碼該空間資 訊、該補償參數及該降混信號與結合該編碼之結果的方式 產生。因此，關於一多通道音訊信號之聲音的品質破壞問 題可避免，其方式為使用該補償參數補償該多通道音訊信 號。 【實施方式】 本發明將配合顯示較佳實施例之圖式更詳細說明如 • 下。 本發明之編碼方法及其裝置與解碼方法及其裝置之 一實施例可被加至一多通道音訊信號的處理上，但卻不僅 限於該種應用。換言之，本發明亦可被加至多通道音訊信 號外之其它信號的處理上。 第二圖為本發明之一編碼裝置實施例的方塊圖。請參 閱第二圖，該編碼裝置包含一降混單元110、一補償參數 計算單元120、一空間資訊計算單元130及一位元流產生單 1330992 元170，其中該位元流產生單元170包含一核心編碼器 140、一參數編碼器150及一多工器160。 該降混單元110利用降混一輸入多通道音訊信號成一 . 單聲道信號或一立體聲信號的方式產生一降混信號。該補 償參數計算單元120將該降混單元110產生之降混信號或 一輸入人工降混信號之包封或位準與一用以產生該降混 信號或該輸入人工降混信號之多通道音訊信號之包封或 位準進行比較，並根據該比較結果計算一補償參數。該空 參間資訊計算單元130計算一多通道音訊信號的空間資訊。 位元流產生單元170之核心編碼器140編碼一降混信 號，參數編碼器150藉編碼該補償參數與該空間資訊的方 式產生額外資訊。接著，多工器160藉結合該經編碼降混 信號及該額外資訊產生一位元流。更詳而言之，該降混單 元110藉降混該輸入多通道音訊信號產生一降混信號。舉 例而言，在對一具有五個通道（即通道1至通道5)之多通道 音訊信號加以降混至一立體聲信號時，降混通道1可藉結 • 合該多通道音訊信號之通道1、3及4的方式獲得，且降混 通道2可藉結合該多通道音訊信號之通道2、3及5的方式獲 得。 當一降混信號一經產生，補償參數計算單元120便計 算一補償參數，且該補償參數可藉不同方法計算得出。舉 例而言，假設一多通道音訊信號包含五個屬於一預定頻帶 的不同通道，即通道1、2、3、4及5，1^1、1^2、1^3、1>4及 L5分別為通道1、2、3、4及5的位準，降混通道1包含通道 1330992 卜3及4，且降混通道2包含通道2、3及5，此時降混通道1 的位準D L13 4與降混通道2的位準D L2 3 5可以方程式（1)代 表之： [數學圖式1] DL134SLl+g3*L3+g4 氺 L4 DL235 ^ L3+g3*L3+g3 * L3+g5*L5 其中g3、g4及g5為在一降混過程中所產生的增益。若藉由 解碼而根據一降混信號產生一多通道音訊信號，則所產生 之多通道音訊信號之五個通道的位準LI’、L2’、L3’、L4’ 及L5’理想上分別同於一原始多通道音訊信號之五個通道 的位準L1、L2、L3、L4及L5。為達此一目的，降混通道1 之一補償參數CF123及降混通道2之一補償參數CF235可 利用方程式（2)計算之： [數學圖式2] CF134=(Ll+g3*L3+g4*L4)/DL134 CF235=(L2+g3*L3+g5*L5)/DL235 在本實施例中，每一降混通道的補償參數皆被計算， 藉以減少待傳輸之資料的量，但一多通道音訊信號之每一 通道的補償參數亦可被計算。易言之，一補償參數可被計 算為一降混信號之能量與一多通道音訊信號之每一通道 之能量的比，或可被計算為一降混信號之包封與一多通道 音訊信號之每一通道之包封的比。 1330992 空間資訊計算單元130計算空間資訊，其中空間資訊 可包含通道位準差（CLD)資訊、通道間互相關聯（ICC)資訊 及通道預測係數（CPC)資訊。 - 核心編碼器140對一降混信號加以編碼，參數編碼器 -150藉編碼空間資訊及一補償參數的方式產生額外資訊， 其中補償參數可利用對通道位準差加以編碼的相同方法 編碼。舉例而言，補償參數可利用一時間或頻率差編碼方 鲁法、分類脈碼調變（PCM)編碼方法、前驅式（pilot-based) 編碼方法、或霍夫曼碼薄方法加以編碼。多工器160藉結 合一經編碼之降混信號與額外資訊的方式產生一位元 流。以此方式為之，一包含作為額外資訊、並用以補償降 混信號之位準衰減的補償參數之位元流可被產生。 當無位準補償之需要時，一關於一補償參數的旗標可 被設成為〇值，以降低額外資訊的位元速率。若補償參數 • CF134及CF235的值之間無大差異，則該二補償參數CF134 及CF235中僅一者即可代表該二補償參數CF134及CF235 可被傳送，而非傳送時需將該二補償參數CF134及CF235 皆加傳送。此外，若一補償參數的值不隨時間改變而被均 勻維持，則一預定旗標可被用以指出一前一補償參數值可 被使用。 在本實施例中，一補償參數可根據一輸入多通道音訊 12 1330992 信號之位準及一降混信號之位準的比較結果而被設定，但 補償參數亦可利用一不同於上述之方法設定或估計之。換 言之，由於一補償參數能對一降混信號的位準相對於一用 - 以產生該降混信號之輸入多通道音頻訊號的位準衰減加 - 以模式化，故一補償參數可被定義為一位準比、波格式資 料、或一具有線性/非線性特性的增益補償值。藉將該種 數學模式化值作為一補償參數值，該補償參數值得被以有 ^ 效率的方式傳送。 第三圖為本發明之一解碼裝置實施例的方塊圖。請參 閱第三圖，該解碼裝置包含一解多工器310、一核心解碼 器320、一參數解碼器330、及一多通道合成單元340。 該解多工器310對額外資訊及來自一輸入位元流的一 經編碼降混信號加以解多工，核心解碼器320藉由解碼該 經編碼之降混信號的方式產生一降混信號，參數解碼器 • 330根據解多工器310得到的額外資訊產生空間資訊與一 補償參數，多通道合成單元340根據核心解碼器320得到之 降混信號及參數解碼器3 3 0得到之空間資訊及補償參數產 生一多通道音頻信號。 第四圖為本發明中第三圖所示解碼裝置的運作流程 圖。請參閱第三圖及第四圖，一多通道音訊信號的位元流 在步驟S400時被接收，解多工器310在步驟S405時對一經 13 1330992 編碼之降混信號及自該經接收之位元流所得之額外資訊 加以解多工。在步驟S410中，核心解碼器320藉由對該經 編碼之降混信號加以解碼的方式產生一降混信號。在步驟 • S420中，參數解碼器330藉由對該額外資訊加以解碼的方 - 式產生一補償參數及空間資訊。在步驟S430中，多通道合 成單元340根據該空間資訊及該降混信號產生一多通道音 訊信號。在步驟S440中，多通道合成單元340使用該補償 φ 參數補償該多通道音訊信號。更詳而言之，多通道合成單 元340可補償根據一降混信號及空間資訊藉由解碼所得之 複數個通道之每一者的輸出，此一補償方式如方程式（3) 所指： [數學圖式3] L1，，=L1，*CF134 L2,,-L2,*CF235 • L3”=L3，*(CF124+CF235)/2 L4，，=L4，*CF134 C5,,=L5,*CF235 其中1^’、1^’、1^3’、]^4’及1^5’為通道的能量位準，而€?124 及CF235為補償參數。 以此方式為之，在一預定頻率的信號失真可藉使用在 一解碼處理中與空間資訊同時被接收之補償參數的方式 14 1330992 避免，以使-因解碼處理而得到的多通道音訊信號可得到 適度的補償。在本實施例中，每一 " 的輪出皆可利用一 補·參數而被補償，但本發明還不僅限於此。易古之，當 每一通道的包封被當作-補償參數傳送時 要被傳送，因為空間資訊可根據關於每—通道之1封= 喊生。即便無空間資訊被接收之時，-解碼裝置可利用 -或多降混通道自—輸人降混信號中取出假”資訊，並 可根據該假空間資輯該輸人降混信號加以解竭。 第五圖為本發明之一解碼裝置實施例的方塊圖。請來 閱第五圖’該解碼裝置不使用空間資訊作為額外資訊，且 僅根據-降混信號產生一多通道音訊信號。 請參閱第五圖，該解碼裝置包含一核心解喝器別、 Γ信號框形成單元別、一空間資訊估計單元530、及-多 通道合成單元540。 核心解碼器510藉由解碼一輸入位元流的方式產生一 降混信號’並將該降混信號傳送至該信號框形成單元 520，其中該降混信號可為一利用pr〇㈣c或&㈣得到的 車歹號’但其在本發明中卻不僅限於此。 —信號框形成單元520關於核心解碼器510得到之降混 U虎的貝料加以配置’以使對應之降混信號可被以空間音 訊編碼（SAC)框組成之單元的形式同步化。在此一信號框In a typical method of encoding a multi-channel audio signal, a multi-channel audio signal is downmixed into a mono or stereo signal and the mono or stereo signal is encoded with spatial information, rather than Each channel of the channel audio signal is encoded. In this method, the spatial information is used to return to its original multi-channel audio signal. The first picture is a block diagram of a blood type system that encodes/decodes multi-channel audio signals. Please refer to the figure--the audio signal encoder includes a downmixing module and the _ parameter estimation touch, and the towel mixing module is used to generate the downmix signal, and the downmixing-multichannel audio signal is formed into a calendar. The sound or = channel signal is used for this purpose, and the space estimation module is used to generate spatial information. The system can receive - the