201204049 六、發明說明： 【發明所屬之技術領域】 本發明有關於一種自適應視頻解碼(adaptively performing video decoding)方法，尤其有關於一種自適應視頻解碼方法、 複雜度自適應視頻解碼器（adaptive complexity video deeoderj 以及自適應音頻/視頻播放系統。 【先前技術】 當終端使用者在家使用家庭影院系統觀看視頻節目時，視 頻節目的編碼資料被解碼以顯示在家庭影院系統的顯示面板 上，且編碼資料的解碼結果可擴大或縮小以符合顯示面板的尺 寸和解析度的要求。通常’不管對編碼資料進行解碼的算法多 麼複雜’也0視頻節目是否高崎度，家庭影院㈣的解碼 裝置可輕易地對編碼資料進行解碼。 對於終端制者練，财可在任何財觀看囉視鮮目的可 攜式電子裝置是非常讀的，該可職子錢可例如行動電話或者個 A^^1(Pers〇nalDigitalAss^ , 開錄eSearchandDevelGpment，R&D)騎在設計這種 署、 時’可能出現-些問題。例如，在對編碼資料進行解碼 : ^見頻節峨析度咖航下，蝴侧啸碼能力^可 月,不足。又例如’當可攜電子裝置⑽—個或多個處理電路工作在最 201204049 需要高效率且价格 同可用工作頻率時，功率雜有可能過高 。因此， 以使用具有有限計算能力的可攜電子裝置播放具 可取的視頻解竭器， 有可接受品質的视頻 【發明内容】 有鑑於此，本發明提供一種自適應視頻解碼方法、 自適應視頻解碼n以及自適應音頻/視頻播放系統。’、又 本發明提供—種自適應視頻解碼方法，包含：基於一 雜度自適應視頻解碼器的一輸入位元流的位元流土 : :· :碼，雜度管理’以決定是否降低該複雜度自適應視頻解: 益内多個組件中的至少—組件的解碼複雜度；以及選擇性降 低該複雜度自適應視頻解碼器内多個組 的解碼複雜度。 a h组件 本發明另提供一種複雜度自適應視頻解碼器，包含：夕 個組件’用於執行與視頻解碼有關的多個操作；以及一解石夕 複雜度管理ϋΐ於基於—輸人位元流的位元流資訊執行解 碼複雜度f理’以決定是否降低該複雜度自適應視頻解媽器 内至少一組件的解碼複雜度；其中，該解碼複雜度管理器 擇性降低該複雜度自適應視頻解碼器内多個組件中的一邻 分組件的解碼複雜度。 本發明還提供一種自適應音頻/視頻播放系統，包含· 201204049 3頻解竭器，用於勃行咅并 視頻解碼器田 頻解碼以產生音頻資訊；以及- 一解碼管理$ »4 馬其中，该視頻解碼器包含 訊解碼管理器依據-輸入位元流的位元流資 錢遲該音頻資訊的音頻播放。 ^ 解碼ίΓΓΓΐ供之自適應視頻解媽方法、複雜度自適應視頻 雜度二音頻/視頻播放系統可自適應地降低解石馬複 視頻播放hi W作量和相應的功率消耗，為實現數位 頁播放系統供了的極大靈活性。 以下係依據多個圖 域習知技藝者閱讀後應 式對本發明之較佳實施例進行詳細描述 可明確了解本發明之目的。 ，本領 【實施方式】 在說明書及申請專利範圍當中使用了某些詞彙來指稱特 =件。所屬領域中具有通常知識者應可理解，硬體製造商可 月廿不同名詞來稱呼同—個組件。本說明書及中請專利範圍 、'不以名稱之差異來作為區分組件之方式，而是以組件在功能 之^異來作為區分之準則。在通篇說明書及中請專利範圍當 =提及之&含」為一開放式用語，故應解釋成「包含但不 限定於」。「大致」係指在可接受之誤差範圍内，所屬領域中具 有通常知識者賴在—定誤絲_解決所述技術問題，基本 達到所述技術效果。此外，「減」—詞在此包含任何直接及 間接之電性連接手段。因此，若文中描述一第—裝置搞接於一 201204049201204049 VI. Description of the Invention: [Technical Field] The present invention relates to an adaptively performing video decoding method, and more particularly to an adaptive video decoding method and an adaptive adaptive video decoder (adaptive complexity) Video deeoderj and adaptive audio/video playback system. [Prior Art] When an end user watches a video program using a home theater system at home, the encoded content of the video program is decoded to be displayed on the display panel of the home theater system, and the encoded data is encoded. The decoding result can be expanded or reduced to meet the requirements of the size and resolution of the display panel. Usually, 'how complicated the algorithm for decoding the encoded data' is also 0. Whether the video program is high or not, the decoding device of the home theater (4) can be easily Encoding data for decoding. For terminal manufacturers, it is very readable to watch the fresh portable electronic devices in any financial situation. The available money can be, for example, a mobile phone or an A^^1 (Pers〇nalDigitalAss ^ , start eSearchandDevelGpment, R&D) In the design of such a department, 'may arise some problems. For example, in the decoding of the encoded data: ^ see the frequency of the analysis of the degree of caliber, the side of the whistle code ability ^ can be monthly, insufficient. For example, 'When available With the electronic device (10) - one or more processing circuits working at the most 201204049 need high efficiency and the price is the same as the available operating frequency, the power may be too high. Therefore, it is desirable to play with portable electronic devices with limited computing power. Video Decompressor, Video with Acceptable Quality [Invention] In view of this, the present invention provides an adaptive video decoding method, an adaptive video decoding n, and an adaptive audio/video playback system. An adaptive video decoding method includes: a bit stream based on an input bit stream of a noise adaptive video decoder: :· : code, noise management' to determine whether to reduce the complexity adaptive video solution : at least one of a plurality of components - decoding complexity of the component; and selectively reducing decoding complexity of the plurality of groups within the adaptive video decoder. The present invention further provides a complexity adaptive video decoder, comprising: a component for performing a plurality of operations related to video decoding; and a solution of the complexity of the solution based on the input bit stream The meta-stream information performs decoding complexity to determine whether to reduce the decoding complexity of at least one component in the complexity adaptive video decoder; wherein the decoding complexity manager selectively reduces the complexity adaptive video decoding The decoding complexity of one of the plurality of components in the device. The present invention also provides an adaptive audio/video playback system, including the 201204049 3 frequency decompressor for the gambling and video decoder field frequency decoding. To generate audio information; and - a decoding management $ » 4 horse, wherein the video decoder comprises a video decoding manager according to the bit stream of the input bit stream to play the audio information of the audio information. ^ Decoding 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应 自适应The playback system provides great flexibility. The purpose of the present invention will be apparent from the following detailed description of the preferred embodiments of the invention. [Embodiment] Some terms are used in the specification and the scope of the patent application to refer to the special item. Those of ordinary skill in the art should understand that a hardware manufacturer can refer to the same component by different terms. In this manual and in the scope of patents, 'the difference between the names is not used as a means of distinguishing components, but the component is differentiated according to the function. In the entire specification and in the scope of patents, = "included" is an open term and should be interpreted as "including but not limited to". "Roughly" means that within the acceptable margin of error, those of ordinary skill in the art rely on the problem of solving the technical problems to achieve the technical effects. In addition, the word "subtract" - hereby includes any direct and indirect electrical connection means. Therefore, if the text describes a first device, it is connected to a 201204049
發明之保護範圍當視所附申料利範圍所界定者為準。 為實施本發明之較佳實施方式，然該描述乃以說 般原則為目的’並非用於限定本發明之範圍。本The scope of protection of the invention is subject to the definition of the scope of the attached application. The description of the preferred embodiments of the present invention is intended to be a this
