TW200421880A - Method and apparatus for rate control in moving picture video compression - Google Patents
Method and apparatus for rate control in moving picture video compression Download PDFInfo
- Publication number
- TW200421880A TW200421880A TW92107989A TW92107989A TW200421880A TW 200421880 A TW200421880 A TW 200421880A TW 92107989 A TW92107989 A TW 92107989A TW 92107989 A TW92107989 A TW 92107989A TW 200421880 A TW200421880 A TW 200421880A
- Authority
- TW
- Taiwan
- Prior art keywords
- day
- complexity
- current
- group
- picture
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 15
- 230000006835 compression Effects 0.000 title abstract description 11
- 238000007906 compression Methods 0.000 title abstract description 11
- 230000000694 effects Effects 0.000 claims abstract description 46
- 238000013139 quantization Methods 0.000 claims abstract description 17
- 230000033001 locomotion Effects 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000004519 manufacturing process Methods 0.000 claims 1
- 239000012528 membrane Substances 0.000 claims 1
- 230000001052 transient effect Effects 0.000 claims 1
- 238000005070 sampling Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010012335 Dependence Diseases 0.000 description 1
- 101000973623 Homo sapiens Neuronal growth regulator 1 Proteins 0.000 description 1
- 102100037812 Medium-wave-sensitive opsin 1 Human genes 0.000 description 1
- 235000000434 Melocanna baccifera Nutrition 0.000 description 1
- 241001497770 Melocanna baccifera Species 0.000 description 1
- 102100022223 Neuronal growth regulator 1 Human genes 0.000 description 1
- 241000237503 Pectinidae Species 0.000 description 1
- 241000218998 Salicaceae Species 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000013256 coordination polymer Substances 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 235000020637 scallop Nutrition 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 229910052720 vanadium Inorganic materials 0.000 description 1
- LEONUFNNVUYDNQ-UHFFFAOYSA-N vanadium atom Chemical group [V] LEONUFNNVUYDNQ-UHFFFAOYSA-N 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
200421880200421880
【卷明所屬之技術領域】 本發明係有關於:备柯% 别么貝枓壓縮之領域,特別係指一 於視訊編碼系統執杆私_ * J n^ 種適用 钒仃位兀率控制之方法及裝置。 【先前技術】 位声眾子通訊世界正經歷一次數位革命,以數 位表不貝5孔的主要優點在於能夠幾近無誤地儲#、再生、 傳收處理及運用的貧料位元流(b i t s t r eam )。舉例而古 ,NTSC的彩色視訊影像每秒有2 9 9 7張晝面,每張晝面二 4 8 0條可見掃瞄線,每條掃瞄線約需4 8 〇點以紅、綠、藍三 色呈現的像素(pixel),但若每種色彩成分以8位元編碼, 則所產生的位元率(bitrate)每秒約168百萬位元(Mbits/s ),故各種視訊格式其未經壓縮處理的位元率是非常高而 不經濟,因此不適於多數的應用。 與電腦、電信網路、消費性產品整合的數位音訊和視 訊,更加刺激資訊革命的前進,而這革命的核心則是視、 音訊的數位壓縮技術。許多的壓縮標準,包含以壓縮技術 共通之精髓為基礎的演算法,如:ITU-T(前身係CCITT)建 議書Η.261 和Η.2 6 3,以及IS0/IEC 的MPEG-1、MPEG-2 和 MPEG-4標準。MPEG的演算法是由動態影像專家群組 (Moving Picture Experts Group,MPEG)所發展出來,該 動態影像專家群組係國際標準組織(I n t e r n a t i ο n a 1[Technical field to which Mingming belongs] The present invention relates to: the field of compression and compression, and particularly refers to a type of video coding system that is suitable for controlling the vanadium position. Method and device. [Prior technology] The world of BitSonic communication is undergoing a number of digital revolutions. The main advantage of using a digital table to represent 5 holes is that it can store #, regenerate, transmit, and process lean bitstreams (bitstr) without error. eam). For example, NTSC's color video image has 2 997 daylight surfaces per second, and each daytime surface has 480 visible scanning lines. Each scanning line requires approximately 480 points in red, green, and blue. Pixels rendered in three blue colors, but if each color component is encoded in 8 bits, the resulting bitrate is about 168 million bits per second (Mbits / s), so various video formats Its uncompressed