externally processed human mixed signal, rather than the generated - downmix signal. - Audio signal decoding The spatial information generated by the parameter estimation module is interpreted, and the original multi-channel audio signal is recovered according to the interpretation result, and the audio signal encoder generates a decimation signal or a period of generation of the human-definition mixed signal. The signal level attenuation shape may be generated in the process of adding different channel signals, plus two channels of ^30992 and one level L1 and L2 respectively. If the two channels do not overlap, the i-break is biased. The shift exists, and at this time, the channel level DL12 of the two signal addition is lower than the sum of L1 and L2. In a decoding process, the attenuation of the level of a downmix signal may cause distortion of 4 。. For example, the relationship between the levels of the channels can be based on the channel level difference ((10)) information, wherein the channel level difference information is a type of space | signal indicating the channel level (10) difference. However, when the material of the downmix signal obtained by the channel addition is attenuated, the level of the decoded downmix signal will be lower than the level of the shoulder downmix signal. Due to the existence of the phenomenon, the multi-channel audio signal obtained by the solution horse may be promoted or suppressed at the predetermined frequency, which may cause the destruction of the sound ^. In addition, the level of attenuation due to the partial deviation of the signal due to the other signal may vary with the frequency domain. Therefore, the signal passes through an audio encoder and an audio decoder. The degree of distortion will also vary with frequency, and this problem cannot be completely solved by changing the energy of the down-mixed signal in the predetermined frequency domain. [Disclosure] [Technical Problem] The present invention provides an encoding method and apparatus for encoding a multi-channel audio signal by using additional information of a complementable _team and a θ mussel downmix signal. The present invention also provides a decoding method and apparatus for decoding a multi-channel audio signal using an additional information that compensates for a downmix signal. 1330992 [Technical means] One aspect of the present invention is a decoding method, the decoding method comprising the steps of: taking out a downmix signal and additional information from an input signal; taking out spatial information and a compensation parameter from the amount and the external information; Generating a multi-channel audio signal according to the downmix signal and the spatial information; and compensating the multi-channel audio signal according to the compensation parameter. Another aspect of the present invention is a decoding apparatus, comprising: a demultiplexer, a core decoder, a parametric encoder, and a multi-channel φ synthesizing unit, the demultiplexer extracting from an input signal Encoded downmix signal and additional information; the core decoder generates a downmix signal by decoding the encoded downmix signal; the parameter encoder extracts spatial information and a compensation parameter from the additional information; the multichannel synthesis The unit is configured to generate a multi-channel audio signal according to the downmix signal and the spatial information, and compensate the multi-channel audio signal by using the compensation parameter. A still further aspect of the present invention is an encoding method, the decoding method comprising the steps of: calculating spatial information according to a multi-channel audio signal and a down-mix signal; and calculating one according to the multi-channel audio signal and the down-mix signal Compensation parameters. A further aspect of the present invention is an encoding device comprising a spatial information computing unit, a compensation parameter calculating unit and a bit stream generating unit, wherein the spatial information calculating unit is configured to use a multi-channel audio signal and a The downmix signal calculates spatial information, and the compensation parameter calculation unit is configured to calculate a compensation parameter according to the multichannel audio signal and the downmix signal; the bitstream generating unit is configured to encode the