Length Decoding，VLD)與逆量化模組 11〇、 I解碼器100的方塊示意圖。 包含快速變長解碼（Variable t組11 〇、逆變換單元（inverse 論也⑽㈣、運動補償（M〇ti〇n c〇mpensati〇n，mc)模組13〇 (在第1A圖中標記為rMC模組」）、計算單元刚、重構訊框 輸出單元150及訊框儲存單元16〇，其中’逆變換單元12〇可 例如離散餘弦逆變換（lnverse Discrete c〇sine 丁咖#贿，idct) 單元；運動補償模組13〇包含時間預測單元132以及空間預測 單元134;計算單A i4G可例如加*器；重構訊框輸出單元15〇 包含解區塊（de-blocking)濾波器，例如迴路濾波器（in l〇〇p filter)152,以及訊框儲存單元16〇可例如訊框緩衝器。特別地， 前述的解區塊遽波H (例如迴路遽波器152)係複雜度自適應 解區塊濾波器。請注意，可將時間預測單元132和空間預測單 元134分別稱為訊框間預測（inter predicti〇n)單元與訊框内預測 (intra prediction)單元。如第1B圖所示，複雜度自適應視頻解 碼器1〇〇進一步包含縮放器170、顯示時間調整驅動器175、 201204049 顯示緩衝器180以及解碼複雜度管理器100M。其中，解碼複 雜度管理器100M可採用硬體來實現，亦可採用軟體來實現。 根據本發明的一種實施方式，解碼複雜度管理器100M的内部 結構可配置為硬體計算器/處理機，用於計算音頻/視頻播放系 統所能允許的解碼複雜度，並根據計算結果對複雜度自適應解 碼器100的解碼複雜度進行調整。例如該計算器/處理機可選 擇性降低複雜度自適應視頻解碼器1〇〇内多個組件中至少一部 分組件的解碼複雜度。然以上僅用於說明目的，解碼複雜度管 理器100M的實施方式並不以此為限。 依據此實施例，快速變長解碼與逆量化模組11〇係用於對 輸入位元流H)8執行快速VLD與逆量化以產生多個逆量化結 果m，而逆變換單幻20係用於對多個逆量化絲ιΐ8執行 逆變換以產生多個逆變換結果128。另外，運動補償模組13〇 係用於依據輸人位元流1G8執行運動補償，並產生相庳的多個 預測輸出結果138’其中’時間預測單元132係用純行時間 預測乂及空。間預測單元134係用於執行空間預測。如第Μ 二：二算早^ 140係用於合計多個逆變換結果128和多個 ==138以產生多個補償輸出結果148，而重構訊框 輸出皁兀係用於產生多個重構訊框158 元134依據重構却 _ 兵甲工間預刺早 測。另外，訊框儲二王輸出早疋150輸出的重構資料執行空間預 、 〔儲存單元⑽係用於臨時儲存多個重構訊框158 201204049 的至少一部分，其中，時間預測單元132依據前述的多個重構 訊框158的至少一部分執行時間預測。 參照第1B圖，縮放器17〇係用於在有需要時執行影像縮 放（例如放大或縮小影像）。在執行影像縮放的情況下，多個 影像訊框178包含多個重構訊框158的縮放版本。在不執行影 像縮放的情況了，多個影像訊框178包含多個重構訊框158， 即，縮放器170旁路多個重構訊框158。另外，顯示緩衝器 係用於臨時儲存多個影像訊框178。在顯示時間調整驅動器 的控制下，輸出臨時儲存在顯示緩衝器18〇中的多個影 Π8作為多個輸出訊框188，以顯示在減於音頻資訊的音頻 播放的適當時間點上。因此，經由使用顯示時間調整驅動器 Π5,複雜度自適應解碼器1〇〇可同步音頻/視頻播放系統（第 1A圖和第1B圖中未不）的音頻和視頻播放，以消除視頻解碼 的延遲（若存在）。例如，當存在繁重的視頻解碼工作量時， 可能會發生視頻解竭的延遲。另外，解碼複雜度管理器i〇〇m 可基於輸入位元流108和來自顯示緩衝器18〇的顯示緩衝器狀 態執行解碼複雜度管理。例如，解碼複雜度管理器刚Μ可選 擇性降低複雜度自適應視頻解喝器·内多個組件中的一部分 組件的解碼複雜度。更具體地，解碼複雜度管理器1觀可依 ，輸入位元流108和前述的顯示緩衝器狀態，產生多個控制訊 =如控制訊號。卜㈣……和〜分別控制 複雜度自適應視頻解碼HHK)内的相應組件的解碼複雜度。例 201204049 如，如第1A圖和第1B圖所示，快速變長解碼與逆量化模組 no受控於控制訊號（η，逆變換單元120受控於控制訊號c'2， 運動補償模組130受控於控制訊號C3,重構訊框輸出單元15〇 受控於控制訊號C5，前述的解區塊濾波器（例如迴路濾波器 152)受控於控制訊號C6，以及顯示時間調整驅動器丄乃受控 於控制訊號C7。 工 特別地，複雜度自適應視頻解碼器1〇〇的至少一部分， 例如時間預測單元132、空間預測單元134、逆變換單元 120、重構訊框輸出單元15〇和/或快速VLD與逆量化模組 11〇，可依據多個重構訊框158的解析度進行操作，而不是 僅依據輸入位元流1〇8所代表的多個原始訊框的解析度進 行操作，以降低解碼操作的複雜度。可經由對訊框降低取樣 (down sampling)獲取解析度降低的訊框，例如，經由選擇特 定像素以代表訊框的每個區塊，例如選擇每個2χ2像素區塊 的右下角像素。在一實施例中，時間預測單元132可依據多 個重構訊框158的解析度進行操作，而不是依據多個原始訊 忙的解析度進行操作，以降低時間預測的複雜度。例如，時 間預測單兀132可進一步估計已省略的部分資訊中可用於 執行時間預測的至少一部分資訊。另外，空間預測單元 可依據多個重構訊框158的解析度進行操作，而不是依據多 個原始訊框的解析度進行操作，以降低空間預測的複雜度。 例如，空間預測單元丨34可進一步估計已省略的部分資訊中 201204049 可用=執行空間預測的至少一部分資訊。依據此實施例，逆 變換單元120可依據多個重構訊框I%的解析度進行操作， 而不疋依據多個原始訊框的解析度進行操作， 以降低逆變換 的複雜度。例如，逆變換單元12〇可經由選擇多個預設的用 於逆變換的逆變換函數Fit中的_個以省略逆變換的多個計 算的一部分。另外，重構訊框輸出單元150可依據多個重構 訊框158的解析度進行操作，而不是多個原始訊框的解析度 進灯操作，以降低產生多個重構訊框的複雜度。基於是否滿 足一個或多個條件，迴路濾波器152可相應於多個複雜度級 別中的-個選擇性執行迴路據波用於解區塊，其中，解區塊 操作可例如，全部解區塊、部分解區塊以及跳過（skip)解區 塊另外陕速變長解碼與逆量化模組11 〇可依據多個重構 訊框158的解析度進行操作，而不是依據多個原始訊框的解 析度進行操作，以降低執行快速VLD與逆量化的複雜度。 更特別地，為降低複雜度，快速變長解碼與逆量化模組ιι〇 可在解碼期間使用包含主表和至少一子表(例如一個或多個 子表）的查找表，其中，由於在複雜度自適應視頻解碼器 100的設計階段對查找表的預設安排，使用主表的概率大於 使用至少一子表的概率。 依據此實施例的一些變形，複雜度自適應視頻解碼器 的至少一部分組件可省略該至少一部分組件需處理的部分資 11 201204049 訊’以降低對輸入位元流108進行解蝎的複雜度。簡潔起見， 有關此類變形的類似描述此處不再重複。 第1C圖係依據第一實施例的自適應音頻/視頻播放系統 50的方塊示意圖。自適應音頻/視頻播放系統5〇包含音頻解竭 器52和複雜度自適應視頻解碼器1 〇〇。音頻解碼器52传用於 執行音頻解碼以產生音頻資訊58。另外，複雜度自適應視頻解 碼器100可執行第一實施例中所揭示的視頻解碼。請注意，解 碼複雜度管理器100M係用於在有需要時對輸入位元流1〇8執 行解碼複雜度管理’且解碼複雜度管理器100M選擇性降低複 雜度自適應視頻解碼器100内的多個組件中的一部分組件的解 碼複雜度。例如’解碼複雜度管理器100M降低複雜度自適廯 視頻解碼器100内一個或多個組件的解碼複雜度。又例如，解 碼複雜度管理器100M延遲音頻資訊58的音頻播放，而不降 低複雜度自適應視頻解碼器100内任何組件的解碼複雜度。因 此’在視頻解碼器不能及時地解碼並輸出與音頻播放相匹配的 多個訊框的情況下，解碼複雜度管理器1〇〇Μ可適當延遲自適 應音頻/視頻播放系統50的音頻播放’以保持自適應音頻/視頻 播放系統50的視頻播放和音頻播放之間的音頻/視頻播放同 步。如第1C圖所示，解碼複雜度管理器ιοοΜ產生至少一控 制訊號C0以控制是否延遲音頻資訊58的音頻播放和/或音頻 資§fl 58的音頻播放的延遲量。實際上，解碼複雜度管理器1 oom 可依據輸入位元流108的位元流資訊、代表處理能力的參數及 12 a 201204049 顯示緩衝器狀態的組合或其中之一，決定延遲音頻資訊58的 音頻播放。例如，依據輸入位元流108的位元流資訊、代表處 理能力的參數及顯示緩衝器狀態的組合或其中之一，解碼複雜 度管理器100M決定是否延遲音頻資訊58的音頻播放。又例 如，依據輸入位元流108的位元流資訊、代表處理能力的參數 及顯示緩衝器狀態的組合或其中之一，解碼複雜度管理器 100M決定音頻資訊58的音頻播放的延遲量。簡潔起見，有關 此類變形的類似描述此處不再重複。 第1D圖係依據本發明第二實施例的複雜度自適應視頻解 碼器100-1的方塊示意圖。此實施例係第一實施例的變形。如 第1D圖所示，將前述的解區塊濾波器（例如迴路濾波器152) 安置在此實施例的重構訊框輸出單元150之外，而不是安置在 重構訊框輸出單元150之内。請注意，複雜度自適應視頻解碼 器100-1内的其他組件（例如快速變長解碼與逆量化模組110、 逆變換單元120、運動補償模組130、計算單元140、訊框儲存 單元160)的實施可與第1A圖所示的複雜度自適應視頻解碼 器100内相應的組件相同。在此實施例中，第1B圖中的多個 元件也可與第1D圖耦接。簡潔起見，有關此實施例的類似描 述此處不再重複。 依據此實施例的一些變形，複雜度自適應視頻解碼器 100-1的至少一部分組件可省略該至少一部分需處理的部分資 13 201204049 訊’以降低解碼輸入位元流108的複雜度。簡潔起見，有關此 類變形的類似描述此處不再重複。 第1E圖係依據第二實施例的自適應音頻/視頻播放系統 60的方塊示意圖。自適應音頻/視頻播放系統60包含音頻解碼 器62和複雜度自適應視頻解碼器100-1。音頻解碼器62係用 於執行音頻解碼以產生音頻資訊68。另外，複雜度自適應視頻 解碼器100-1可執行第二實施例中所揭示的視頻解碼。請注 意’解碼複雜度管理器100M係用於在有需要時對輸入位元流 108執行解碼複雜度管理，以及解碼複雜度管理器ι〇〇Μ選擇 性降低複雜度自適應視頻解碼器100-1内的多個組件中的一部 分組件的解碼複雜度。例如，解碼複雜度管理器1〇〇M降低複 雜度自適應視頻解碼器100-1内的一個或多個組件的解褐複雜 度。又例如，解碼複雜度管理器100M延遲音頻資訊68的音 頻播放，而不降低複雜度自適應視頻解碼器100_丨内任何組件 的解碼複雜度。因此，解碼複雜度管理器100M適當延遲自適 應音頻/視頻播放系統60的音頻_，以保持自適應音頻/視頻 播放系統6G的視_放和音賴放之間的音頻/視頻播放同 步，從而解決任何視頻解碼延遲問題。如第1E圖所示，解石馬 複雜度管理器10GM產生至少-㈣訊號，例如前述的至少_ 控制訊號CG’以控制是否延遲音頻資訊68的音頻播放和/或音 頻資訊68的音頻播放的延遲量。實際上，解碼複雜度管理^ 100M可依據例如第1C圖中所示的實施例中所述的輸入位元 201204049 流108的位元流資訊、代表處理能力的參數及顯 。^ 的組合或其中之一，決定延遲音頻資訊68的音 ^ 肖难教^。例如’ 依據輸入位元流108的位元流資訊、代表處理能力 示緩衝器狀態的組合或其中之一，解碼複雜度管理器參數及顯 定是否延遲音頻資訊68的音頻播放。又例如，二〇〇M決 抓108的位兀&貝訊、代表處理能力的參數及顯示緩衝器狀熊 的組合或其中之-，解碼複雜度管理器丨_決定音頻資訊二 的音頻播放的延遲量。簡潔起見，有關此變形的類似描述此處 不再重複。 經由使用上述揭示的任何實施例/變形的架構，可—周整各 個組件的視頻解碼複雜度。例如，在解碼複雜度管理器^漏 了二:Γ:(或其變形之一)的複雜度自適應視頻 解碼态100内或第二實施例（或其變 當決疋不需要降低第一實施例(或其變 雜度自適應視頻解碼器内或第二實施例（或其變形之一） 的硬雜度自適應視頻解碼器1(ΚΜ内任何組件的解碼複雜度 時’複雜度自適應視頻解碼器丨⑽或複雜度自適應視頻解碼器 刚-1内的所有組件可依據最高級別的視頻解碼複雜度進行操 作。參照第2圖’更多的細節進—步描述如下。 201204049 第2圖係依據本發明的—實施例的自適應視頻解碼方法 910的流程圖。自適應視頻解碼方法则可適用於第ia圖和 第1B圖所不的複雜度自適應視頻解竭器議或第⑴圖所示 的複雜度自適應視頻解碼器1G(M，更特別地，適用於上述任 何實施例/變形所揭_解碼複雜”理器議M。自適應視頻 解瑪方法910的描述如下。 在步驟S912中，解碼複雜度管理器1〇觀基於考量中的 複雜度自適應視頻解碼器（例如第1施例的複雜度自適库視 頻解碼器⑽或第二實施例的複雜度自適應視頻解碼器ι〇〇ι) 的輸入位元流108執行解碼複雜度管理，以決定是否降低上述 複雜度自適應視頻解碼器内多個組件中的至少—組件的解碼 複雜度。例如’該多個組件可包含快速⑽與逆量化模組 110、逆變換單元U0、運動補償模組13G、重構訊框輸出單元 150、前述的解區塊濾波器（例如迴路濾波器152)以及 間調整驅動器175。 在步驟S914中，解碼複雜度管理器1〇〇M選擇性降低步 驟別12中所述的複雜度自適應視頻解碼器内多個組件中的1 邰刀組件的解碼複雜度。例如，解碼複雜度管理器1⑻Μ在特 定時間點可選擇性地降低快速VLD與逆量化模組11G、逆變換 單,120運動補償模組13〇、重構訊框輸出單元150、前述的 解區塊渡波器（例如迴路滤波器152)及顯示時間調整驅動器 201204049 175中的至少一部分（例如一部分或全部）的解碼複雜度。又 例如，解碼複雜度管理器100M在另一時間點不降低快速VLD 與逆量化模組11〇、逆變換單元120、運動補償模組13〇、重構 訊框輸出單元150、前述的解區塊濾波器（例如迴路滤波器152) 及顯示時間調整驅動器175中任何一個的解碼複雜度。 依據此實施例，解碼複雜度管理器100M可經由參考顯示 緩衝器180的顯示緩衝器狀態、步驟S912中所述的複雜度自 適應視頻解碼器的一個或多個代表處理能力的參數以及輸入 位元流10 8的位元流資訊的組合或其中之一來執行解碼複雜度 管理，其中，位元流資訊可包含位元率、解析度和/或訊框類 型。位元流資訊在一些其他實例中可包含多個運動向量、多個 殘差（residual)、多個IDCT係數、位元流長度或可獲得的位元 流資訊，其中，多個殘差可例如直流電流(Direct Current，DC) 值’可獲得的位元流資訊可基於例如比較和/或歷史統計由上 述已知位元流的資訊所獲得。在一實施例中，解碼複雜度管理 器100M依據多個運動向量的變化調整解碼複雜度。代表處理 能力的參數可包含每個訊框所需的解碼時間的統計，例如，解 碼複雜度管理器100M可使用每個訊框的平均解碼時間以判斷 當前系統能力，從而適應性調整解碼複雜度。特別地，解碼複 雜度管理器100M可基於前述的顯示緩衝器18〇的顯示緩衝器 狀態、前述的至少一參數以及輸入位元流108的位元流資訊的 全部執行解碼複雜度管理。 