spatial information, the compensation parameter, and the 1330992 The downmix signal produces a one-bit stream in a manner that combines the results of the encoding. Still another aspect of the present invention is a computer readable recording medium having a program for executing the above decoding method recorded thereon. Further, still another aspect of the invention is a computer readable recording medium having a program for executing the above encoding method recorded thereon. [Effect] In the encoding method, spatial information is calculated based on a multi-channel audio signal and a down-mix signal, and a compensation parameter is calculated based on the multi-channel audio signal and the down-mix signal parameter. Thereafter, a bit stream is generated by encoding the spatial information, the compensation parameter, and the downmix signal in combination with the result of the encoding. Therefore, the quality destruction problem of the sound of a multi-channel audio signal can be avoided by using the compensation parameter to compensate the multi-channel audio signal. [Embodiment] The present invention will be described in more detail in conjunction with the drawings showing the preferred embodiments. An embodiment of the encoding method and apparatus and decoding method thereof and apparatus therefor of the present invention can be applied to the processing of a multi-channel audio signal, but is not limited to such an application. In other words, the invention can also be applied to the processing of other signals outside of the multi-channel audio signal. The second figure is a block diagram of an embodiment of an encoding apparatus of the present invention. Referring to the second figure, the encoding apparatus includes a downmixing unit 110, a compensation parameter calculation unit 120, a spatial information calculation unit 130, and a bit stream generation unit 1330992 element 170, wherein the bit stream generation unit 170 includes a The core encoder 140, a parameter encoder 150 and a multiplexer 160. The downmixing unit 110 generates a downmix signal by downmixing an input multichannel audio signal into a mono signal or a stereo signal. The compensation parameter calculation unit 120 combines the encapsulation or level of the downmix signal generated by the downmixing unit 110 or an input artificial downmix signal with a multi-channel audio used to generate the downmix signal or the input artificial downmix signal. The envelope or level of the signal is compared and a compensation parameter is calculated based on the comparison. The inter-parameter information calculation unit 130 calculates spatial information of a multi-channel audio signal. The core encoder 140 of the bit stream generation unit 170 encodes a downmix signal, and the parameter encoder 150 generates additional information by encoding the compensation parameter and the spatial information. Next, multiplexer 160 generates a bit stream by combining the encoded downmix signal with the additional information. More specifically, the downmix unit 110 generates a downmix signal by downmixing the input multichannel audio signal. For example, when a multi-channel audio signal having five channels (ie, channel 1 to channel 5) is downmixed to a stereo signal, the downmix channel 1 can be borrowed to include the channel 1 of the multi-channel audio signal. The methods of 3, 4 and 4 are obtained, and the downmix channel 2 can be obtained by combining the channels 2, 3 and 5 of the multi-channel audio signal. When a downmix signal is generated, the compensation parameter calculation unit 120 calculates a compensation parameter, and the compensation parameter can be calculated by different methods. For example, suppose a multi-channel audio signal contains five different channels belonging to a predetermined frequency band, namely channels 1, 2, 3, 4 and 5, 1^1, 1^2, 1^3, 1>4 and L5. The levels of channels 1, 2, 3, 4, and 5 are respectively, the downmix channel 1 includes channels 1330992, 3 and 4, and the downmix channel 2 includes channels 2, 3, and 5, and the level of downmix channel 1 is at this time. D L13 4 and the level of the downmix channel 2 D L2 3 5 can be represented by equation (1): [Mathematical pattern 1] DL134SLl+g3*L3+g4 氺L4 DL235 ^ L3+g3*L3+g3 * L3+ G5*L5 where g3, g4 and g5 are the gains produced during a downmixing process. If a multi-channel audio signal is generated according to a downmix signal by decoding, the positions of the five channels of the generated multi-channel audio signal are ideally the same as LI', L2', L3', L4' and L5'. The levels L1, L2, L3, L4, and L5 of the five channels of the original multi-channel audio signal. To achieve this goal, one of the compensation parameters CF123 and one of the downmix channel 2 compensation parameters CF235 can be calculated using equation (2): [Mathematical Figure 2] CF134=(Ll+g3*L3+g4 *L4)/DL134 CF235=(L2+g3*L3+g5*L5)/DL235 In this embodiment, the compensation parameters of each downmix channel are calculated to reduce the amount of data to be transmitted, but more The compensation parameters for each channel of the channel audio signal can also be calculated. In other words, a compensation parameter can be calculated as the ratio of the energy of a downmix signal to the energy of each channel of a multichannel audio signal, or can be calculated as a packet of a downmix signal and a multichannel audio signal. The ratio of the envelope of each channel. 1330992 The spatial information calculation unit 130 calculates spatial information, wherein the spatial information may include channel level difference (CLD) information, inter-channel correlation (ICC) information, and channel prediction coefficient (CPC) information. - The core encoder 140 encodes a downmix signal, and the parameter encoder -150 generates additional information by means of coded spatial information and a compensation parameter, wherein the compensation parameters can be encoded using the same method of encoding the channel level difference. For example, the compensation parameters may be encoded using a time or frequency difference coding method, a classification pulse code modulation (PCM) coding method, a pilot-based coding method, or a Huffman code thin method. The multiplexer 160 generates a bit stream by combining the encoded downmix signal with additional information. In this manner, a bit stream containing compensation parameters as additional information and used to compensate for the level attenuation of the downmix signal can be generated. When no level compensation is required, a flag for a compensation parameter can be set to a threshold value to reduce the bit rate of additional information. If there is no large difference between the compensation parameters and the values of CF134 and CF235, only one of the two compensation parameters CF134 and CF235 can represent that the two compensation parameters CF134 and CF235 can be transmitted, instead of the two compensations when transmitting The parameters CF134 and CF235 are transmitted. Moreover, if the value of a compensation parameter is not uniformly maintained over time, a predetermined flag can be used to indicate that a previous compensation parameter value can be used. In this embodiment, a compensation parameter may be set according to a comparison result of a level of an input multi-channel audio 12 1330992 signal and a level of a downmix signal, but the compensation parameter may also be set by using a method different from the above method. Or estimate it. In other words, since a compensation parameter can be used to model the level of a downmix signal relative to the level attenuation of the input multichannel audio signal used to generate the downmix signal, a compensation parameter can be defined as A quasi-ratio, wave format data, or a gain compensation value with linear/nonlinear characteristics. By using the mathematical model value as a compensation parameter value, the compensation parameter value is transmitted in an efficient manner. The third figure is a block diagram of an embodiment of a decoding apparatus of the present invention. Referring to the third figure, the decoding apparatus includes a demultiplexer 310, a core decoder 320, a parameter decoder 330, and a multi-channel synthesizing unit 340. The demultiplexer 310 demultiplexes additional information and a coded downmix signal from an input bit stream, and the core decoder 320 generates a downmix signal by decoding the encoded downmix signal. The decoder 330 generates spatial information and a compensation parameter according to the additional information obtained by the demultiplexer 310. The multichannel synthesis unit 340 obtains the spatial information and compensation obtained by the downmix signal obtained by the core decoder 320 and the parameter decoder 303. The parameters