17 201204049Length Decoding, VLD) and block diagram of the inverse quantization module 11〇, I decoder 100. Includes fast variable length decoding (Variable t group 11 〇, inverse transform unit (inverse also (10) (four), motion compensation (M〇ti〇nc〇mpensati〇n, mc) module 13〇 (marked as rMC mode in Figure 1A) a group, a calculation unit, a reconstructed frame output unit 150, and a frame storage unit 16A, wherein the 'inverse transform unit 12' can be, for example, a discrete cosine inverse transform (lnverse Discrete c〇sine Ding## bribe, idct) unit The motion compensation module 13A includes a temporal prediction unit 132 and a spatial prediction unit 134; the calculation unit A i4G may be, for example, a processor; the reconstruction frame output unit 15A includes a de-blocking filter, such as a loop. The filter (in l〇〇p filter) 152, and the frame storage unit 16 can be, for example, a frame buffer. In particular, the aforementioned solution block chopping H (for example, the chopper chopper 152) is adaptive to complexity. The deblocking filter is noted. Note that the temporal prediction unit 132 and the spatial prediction unit 134 may be referred to as inter-inter prediction units and intra prediction units, respectively, as shown in FIG. 1B. Complexity adaptive video decoder 1 The device further includes a scaler 170, a display time adjustment driver 175, a 201204049 display buffer 180, and a decoding complexity manager 100M. The decoding complexity manager 100M may be implemented by using hardware or by using software. In an embodiment of the invention, the internal structure of the decoding complexity manager 100M can be configured as a hardware calculator/processor for calculating the decoding complexity allowed by the audio/video playback system, and the complexity is calculated according to the calculation result. Adapting to the decoding complexity of the decoder 100. For example, the calculator/processor can selectively reduce the decoding complexity of at least some of the components of the complexity adaptive video decoder 1 。. For the purpose of description, the implementation of the decoding complexity manager 100M is not limited thereto. According to this embodiment, the fast variable length decoding and inverse quantization module 11 is used to perform fast VLD on the input bit stream H)8. Inverse quantization to generate a plurality of inverse quantization results m, and inverse transform single magic 20 is used to perform inverse transformation on the plurality of inverse quantization wires 以 8 to generate multiple The inverse transform result 128. In addition, the motion compensation module 13 is configured to perform motion compensation according to the input bit stream 1G8, and generate a plurality of predicted output results 138', wherein the 'time prediction unit 132 uses the pure line time. The prediction 乂 空 。 。 。 。 。 。 。 。 。 。 。 。 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 140 The reconstructed frame output saponin is used to generate a plurality of reconstructed frames 158 yuan 134 according to the reconstruction _ armored work pre-stab early detection. In addition, the frame storage second king outputs the reconstructed data execution space pre-received by 150, and the storage unit (10) is configured to temporarily store at least a part of the plurality of reconstructed frames 158 201204049, wherein the temporal prediction unit 132 is based on the foregoing At least a portion of the plurality of reconstructed frames 158 perform temporal prediction. Referring to Fig. 1B, the scaler 17 is used to perform image zooming (e.g., zooming in or out) when necessary. In the case of performing image scaling, the plurality of video frames 178 include a scaled version of the plurality of reconstructed frames 158. In the event that image scaling is not performed, the plurality of video frames 178 include a plurality of reconstructed frames 158, i.e., the scaler 170 bypasses the plurality of reconstructed frames 158. In addition, the display buffer is used to temporarily store a plurality of video frames 178. Under the control of the display time adjustment driver, a plurality of images 8 temporarily stored in the display buffer 18A are output as a plurality of output frames 188 for display at appropriate points in time for audio playback reduced to audio information. Therefore, by using the display time adjustment driver Π5, the complexity adaptive decoder 1 can synchronize audio and video playback of the audio/video playback system (not shown in FIGS. 1A and 1B) to eliminate the delay of video decoding. (if it exists). For example, when there is a heavy video decoding effort, a delay in video exhaustion may occur. Additionally, the decoding complexity manager i〇〇m can perform decoding complexity management based on the input bitstream 108 and the display buffer state from the display buffer 18A. For example, the Decoding Complexity Manager has been chosen to selectively reduce the decoding complexity of some of the components within the Complex Adaptive Video Decompositioner. More specifically, the decoding complexity manager 1 can input a bit stream 108 and the aforementioned display buffer state to generate a plurality of control signals, such as control signals. Bu (four) ... and ~ respectively control the decoding complexity of the corresponding components within the complexity adaptive video decoding HHK). Example 201204049 For example, as shown in FIG. 1A and FIG. 1B, the fast variable length decoding and inverse quantization module no is controlled by the control signal (n, the inverse transform unit 120 is controlled by the control signal c'2, the motion compensation module The 130 is controlled by the control signal C3, and the reconstructed frame output unit 15 is controlled by the control signal C5. The aforementioned deblocking filter (for example, the loop filter 152) is controlled by the control signal C6, and the time adjustment driver is displayed. It is controlled by the control signal C7. Specifically, at least a part of the complexity adaptive video decoder 1 ,, such as the temporal prediction unit 132, the spatial prediction unit 134, the inverse transform unit 120, and the reconstructed frame output unit 15 And/or the fast VLD and inverse quantization module 11 〇 can operate according to the resolution of the plurality of reconstructed frames 158, instead of only relying on the resolution of the plurality of original frames represented by the input bit stream 1 〇 8 The operation is performed to reduce the