produce a multi-channel audio signal. The fourth figure is a flow chart showing the operation of the decoding apparatus shown in the third figure of the present invention. Referring to the third and fourth figures, the bit stream of a multi-channel audio signal is received at step S400, and the demultiplexer 310 processes the downmix signal encoded by 13 1330992 at step S405 and receives the downmix signal. Additional information from the bitstream is used to solve the multiplex. In step S410, core decoder 320 generates a downmix signal by decoding the encoded downmix signal. In step S420, the parameter decoder 330 generates a compensation parameter and spatial information by decoding the additional information. In step S430, the multi-channel synthesizing unit 340 generates a multi-channel audio signal based on the spatial information and the downmix signal. In step S440, the multi-channel synthesis unit 340 compensates the multi-channel audio signal using the compensation φ parameter. In more detail, the multi-channel synthesis unit 340 can compensate the output of each of the plurality of channels obtained by decoding according to a downmix signal and spatial information, and the compensation method is as indicated by equation (3): [Mathematics Figure 3] L1,, =L1, *CF134 L2,, -L2, *CF235 • L3"=L3,*(CF124+CF235)/2 L4,,=L4,*CF134 C5,,=L5,*CF235 Where 1^', 1^', 1^3', ]^4' and 1^5' are the energy levels of the channel, and €124 and CF235 are the compensation parameters. In this way, at a predetermined frequency The signal distortion can be avoided by using the compensation parameter 14 1330992 which is received at the same time as the spatial information in the decoding process, so that the multi-channel audio signal obtained by the decoding process can be moderately compensated. In this embodiment The rotation of each " can be compensated by using a supplemental parameter, but the invention is not limited to this. Yiguzhi, when the envelope of each channel is transmitted as a compensation parameter, is transmitted. Because the spatial information can be shouted according to one per channel - even if no spatial information is received, the - decoding device can be used - or the multi-downmix channel takes the false information from the input and subtraction signal, and can be depleted according to the false space signal. Figure 5 is a block diagram of an embodiment of a decoding apparatus of the present invention. Please refer to the fifth figure'. The decoding device does not use spatial information as additional information, and only generates a multi-channel audio signal based on the -downmix signal. Referring to FIG. 5, the decoding apparatus includes a core decanter, a signal frame forming unit, a spatial information estimating unit 530, and a multi-channel synthesizing unit 540. The core decoder 510 generates a downmix signal by decoding an input bit stream and transmits the downmix signal to the signal frame forming unit 520, wherein the downmix signal can be a pr 〇 (4) c or & (4) The obtained car number 'but it is not limited to this in the present invention. The signal frame forming unit 520 is configured with respect to the downmix of the U-hub obtained by the core decoder 510 so that the corresponding downmix signal can be synchronized in the form of a unit composed of a spatial audio coding (SAC) frame. In this signal box
15 1330992 形成處理中，若正交鏡象濾波器（QMF)及混合帶域信號利 用一分析濾波帶並根據核心解碼器510所得之降混信號被 產生，則信號框形成單元520可將混合帶域信號傳送至多 - 通道合成單元540，因混合帶域信號可被輕易使用於一解 - 碼處理中。 空間資訊估計單元530根據一由信號框形成單元520 得到的降混信號產生通道位準差資訊、通道間互相關聯資 φ 訊及通道預測係數資訊等空間資訊。更詳而言之，空間資 訊估計單元530產生每一空間音訊編碼信號框所需之空間 資訊。此時，空間資訊估計單元530可收集一降混信號的 資料，直至所收集之經結合資料的長度變成與一信號框者 相同為止，並接著處理該經收集之降混信號資料。或者， 該空間資訊估計單元亦可產生每一脈碼調變（PCM)樣品 所需之空間資訊。空間資訊估計單元530所產生的空間資 _ 訊並非為待傳送之資料，故不需被加以量化等壓縮處理， 也因此空間資訊估計單元530產生的空間資訊可包含盡可 能多的資訊。 多通道合成單元540根據信號框形成單元520得到之 降混信號與空間資訊估計單元530產生之空間資訊產生一 多通道音訊信號。 在本實施例中，其位元速率可較傳送空間信號為額外 16 1330992 資訊的傳統方法所得者降低。此外，亦可使用一般用以產 生陣列式降混内容的方法產生一多通道信號。 第六圖為本發明之一解碼裝置實施例的方塊圖。請參 - 閱第六圖，當一同時包含一降混式音訊信號及空間資訊的 - 位元流被接收得時，解碼裝置根據包含於該經接收得之位 元流中的空間資訊產生額外空間資訊，並使用該額外空間 資訊解碼該降混音訊信號。 φ 請參閱第六圖，該解碼裝置包含一解多工器610、一 核心解碼器620、一信號框形成單元630、一空間資訊估計 單元640、一多通道合成單元650及一結合單元660。 解多工器610對自一輸入位元流得到之空間資訊及一 經編碼降混信號加以解多工，核心解碼器620藉由對該經 編碼之降混信號解碼方式產生一降混信號，信號框形成單 元630對關於核心解碼器510得到之降混信號的資料加以 • 配置，以使對應降混信號可以空間音訊編碼信號框組成之 單元的形式被同步化。空間資訊估計單元640根據解多工 器610得到之空間資訊透過估計方式產生額外空間資訊。 結合單元660結合解多工器610得到之空間資訊與空間資 訊估計單元640產生之額外空間資訊，並將該結合得之空 間資訊傳送至多通道合成單元650。接著，多通道合成單 元650根據該核心解碼器620產生之降混信號及該結合單 17 1330992 元660傳送之空間資訊產生一多通道音訊信號。 在本實施例中，包含在一輸入位元流之空間資訊及透 過估計處理而自一降混信號得到的額外空間資訊皆可被 ·· 使用，且隨包含於一輸入位元流中之空間資訊之類型的不 ，- 同，各種應用皆可形成，此將更詳細描述於後文中。 當包含僅數時間長度及資料帶之空間資訊被接收 時，即當空間資訊之位元速率低至使得該空間資訊之資料 φ 帶數小或該空間資訊之傳送頻率低時，空間資訊估計單元 640根據該空間資訊及一降混脈碼調變信號產生無該空間 資訊的資訊，藉以提升一多通道音訊信號的品質。舉例而 言，若包含僅五個資料帶的空間資訊被接收，則空間資訊 估計單元6 4 0可藉參考與該空間資訊一塊被接收之降混信 號將該空間資訊轉換成包含二十八個資料帶的空間資 訊。若僅包含二時間長度的空間資訊被接收，則空間資訊 • 估計單元640可藉參考一與該空間資訊一塊被接收之降混 信號並透過内插法產生總計八個時間長度。 當包含通道位準差資訊、通道間互相關聯資訊及通道 預測係數資訊之空間資訊之僅部份被接收（如僅通道間互 相關聯資訊被接收），則空間資訊估計單元640可透過估計 產生通道位準差資訊及通道預測係數資訊，藉以提升一多 通道音訊信號的品質。同樣地，當僅通道預測係數資訊被 1330992 接收時，空間資訊估計單元640可透過估計產生通道間互 相關聯貧訊。 一種編碼裝置利用一對二（OTT)或二對三（TTT)盒將 ·· 一輸入多通道信號降混成一降混信號。當對應僅部份之一 - 對二或二對三盒之空間資訊被接收時，空間資訊估計單元 640可透過估計產生對應其它一對二及二對三盒的空間資 訊，並根據該經接收之空間資訊及該經產生之空間資訊產 φ 生一多通道音訊信號，此時空間資訊的估計可在對經接收 之空間資訊加以空間音訊編碼解碼之後被執行。舉例而 言，若一具有二通道（即左通道與右通道）的降混信號及對 應二對三盒之空間資訊被接收，則空間資訊估計單元640 可根據該經接收得之降混信號的左及右通道信號產生L -通道、C(中通道及R-通道信號。 其後，空間資訊估計單元640可產生對應一對二盒之 • 空間資訊。接著，多通道合成單元650根據該經接收之空 間資訊與該空間資訊估計單元640產生之空間資訊產生一 多通道音訊信號，該方法可在輸出通道數大時應用。舉例 而言，當一具有一 525格式之位元流被輸入至一可提供多 至七個通道的解碼裝置中時，該解碼裝置透過空間音訊編 碼解碼處理產生五個通道信號（混域）、透過估計產生拓展 五個通道信號至七個通道所需之空間資訊、並另外執行解 1330992 碼處理，藉以產生一具較一單一位元流所提供者為多的通 道之信號。 本發明可以電腦可讀碼的形式實施，且該電腦可讀碼 - 係寫於一電腦讀記錄媒體中，其中該電腦可讀記錄媒體可 - 為任何一種能使資料以一電腦可讀取之方式儲存的記錄 裝置.，如可為ROM、RAM、CD-ROM、磁帶、軟碟片、 光資料儲存裝至及一載波（如透過網際網路所為之資料傳 φ 輸）等。此外，電腦可讀記錄媒體可被分散於複數個與一 網路相接的電腦系統中，此時電腦可讀碼可被寫至該等電 腦系統中，並得由該等系統處以分散方式執行。關於實施 本發明所需之功能程式、碼及碼段，其可為熟習該項技術 者所輕易完成。 本發明已透過其較佳實施例加以說明及顯示，熟習該 項技術者可知該等較佳實施例可在不違本發明之精神與 φ 特徵的條件下被加以各種改變，其中本發明之精神與特徵 將定義於下列申請專利範圍中。 [產業可利用性] 在本發明中，一解碼所得之多通道音訊信號可使用 一補償參數作為額外資訊加以補償，其中該補償參數係藉 由比較一輸入多通道音訊信號與一降混信號之位準的方 式計算。此外，本發明可根據輸入空間資訊與一輸入降混 20 1330992 信號產生餐”資訊’故―由解碼所得之多通道音訊产 號可避免在—預定頻率上產生失真，故能提升該多通道音 訊信號的品質。 在本心明中’聲音品質的破壞得藉由在-多通道音訊 U之、，扁碼及/或解碼期間使用—補償參數補償多通道音 訊信號的方式避免之。 【圖式簡單說明】 在？=例f生只化例配以所附圖式加以詳細說明後，本 及其匕特徵與優點將變得·更為明顯易懂，其中 琢等圖式之： 所示為本發明之用以解碼/編碼—多通道音訊 L 5虎之典型系統的方塊圖； =二=所4树明之—編碼U實施㈣方塊圖； =二=所不林發明之—解衫置實施例的方塊圖； 之摔ίΐϊ所不為本發明之在第三圖所示之解碼裝置上 <钿作的流程圖； 第五圖所示為本發明之一 方塊圖;以及 •馬裝置之另-實施例的 第六圖所示為本發明之一解 方塊圖。 ㉟碼裴置之另-實施例的 【主要元件符號說明】 21 降混單元 補償參數計算單元 空間資訊計算單元 核心編碼 參數編碼器 多工器 位元流產生單元 解多工器 核心解碼器 參數解碼器 多通道合成單元 步驟 步驟 步驟 步驟 步驟 步驟 核心解碼器 信號框形成單元 空間資訊估計單元 多通道合成單元 解多工器 核心解碼器 信號框形成單元15 1330992 In the forming process, if the quadrature mirror filter (QMF) and the mixed band signal are generated using an analysis filter band and the downmix signal obtained from the core decoder 510, the signal frame forming unit 520 can mix the band The domain signal is transmitted to the multi-channel synthesis unit 540, since the mixed band signal can be easily used in a solution-code process. The spatial information estimating unit 530 generates spatial information such as channel level difference information, inter-channel correlation information, and channel prediction coefficient information based on a downmix signal obtained by the signal frame forming unit 520. More specifically, the spatial information estimating unit 530 generates spatial information required for each spatial audio coded signal frame. At this time, the spatial information estimating unit 530 can collect the data of a downmix signal until the length of the collected combined data becomes the same as that of a signal frame, and then processes the collected downmix signal data. Alternatively, the spatial information estimating unit may also generate spatial information required for each pulse code modulation (PCM) sample. The spatial information generated by the spatial information estimating unit 530 is not the data to be transmitted, so it does not need to be quantized or the like, and therefore the spatial information generated by the spatial information estimating unit 530 can contain as much information as possible. The multi-channel synthesizing unit 540 generates a multi-channel audio signal based on the downmix signal obtained by the signal frame forming unit 520 and the spatial information generated by the spatial information estimating unit 530. In this embodiment, the bit rate can be reduced compared to the conventional method of transmitting spatial signals to an additional 16 1330992 information. In addition, a multi-channel signal can be generated using a method generally used to generate array downmix content. Figure 6 is a block diagram of an embodiment of a decoding apparatus of the present invention. Referring to the sixth figure, when a bit stream containing both a downmix audio signal and spatial information is received, the decoding device generates an extra based on the spatial