complexity of the decoding operation. The reduced resolution frame can be obtained by down sampling, for example, by selecting a specific pixel to represent each block of the frame, for example, selecting each 2χ2 The pixel in the lower right corner of the prime block. In an embodiment, the time prediction unit 132 can operate according to the resolution of the plurality of reconstructed frames 158 instead of operating according to the resolution of the plurality of original messages to reduce the time. The complexity of the prediction. For example, the temporal prediction unit 132 may further estimate at least a portion of the information that may be used to perform temporal prediction in the omitted partial information. In addition, the spatial prediction unit may operate according to the resolution of the plurality of reconstructed frames 158. Instead of operating according to the resolution of multiple original frames, the complexity of the spatial prediction is reduced. For example, the spatial prediction unit 丨34 may further estimate at least a portion of the information of the 201204049 available = performing spatial prediction in the partially omitted information. According to this embodiment, the inverse transform unit 120 can operate according to the resolution of the plurality of reconstructed frames 1% without operating according to the resolution of the plurality of original frames to reduce the complexity of the inverse transform. For example, The inverse transform unit 12〇 may omit the inverse transform by selecting a plurality of preset inverse transform functions Fit for inverse transform In addition, the reconstructed frame output unit 150 can operate according to the resolution of the plurality of reconstructed frames 158 instead of the resolution of the plurality of original frames to reduce the number of generated signals. Reconstructing the complexity of the frame. Based on whether one or more conditions are met, the loop filter 152 may selectively perform the loop data for the deblocking corresponding to one of the plurality of complexity levels, wherein the deblocking The operations may be, for example, all deblocking, partial deblocking, and skipping deblocking. Further, the decoding and inverse quantization module 11 may operate according to the resolution of the plurality of reconstruction frames 158. Rather than relying on the resolution of multiple original frames to reduce the complexity of performing fast VLD and inverse quantization. More specifically, to reduce complexity, the fast variable length decoding and inverse quantization module may use a lookup table containing a primary table and at least one child table (eg, one or more child tables) during decoding, wherein In the design phase of the adaptive video decoder 100, the preset arrangement of the lookup table, the probability of using the primary table is greater than the probability of using at least one child table. In accordance with some variations of this embodiment, at least a portion of the components of the complexity adaptive video decoder may omit the portion of the component that is to be processed by the at least a portion of the component to reduce the complexity of decoding the input bitstream 108. For the sake of brevity, a similar description of such variations will not be repeated here. The 1C is a block diagram of an adaptive audio/video playback system 50 in accordance with the first embodiment. The adaptive audio/video playback system 5 includes an audio decanter 52 and a complexity adaptive video decoder 1 〇〇. Audio decoder 52 is passed to perform audio decoding to produce audio information 58. Additionally, the complexity adaptive video decoder 100 can perform the video decoding disclosed in the first embodiment. Note that the decoding complexity manager 100M is for performing decoding complexity management on the input bit stream 1 〇 8 when needed and the decoding complexity manager 100M selectively reduces the complexity within the adaptive video decoder 100 The decoding complexity of some of the multiple components. For example, the Decoding Complexity Manager 100M reduces the complexity of decoding the complexity of one or more components within the video decoder 100. As another example, the decoding complexity manager 100M delays audio playback of the audio information 58 without reducing the decoding complexity of any component within the complexity adaptive video decoder 100. Therefore, in the case where the video decoder cannot decode and output a plurality of frames matching the audio playback in time, the decoding complexity manager 1 can appropriately delay the audio playback of the adaptive audio/video playback system 50. To maintain audio/video playback synchronization between video playback and audio playback of the adaptive audio/video playback system 50. As shown in FIG. 1C, the decoding complexity manager ιοοΜ generates at least one control signal C0 to control whether to delay the audio playback of the audio information 58 and/or the amount of delay in the audio playback of the audio device §fl 58. In fact, the decoding complexity manager 1 oom can decide to delay the audio of the audio information 58 according to the bit stream information of the input bit stream 108, the parameter representing the processing capability, and the combination of the display buffer status of 12 a 201204049. Play. For example, based on either or both of the bitstream information of the input bitstream 108, a parameter representing the processing capability, and a display buffer state, the decoding complexity manager 100M determines whether to delay audio playback of the audio information 58. For another example, the decoding complexity manager 100M determines the amount of delay in audio playback of the audio information 58 based on either or both of the bitstream information of the input bitstream 108, the parameter representing the processing capability, and the display buffer state. For the sake of brevity, a similar description of such variations will not be repeated here. Fig. 1D is a block diagram showing a complexity adaptive video decoder 100-1 in accordance with a second embodiment of the present invention. This embodiment is a modification of the first embodiment. As shown in FIG. 1D, the aforementioned deblocking filter (for example, loop filter 152) is disposed outside the reconstructed frame output unit 150 of this embodiment, instead of being disposed in