information contained in the received bit stream. Spatial information, and use the additional spatial information to decode the downmixed audio signal. φ Referring to FIG. 6, the decoding apparatus includes a demultiplexer 610, a core decoder 620, a signal frame forming unit 630, a spatial information estimating unit 640, a multi-channel synthesizing unit 650, and a combining unit 660. The demultiplexer 610 demultiplexes the spatial information obtained from an input bit stream and the encoded downmix signal, and the core decoder 620 generates a downmix signal by decoding the encoded downmix signal. The frame forming unit 630 configures the data about the downmix signal obtained by the core decoder 510 so that the corresponding downmix signal can be synchronized in the form of a unit composed of spatial audio coded signal frames. The spatial information estimating unit 640 generates additional spatial information by estimating the spatial information obtained by the demultiplexing 610. The combining unit 660 combines the spatial information obtained by the multiplexer 610 and the additional spatial information generated by the spatial information estimating unit 640, and transmits the combined spatial information to the multi-channel synthesizing unit 650. Then, the multi-channel synthesis unit 650 generates a multi-channel audio signal according to the downmix signal generated by the core decoder 620 and the spatial information transmitted by the combination unit 17 1330992 660. In this embodiment, the spatial information contained in an input bit stream and the additional spatial information obtained from a downmix signal through the estimation process can be used and included in the space in an input bit stream. The types of information are not, - the same, various applications can be formed, which will be described in more detail later. When the spatial information including the length of time and the data band is received, that is, when the bit rate of the spatial information is so low that the data φ band of the spatial information is small or the transmission frequency of the spatial information is low, the spatial information estimating unit The 640 generates information without the spatial information according to the spatial information and a down-mix pulse code modulation signal, thereby improving the quality of a multi-channel audio signal. For example, if spatial information including only five data bands is received, the spatial information estimating unit 600 may convert the spatial information into twenty-eight by referring to the downmix signal received with the spatial information. Spatial information of the data band. If spatial information containing only two lengths of time is received, the spatial information estimating unit 640 can generate a downmix signal that is received along with the spatial information and generate a total of eight time lengths by interpolation. When only part of the spatial information including the channel bit difference information, the inter-channel inter-correlation information, and the channel prediction coefficient information is received (eg, only the inter-channel inter-correlation information is received), the spatial information estimating unit 640 can generate the channel through the estimation. The position difference information and the channel prediction coefficient information are used to improve the quality of a multi-channel audio signal. Similarly, when only the channel prediction coefficient information is received by 1330992, the spatial information estimating unit 640 can generate inter-channel cross-correlation by estimation. An encoding device downmixes an input multichannel signal into a downmix signal using a pair of two (OTT) or two to three (TTT) boxes. When spatial information corresponding to only one of the two-to-two or two-to-three-boxes is received, the spatial information estimating unit 640 can generate spatial information corresponding to the other pair of two and two pairs of three boxes by estimation, and according to the received The spatial information and the generated spatial information product produce a multi-channel audio signal, and the spatial information estimate can be performed after spatially encoding and decoding the received spatial information. For example, if a downmix signal having two channels (ie, a left channel and a right channel) and spatial information corresponding to two pairs of three boxes are received, the spatial information estimating unit 640 may be configured according to the received downmix signal. The left and right channel signals generate L-channel, C (middle channel and R-channel signals. Thereafter, the spatial information estimating unit 640 can generate spatial information corresponding to a pair of two boxes. Then, the multi-channel synthesizing unit 650 according to the The received spatial information and the spatial information generated by the spatial information estimating unit 640 generate a multi-channel audio signal, and the method can be applied when the number of output channels is large. For example, when a bit stream having a 525 format is input to When a decoding device of up to seven channels is provided, the decoding device generates five channel signals (mixed domain) through spatial audio codec decoding processing, and generates spatial information required to expand five channel signals to seven channels through estimation. And additionally performing the 1330992 code processing to generate a signal that is more than one channel provided by the single bit stream. The invention can be read by a computer The code is implemented in the form of a computer readable code - written in a computer read recording medium, wherein the computer readable recording medium can be any recording device that enables the data to be stored in a computer readable manner. For example, it can be ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage and a carrier (such as data transmission through the Internet), etc. In addition, the computer readable recording medium can be Distributed among a plurality of computer systems connected to a network, at which time computer readable codes can be written to the computer systems and executed in a distributed manner by the systems. Functions required to implement the present invention The present invention has been described and illustrated by the preferred embodiments thereof, and those skilled in the art will appreciate that the preferred embodiments may be Various changes are made in the spirit of the invention and the characteristics of φ, and the spirit and features of the present invention will be defined in the following claims. [Industrial Applicability] In the present invention, a decoded multi-channel audio is obtained. The signal can be compensated for by using a compensation parameter that is calculated by comparing the level of an input multi-channel audio signal with a down-mix signal. Furthermore, the present invention can be based on input spatial information and an input. Downmix 20 1330992 signal to generate meal "information" Therefore, the multi-channel audio production number obtained by decoding can avoid distortion at the predetermined frequency, so the quality of the multi-channel audio signal can be improved. The destruction can be avoided by using the compensation parameter to compensate the multi-channel audio signal during the multi-channel audio U, flat code and/or decoding. [Simplified illustration] In the case of ? = example The features and advantages of the present invention will become more apparent and obvious, and the drawings and the advantages of the present invention are shown in the drawings for decoding/encoding-multi-channel audio. Block diagram of a typical system of L 5 Tiger; = 2 = 4 trees - code U implementation (four) block diagram; = 2 = Suzuki invention - block diagram of the embodiment of the shirt; The flowchart of the invention is shown in the third embodiment; the fifth diagram is a block diagram of the present invention; and the sixth embodiment of the horse apparatus is shown in the sixth diagram. One of the inventions is a block diagram. The other part of the 35-coded device--the main component symbol description 21 The downmixing unit compensation parameter calculation unit space information calculation unit core coding parameter encoder multiplexer bit stream generation unit solution multiplexer core decoder parameter decoding Multi-channel synthesis unit step step step step step step core decoder signal frame formation unit space information estimation unit multi-channel synthesis unit solution multiplexer core decoder signal frame forming unit
22 1330992 640 空間資訊估計單元 650 多通道合成單元 660 結合單元22 1330992 640 Spatial Information Estimation Unit 650 Multi-Channel Synthesis Unit 660 Combination Unit
23twenty three