the reconstructed frame output unit 150. Inside. Please note that other components in the complexity adaptive video decoder 100-1 (eg, fast variable length decoding and inverse quantization module 110, inverse transform unit 120, motion compensation module 130, computing unit 140, frame storage unit 160) The implementation may be the same as the corresponding components in the complexity adaptive video decoder 100 shown in FIG. 1A. In this embodiment, the plurality of elements in Fig. 1B can also be coupled to the first DD. For the sake of brevity, a similar description of this embodiment will not be repeated here. In accordance with some variations of this embodiment, at least a portion of the components of the complexity adaptive video decoder 100-1 may omit the at least a portion of the portion of the processing that is to be processed to reduce the complexity of decoding the input bitstream 108. For the sake of brevity, a similar description of such variants is not repeated here. Figure 1E is a block diagram of an adaptive audio/video playback system 60 in accordance with a second embodiment. The adaptive audio/video playback system 60 includes an audio decoder 62 and a complexity adaptive video decoder 100-1. Audio decoder 62 is operative to perform audio decoding to produce audio information 68. Additionally, the complexity adaptive video decoder 100-1 can perform the video decoding disclosed in the second embodiment. Note that the 'Decoding Complexity Manager 100M is used to perform decoding complexity management on the input bitstream 108 when needed, and the Decoding Complexity Manager ι〇〇Μ Selectively Reduce Complexity Adaptive Video Decoder 100- The decoding complexity of some of the multiple components within 1. For example, the decoding complexity manager 1 〇〇M reduces the de-cracking complexity of one or more components within the complex adaptive video decoder 100-1. As another example, the decoding complexity manager 100M delays audio playback of the audio information 68 without reducing the decoding complexity of any component within the complexity adaptive video decoder 100_丨. Therefore, the decoding complexity manager 100M appropriately delays the audio_ of the adaptive audio/video playback system 60 to maintain the audio/video playback synchronization between the video playback and audio playback of the adaptive audio/video playback system 6G. Any video decoding delay issues. As shown in FIG. 1E, the solution processing horse complexity manager 10GM generates at least a (four) signal, such as the aforementioned at least _ control signal CG' to control whether to delay audio playback of the audio information 68 and/or audio playback of the audio information 68. The amount of delay. In practice, the decoding complexity management ^ 100M may depend on, for example, the bit stream information of the input bit 201204049 stream 108, the parameters representing the processing capabilities, and the display as described in the embodiment shown in FIG. 1C. A combination of ^ or one of them, decided to delay the sound of the audio information 68 ^ Xiao difficult to teach ^. For example, the complexity manager parameters are decoded and the audio playback of the audio information 68 is delayed based on either or both of the bit stream information of the input bit stream 108, the combination of the processing capability buffer states. For another example, the 〇〇M 决 的 的 的 贝 贝 贝 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 贝 贝 贝 贝 贝 贝 贝 解码 解码 解码 解码 解码 解码 解码 解码 解码 解码 解码 解码 解码 解码The amount of delay. For the sake of brevity, a similar description of this variant is not repeated here. By using the architecture of any of the embodiments/variations disclosed above, the video decoding complexity of each component can be rounded off. For example, within the complexity adaptive video decoding state 100 of the decoding complexity manager: (or one of its variants) or the second embodiment (or its variants do not need to lower the first implementation) Hard randomness adaptive video decoder 1 (either in its variable-noise adaptive video decoder or in the second embodiment (or one of its variants)) (complexity adaptation in the decoding complexity of any component within the frame) All components within the video decoder 丨(10) or complexity adaptive video decoder just-1 can operate according to the highest level of video decoding complexity. Refer to Figure 2 for more details. The following is described as follows: 201204049 2 The figure is a flowchart of an adaptive video decoding method 910 according to an embodiment of the present invention. The adaptive video decoding method is applicable to the complexity adaptive video decommissioning device or the first in the ia diagram and the first diagram. (1) The complexity adaptive video decoder 1G (M, more particularly, applicable to any of the above embodiments/deformation) is described in the figure. The description of the adaptive video gamma method 910 is as follows. In step S912, decoding The complexity manager 1 is based on the complexity adaptive video decoder in consideration (for example, the complexity adaptive library video decoder (10) of the first embodiment or the complexity adaptive video decoder ι〇〇ι of the second embodiment) The input bit stream 108 performs decoding complexity management to determine whether to reduce the decoding complexity of at least one of the plurality of components within the complexity adaptive video decoder. For example, the plurality of components can include fast (10) and The inverse quantization module 110, the inverse transform unit U0, the motion compensation module 13G, the reconstructed frame output unit 150, the aforementioned deblocking filter (for example, the loop filter 152), and the inter-drive driver 175. In step S914, The Decoding Complexity Manager 1M selectively reduces the decoding complexity of the 1 knives component of the plurality of components within the Complex Adaptive Video Decoder described in step 12. For example, the Decoding Complexity Manager 1 (8) is now The VLD and inverse quantization module 11G, the inverse transform unit, the 120 motion compensation module 13〇, the reconstructed frame output unit 150, and the aforementioned solution block waver can be selectively reduced at a specific time point. The loop filter 152) and the display complexity adjust the decoding complexity of at least a portion (eg, a portion or all) of the driver 201204049 175. For another example, the decoding complexity manager 100M does not reduce the fast VLD and the inverse quantization module at another point in time. 11解码, inverse transform unit 120, motion compensation module 13〇, reconstructed frame output unit 150, decoding complexity of any of the aforementioned deblocking filters (eg, loop filter 152) and display time adjustment driver 175 According to this embodiment, the decoding complexity manager 100M can refer to the display buffer state of the display buffer 180, the parameters of the one or more representative processing capabilities of the complexity adaptive video decoder described in step S912, and the input. The decoding complexity management is performed by a combination or one of the bit stream information of the bit stream 108, wherein the bit stream information may include a bit rate, a resolution, and/or a frame type. The bitstream information may include a plurality of motion vectors, a plurality of residuals, a plurality of IDCT coefficients, a bitstream length, or available bitstream information in some other instances, wherein the plurality of residuals may be, for example The bit stream information available for the Direct Current (DC) value may be obtained from information of the known bit stream described above based on, for example, comparison and/or historical statistics. In an embodiment, the decoding complexity manager 100M adjusts the decoding complexity based on changes in the plurality of motion vectors. The parameter representing the processing capability may include statistics of the decoding time required for each frame. For example, the decoding complexity manager 100M may use the average decoding time of each frame to determine the current system capability, thereby adaptively adjusting the decoding complexity. . In particular, the decoding complexity manager 100M can perform decoding complexity management based on the aforementioned display buffer state of the display buffer 18A, the aforementioned at least one parameter, and all of the bitstream information of the input bitstream 108. 17 201204049
第3圖係依據本發明一實施例的解碼複雜度管理器1〇〇M 可觸發的一些操作示意圖，其中，這些操作係關於如第2圖所 示的自適應視頻解碼方法910的解碼複雜度管理。例如，解碼 複雜度管理器100M可觸發的操作包含：顯示時間調整、丢棄 訊框、解區塊濾波器複雜度調整、運動補償複雜度調整（在第 3圖中標記為「MC複雜度調整」）、訊框内預測複雜度調整、 逆變換複雜度調整以及VLD複雜度調整。基於前述的顯示緩 衝器180的顯示緩衝器狀態、代表處理能力的參數以及輸入位 元流108的位元流資訊的組合或其中之一，解碼複雜度管理器 100M執行解碼複雜度管理，並選擇性地觸發這些操作中的一 個或多個。 第4圖係依據本發明一實施例的解碼複雜度管理器i〇〇m 可觸發的—些操作/步驟的相關品質和處理速度的示意圖，其 中，逆些操作係關於如第2圖所示的自適應視頻解碼方法 的解碼複雜度管理。例如，這些操作/步驟可包含：標準解碼 流程“、自適應雙向訊框(Bi-directional frame，B frame)(以下簡 稱為“B訊框”）解區塊（即自適應執行全部或部分b訊框解區 塊或跳過B訊框解區塊，其巾區塊指對多個b訊框 執行解區塊操作）、B訊框簡化運動補償（在第4 ®中標記為 B代忙簡化MC」）、快速B訊框解碼（例如在降低取樣域中 解碼B況枢）、丢棄B訊框、自適應預測訊框听―frame. 201204049 P frame)(以下簡稱為“p訊框，，）解區塊（例如，自適應執行 全部或部分P訊框解區塊或跳過p訊框解區塊，其中p訊框解 區塊指對多個P訊框執行解區塊操作）、p訊框簡化運動補償 (在第4圖中標記為「P訊框簡化Mc」）、快速p訊框解碼、 丟棄^訊框〜自適應訊框内㈣^咖以誕^以 ==解區塊(例如，自適應執行全部或部分1訊框解 £塊或跳過I訊框解區塊，其中 執行解區塊操作）以及快速^忙链框解區塊指對多個1訊框 .用，本發明並非以此為限垂所舉之例僅為說明之 '像口曾、冷心巧限柔直軸上的「品質」標記可表示影 指:的綜:評工二系統資源㈣^ 、依嫘此實細*例的一此轡开4j日庙私.士士 /步驟的品質和處理速戶在…—反"應於每些操作 第4圖所示的操作/步驟中在不可有所不 標準解石馬流程 f Β _的操作/步驟可插入在 之間。又例如，在第/圖ΓΛ 訊框解區塊的操作/步驟 區塊的操作/步驟訊框解 快速P訊框解碼3化運動補償的操作/步驟和 操作/步驟中狀間。又例如，在第4圖所示的 的操作/步驟，因此 =#作/步驟可鄰近快速1訊框解碼 -個操作/步驟。為第4圖所示的操作/步驟中最底層的 在此貫施例令，其笛— 201204049 可觸發標準解碼流程的操作/步驟。然而，當確定會發生視頻 解碼的延遲時（例如發現硬體資源不足和/或解碼工作量大於 預設的閾值），解碼複雜度管理器100M可觸發一個或多個操 作/步驟，以防止嚴重的視頻解碼延遲，並實現當前情況下解 碼器可提供的最佳品質。例如，在第4圖所示的操作/步驟中， 解碼複雜度管理器100M可觸發下一適當的操作/步驟（例如自 適應B訊框解區塊的操作/步驟，B訊框簡化運動補償的操作/ 步驟等），以釋放硬體資源和/或減少解碼工作量，其中，當降 低解碼複雜度時，相應於較高品質級別的操作/步驟擁有較高 的執行優先權，而相應於較低品質級別的操作/步驟會在最後 執行。所舉之例僅為說明之用，本發明並非以此為限。依據此 實施例的變形，解碼複雜度管理器100M可基於第二策略-實現 預設的品質級別進行操作，而不是基於實現最佳品質進行操 作。依據此實施例的一些其他變形，解碼複雜度管理器100M 可基於實現最快處理速度或預設的處理速度的一些策略進行 操作。 第5圖係依據本發明另一實施例的解碼複雜度管理器 100M可觸發的一些操作/步驟的相關品質和處理速度的示意 圖，其中，這些操作係關於如第2圖所示的自適應視頻解碼方 法910的解碼複雜度管理。特別地，此實施例中考量的操作/ 步驟包含：正常解區塊、簡單解區塊以及禁用解區塊。請注意， 解碼複雜度管理器100M的控制機制可依據多個預設的閾值 201204049 (例如，第5圖所示的兩個閾值Τ1和Τ2 )在這些操作/步驟之 間進行切換，其中’預設的閾值T1和T2代表品質的相應級別。 第6圖係依據本發明另一實施例的解碼複雜度管理器 100M在不同情況下可觸發的一些操作/步驟的示意圖，這些操 作係關於如第2圖所示的自適應視頻解碼方法91〇的解碼複雜 度管理。特別地，此實施例中考量的操作/步驟包含：停止这 訊框解區塊、停止P訊框解區塊、啟動自適應p訊框解區塊、 啟動自適應I訊框解區塊以及停止解區塊。請注意，解碼複雜 度管理器100M的控制機制可依據多個預設的閾值（例如，第 6圖所示的三個閾值T!、TP和TB)在這些操作/步驟之間進行 切換，其中，預設的閾值1、TP和TB代表顯示緩衝器狀態的 相應級別。在此實施例中，預設的閾值τ : 決定是否啟動/停止B訊框解區塊、P訊框解‘二== 區塊。特別地，在顯示緩衝器18()為滿或幾乎為滿的情 ::了二1器_不停止1訊框解區塊、p訊框解區塊 為空錢塊的任何一個。相反地，在顯示緩衝器180 f A 的情訂，解碼複雜度管理111醒停止1訊 1 A p訊框解區塊以及心框解區塊的每一個。 顯干ί衝在顯示緩衝器18G的緩衝11狀態（更特別地， ==的緩衝級別）在預設的間值了糾之間的情 .....理1^麵停止B訊框解區塊，並觸發啟 201204049 動自適應P訊框解區塊的操作/步驟（即具有複雜度自適應的p 訊框解區塊）。另外，在顯示緩衝器⑽的緩衝器狀q更特 別地，顯示緩衝器⑽的緩衝級別）在預設的閾值T。之 間的情況下，解碼複雜度管理器⑽M停止p訊框解區塊矛:B 訊框解區塊，並觸發啟動自適應1訊框解區塊的_/轉（即 具有複雜度自適應的：[訊框解區塊）。另外，在顯示緩衝器⑽ 的緩衝器狀態（更特別地’顯示緩衝器18〇的緩衝級別）在預 設的閾值T>G之間的情況下，解碼複雜度管㈣丨麵停止 I afl框解區塊、p訊框解區塊和B訊框解區塊。 在此實施例中，關於複雜度自適應解區塊的控制機制，解 碼複雜度管理器HK)M可基於巨集區塊⑽⑽吨_为基 礎的解區塊方案进行操作。例如，解碼複雜度管理器i〇〇m可 依據緩衝器狀態和輸入位元流1〇8的至少一部分（例如一部分 或全部）的位元流長度，自適應啟動/停止〗訊框解區塊、p訊 框解區塊和/或B訊框解區塊。所舉之例僅為說明之用，本發 明並非以此為限。依據此實施例的變形，關於複雜度自適應解 區塊的控制機制，解碼複雜度管理器100M可基於訊框为基礎 的解區塊方案进行操作。例如，解碼複雜度管理器100M可依 據緩衝器狀態 '輸入位元流108的至少一部分（例如一部分或 全部）的位元流長度、量化步长、MB類型和/或邊緣強度來自 適應地啟動/停止I訊框解區塊、P訊框解區塊和/或B訊框解3 is a schematic diagram of some operations that the decoding complexity manager 1 〇〇M can trigger according to an embodiment of the present invention, wherein the operations are related to the decoding complexity of the adaptive video decoding method 910 as shown in FIG. management. For example, the operations that the decoding complexity manager 100M can trigger include: display time adjustment, discard frame, deblocking block complexity adjustment, and motion compensation complexity adjustment (labeled as "MC complexity adjustment" in FIG. "), intra-frame prediction complexity adjustment, inverse transformation complexity adjustment and VLD complexity adjustment. Based on the aforementioned display buffer status of the display buffer 180, a parameter representing the processing capability, and a combination of bit stream information of the input bit stream 108, the decoding complexity manager 100M performs decoding complexity management and selects One or more of these operations are triggered sexually. 4 is a schematic diagram showing the correlation quality and processing speed of some operations/steps that can be triggered by the decoding complexity manager i〇〇m according to an embodiment of the present invention, wherein the operations are reversed as shown in FIG. Decoding complexity management of adaptive video decoding methods. For example, the operations/steps may include: a standard decoding process, a Bi-directional frame (B frame) (hereinafter referred to as a "B frame") solution block (ie, adaptive execution of all or part of b) Frame deblocking or skipping the B frame to solve the block, the towel block refers to performing deblocking operation on multiple b frames, and B frame simplifies motion compensation (marked as B generation busy in 4th ® Simplify MC"), fast B frame decoding (for example, decoding B status in the reduced sampling domain), discarding B frames, adaptive prediction frame listening - frame. 201204049 P frame) (hereinafter referred to as "p frame" , ,) deblocking (for example, adaptively executing all or part of the P-frame deblocking or skipping the p-frame deblocking, where the p-frame deblocking refers to performing deblocking operations on multiple P-frames ), p frame to simplify motion compensation (marked as "P frame simplification Mc" in Figure 4), fast p frame decoding, discarding frame ~ adaptive frame (4) ^ coffee to birth ^ to == Deblocking blocks (for example, adaptively performing all or part of the 1 frame cancellation block or skipping the I frame deblocking, where the deblocking operation is performed) and fast ^Busy chain frame solution block refers to multiple 1 frame. For the purpose of this invention, it is not only limited to the description of the "quality" on the straight axis. The mark can represent the shadow: the comprehensive: the evaluation of the second system resources (four) ^, relying on this actual fine * example of the opening of the 4j day temple private. The quality of the gentleman / step and processing fast-speed in the ... - anti " In the operation/step shown in Figure 4 of each operation, the operation/step of the non-standard stone-removing process f Β _ can be inserted between. For another example, in the operation/step block operation/step block of the block/decode block, the fast P frame decoding 3 motion compensation operation/step and operation/step mode. For another example, the operation/step shown in Fig. 4, therefore, the =# doing/step can be decoded adjacent to the fast 1 frame-operation/step. For the lowest level of the operations/steps shown in Figure 4, the flute - 201204049 can trigger the operation/step of the standard decoding process. However, when it is determined that a delay in video decoding will occur (eg, insufficient hardware resources are found and/or the decoding effort is greater than a predetermined threshold), the decoding complexity manager 100M may trigger one or more operations/steps to prevent serious The video decode delays and achieves the best quality that the decoder can provide in the current situation. For example, in the operation/step shown in FIG. 4, the decoding complexity manager 100M can trigger the next appropriate operation/step (for example, the operation/step of the adaptive B frame deblocking, the B frame simplifies the motion compensation). Operation/steps, etc.) to free up hardware resources and/or reduce decoding effort, wherein when decoding complexity is reduced, operations/steps corresponding to higher quality levels have higher execution priority, corresponding to Lower quality level operations/steps will be performed at the end. The examples are for illustrative purposes only and the invention is not limited thereto. According to a variation of this embodiment, the decoding complexity manager 100M can operate based on the second policy-implementing a predetermined quality level, rather than operating based on achieving the best quality. According to some other variations of this embodiment, the decoding complexity manager 100M can operate based on some strategies that achieve the fastest processing speed or preset processing speed. 5 is a schematic diagram showing the correlation quality and processing speed of some operations/steps that can be triggered by the decoding complexity manager 100M according to another embodiment of the present invention, wherein the operations are related to the adaptive video as shown in FIG. Decoding complexity management of decoding method 910. In particular, the operations/steps considered in this embodiment include: normal deblocking, simple deblocking, and disabling of the deblocking. Please note that the control mechanism of the decoding complexity manager 100M can switch between these operations/steps according to a plurality of preset thresholds 201204049 (for example, two thresholds Τ1 and Τ2 shown in FIG. 5), where The thresholds T1 and T2 are set to represent the corresponding levels of quality. Figure 6 is a diagram showing some of the operations/steps that the decoding complexity manager 100M can trigger in different situations in accordance with another embodiment of the present invention, relating to the adaptive video decoding method 91 as shown in Figure 2〇 Decoding complexity management. In particular, the operations/steps considered in this embodiment include: stopping the frame deblocking, stopping the P-frame deblocking, starting the adaptive p-frame deblocking, starting the adaptive I-frame deblocking, and Stop solving the block. Please note that the control mechanism of the decoding complexity manager 100M can switch between these operations/steps according to a plurality of preset thresholds (for example, three thresholds T!, TP and TB shown in FIG. 6), wherein The preset thresholds 1, TP and TB represent the respective levels of the display buffer status. In this embodiment, the preset threshold τ: determines whether to start/stop the B-frame solution block, and the P-frame solution ‘2== block. In particular, in the case where the display buffer 18() is full or almost full, the second blocker _ does not stop the 1-frame demapping block, and the p-frame demapping block is any one of the empty money blocks. Conversely, in the display buffer 180 f A, the decoding complexity management 111 wakes up and stops each of the 1 frame and the heart frame solution block. Explicitly rushing in the buffer 11 state of the display buffer 18G (more specifically, the buffer level of ==) is between the preset inter-valued corrections. Block, and trigger the operation/step of the 201204049 dynamic adaptive P-frame to solve the block (that is, the p-frame solution block with complexity adaptation). Further, in the buffer shape q of the display buffer (10), more specifically, the buffer level of the display buffer (10) is at a preset threshold T. In the case of the case, the decoding complexity manager (10) M stops the p-frame to solve the block spear: the B frame deblocks, and triggers the start of the adaptive 1 frame solution block _ / turn (ie has complexity adaptation : [frame decoding block). In addition, in the case where the buffer state of the display buffer (10) (more specifically, the buffer level of the display buffer 18A) is between the preset threshold T > G, the decoding complexity tube (4) is stopped. Solution block, p-frame solution block and B frame solution block. In this embodiment, with respect to the control mechanism of the complexity adaptive deblocking, the decoding complexity manager HK)M can operate based on the macroblock (10) (10) ton-based solution block scheme. For example, the decoding complexity manager i〇〇m may adaptively start/stop the frame deblocking according to the buffer status and the bit stream length of at least a portion (eg, part or all) of the input bit stream 1〇8. , p frame decryption block and / or B frame solution block. The examples are for illustrative purposes only and the invention is not limited thereto. According to a variation of this embodiment, with respect to the control mechanism of the complexity adaptive deblocking, the decoding complexity manager 100M can operate based on the frame-based deblocking scheme. For example, the decoding complexity manager 100M may adaptively enable/initiate/in accordance with the bitstream length, quantization step size, MB type, and/or edge strength of the buffer state 'input portion (eg, part or all) of the bit stream 108. Stop I frame decryption block, P frame solution block and / or B frame solution
22 201204049 區塊’其中，量化步長即所謂的量化參數(Quantizati〇n Parameter，QP)值。 本發明的優勢在於，上述揭示的每個實施例/變形的自適 應視頻解碼方法、相關的複雜度自適應視頻解碼器及相應的自 適應音頻/視頻播放系統可適用於各種數位视頻應用，其中， 複雜度自適應視頻解碼器内多個組件中的至少一部分（例如一 部分或全部）可以低級別的視頻解碼複雜度進行操作。由於可 調整一個或多個組件的視頻解碼複雜度，解碼工作量和相應的 功率消耗可自適應地減少。因此，與相關的技藝相比，本發明 為實現數位視頻播放系統提供了的極大靈活性。 上述之實施例僅用於例舉本發明之實施態樣，以及闡釋本 發明之技術特徵，並非用於限制本發明之範驚。任何習知技藝 者可依據本發明之精神輕易完成之改變或均等性之安排均屬 於本發明所主張之範圍，本發明之權利範圍應以申請專利範圍 為準。 « 【圖式簡單說明】 第1A圖係依據本發明第一實施例的複雜度自適應視頻解 碼器的方塊示意圖。 第1B圖係依據本發明第—實施例的複雜度自適應視頻解 碼器的方塊示意圖。 23 201204049 第ic圖係依據第一實施例的自適應音頻/視頻播放系統 的方塊示意圖。 第1D圖係依據本發明第二實施例的複雜度自適應視頻解 碼器的方塊示意圖。 圖係依據第一實施例的自適應音頻/視頻播放系統 的方塊示意圖。 '’ 第2圖係依據本發明的一實施例的自適應視頻 的流程圖。 电 第3圖係依據本發明一實施例的解碼複雜度管理 發的一些操作示意圖。 ^ 了觸 第4圖係依據本發明一實施例的解碼複雜度管理器可 發的-些操作/步驟的相關品質和處理速度的示意圖。 第5圖係依據本發明另一實施例的解碼複雜度管理器 觸發的-些操作/步驟的相關品質和處理速度的示意圖。 第6圖係依據本發明另—實施例的解碼複雜度管㈣在不同 下可觸發的-錄作/倾的示意gj。 4 【主要元件符號說明】 100〜複雜度自適應視頻解碼器；1〇8〜輸入位元流； 110〜快速VLD與逆量化模組；118〜逆量化結果； 120〜逆變換單元；128〜逆變換結果； 130〜MC模組；132〜時間預測單元； 134〜空間預測單元；138〜預測輸出結果；14〇〜計算 24 201204049 單元； 148〜補償輸出結果；150〜重構訊框輸出單元； 152〜迴路濾波器；158〜重構訊框； 160〜訊框儲存單元；170〜縮放器； 175〜顯示時間調整驅動器；178〜影像訊框； 180〜顯示緩衝器；188〜輸出訊框； 100M〜解碼複雜度管理器； 50〜自適應音頻/視頻播放系統；52〜音頻解碼器； 58〜音頻資訊；100-1〜複雜度自適應視頻解碼器； 62〜音頻解碼器；68〜音頻資訊； 910〜自適應視頻解碼方法； C1-C7〜控制信號； S912- S914〜步驟。 2522 201204049 Block ' where the quantization step size is the so-called quantization parameter (QP) value. An advantage of the present invention is that each of the embodiments/variable adaptive video decoding methods, associated complexity adaptive video decoders, and corresponding adaptive audio/video playback systems disclosed above are applicable to a variety of digital video applications, wherein At least a portion (eg, a portion or all) of the plurality of components within the complexity adaptive video decoder may operate at a low level of video decoding complexity. Since the video decoding complexity of one or more components can be adjusted, the decoding effort and corresponding power consumption can be adaptively reduced. Thus, the present invention provides great flexibility in implementing a digital video playback system as compared to related art. The embodiments described above are only intended to illustrate the embodiments of the present invention, and to explain the technical features of the present invention, and are not intended to limit the scope of the present invention. Any change or singularity of the present invention in light of the spirit of the present invention is intended to be within the scope of the present invention. The scope of the invention should be determined by the scope of the claims. «FIG. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1A is a block diagram showing a complexity adaptive video decoder in accordance with a first embodiment of the present invention. Fig. 1B is a block diagram showing a complexity adaptive video decoder in accordance with a first embodiment of the present invention. 23 201204049 The ic diagram is a block diagram of an adaptive audio/video playback system in accordance with the first embodiment. Fig. 1D is a block diagram showing a complexity adaptive video decoder in accordance with a second embodiment of the present invention. The figure is a block diagram of an adaptive audio/video playback system in accordance with the first embodiment. '' Fig. 2 is a flow chart of adaptive video in accordance with an embodiment of the present invention. Figure 3 is a schematic diagram of some operations of decoding complexity management in accordance with an embodiment of the present invention. The touch diagram 4 is a schematic diagram of the correlation quality and processing speed of the operations/steps that can be performed by the decoding complexity manager in accordance with an embodiment of the present invention. Figure 5 is a diagram showing the correlation quality and processing speed of the operations/steps triggered by the decoding complexity manager in accordance with another embodiment of the present invention. Fig. 6 is a schematic diagram showing the recording/tilting of the decoding complexity tube (4) in accordance with another embodiment of the present invention. 4 [Main component symbol description] 100~complexity adaptive video decoder; 1〇8~ input bit stream; 110~ fast VLD and inverse quantization module; 118~ inverse quantization result; 120~ inverse transform unit; 128~ Inverse transform result; 130~MC module; 132~time prediction unit; 134~space prediction unit; 138~predictive output result; 14〇~calculation 24 201204049 unit; 148~compensated output result; 150~reconstructed frame output unit 152~loop filter; 158~reconstruction frame; 160~frame storage unit; 170~scaler; 175~ display time adjustment driver; 178~image frame; 180~ display buffer; 188~output frame ; 100M ~ decoding complexity manager; 50 ~ adaptive audio / video playback system; 52 ~ audio decoder; 58 ~ audio information; 100-1 ~ complexity adaptive video decoder; 62 ~ audio decoder; 68 ~ Audio information; 910~Adaptive video decoding method; C1-C7~ control signal; S912-S914~ steps. 25