bit rate is very high and uneconomical, so it is not suitable for most applications. Digital audio and video integrated with computers, telecommunications networks, and consumer products have further stimulated the information revolution. The core of this revolution is digital compression technology for video and audio. Many compression standards, including algorithms based on the essence of common compression technology, such as: ITU-T (formerly CCITT) Recommendations Η.261 and Η.2 6 3, and ISO / IEC MPEG-1, MPEG -2 and MPEG-4 standards. The MPEG algorithm was developed by the Moving Picture Experts Group (MPEG), which is an international standard organization (I n t e r n a t i ο n a 1
Standards Organization,ISO)及國際電子技術委員會 (International Electrotechnical Commission , IEC)的Standards Organization (ISO) and International Electrotechnical Commission (IEC)
0702-7302TWf(nl);90P97;Lin.ptd 第6頁 200421880 五、發明說明(2) 聯合技術委員會,致力於發展視、音訊的壓縮及多路傳殊 的表現方式,這些標準規定了壓縮位元流的語法(syntax) 以及解碼的方法,但對於編碼器所使用的演算法而言,卻 為編碼的新穎性與多樣性保留了相當多的自由度。 根據MPEG,一連串的視訊晝面(picture)分成一序列 的晝面群組(group ofpicture,G0P),其中每組G0P以I-晝面開始,後面跟著P-晝面和B-晝面的安排,第1圖以顯 示的順序說明一組典型的G0P。I -晝面之編碼毋須參考先 前或將來的晝面,P-晝面則參考連串的視訊晝面中在時間 上最接近的I -晝面或p -晝面來進行編碼,而B -畫面散佈於 I -晝面和P -晝面之間。並且利用先前的、將來的或兩者兼 具的緊鄰I -晝面和p_畫面來編碼。雖然好幾張B-畫面可以 緊接著連續出現,但絕不能以B-晝面預估其他的晝面。 每張晝面具有三種成分:亮度值(luminance,以Y表 示)’紅色差值(red color difference,以Cr表示),以 及監色差值(blue color difference,以Cb表示)。對於 MPEG的4 : 2 : 0的格式而言,每一種cr和Cb成分的取樣點在 水平與垂直方向只有γ成分的一半。如第2圖所示,一張 MPEG的晝面其基本構成方塊係大區塊,以MB 表示)。以4 : 2 : 〇的視訊為例,每個MB包含一個Y成分16x1 6 取樣點的陣列,以及兩個Cr和Cb成分8x8取樣點的區塊, 其中Y成分16x16取樣點的陣列實際上由4個8x8取樣點的區 塊所組成。 編碼器的作用在於決定何種畫面編碼型態以及何種預0702-7302TWf (nl); 90P97; Lin.ptd Page 6 200421880 V. Description of the invention (2) Joint technical committee is committed to the development of video and audio compression and multi-channel transmission. These standards specify the compression bit The syntax and decoding method of the metastream, but for the algorithm used by the encoder, it retains a considerable amount of freedom for the novelty and diversity of the encoding. According to MPEG, a series of video diurnal pictures are divided into a series of diurnal groups (G0P), where each G0P starts with I-dimension, followed by P-dimension and B-dimension Figure 1 illustrates a typical set of GOPs in the order shown. The encoding of the I-day plane does not need to refer to the previous or future day plane, the P-day plane refers to the I-day plane or p-day plane closest to the time in the series of video day planes, and B- The picture is interspersed between I-day and P-day. And use the previous, future, or both of the immediate I-day plane and p_picture to encode. Although several B-pictures can appear next to each other in succession, the other day-surfaces must not be estimated from the B-day-surface. Each daylight surface has three components: a brightness value (represented by Y), a red color difference (represented by Cr), and a monitor color difference (represented by Cb). For the MPEG 4: 2: 0 format, the sampling points for each cr and Cb component are only half of the γ component in the horizontal and vertical directions. As shown in Figure 2, the basic composition of a MPEG diurnal surface is a large block, expressed in MB). Taking the video of 4: 2: 〇 as an example, each MB contains an array of 16x16 sampling points of the Y component and two blocks of 8x8 sampling points of the Cr and Cb components, where the array of 16x16 sampling points of the Y component is actually composed of It consists of 4 blocks of 8x8 sampling points. The role of the encoder is to determine what type of picture encoding and what kind of prediction
07〇2-7302TWf(nl);90P97;Lin.ptd 第7頁 200421880 五、發明說明(3) 測模式為最佳。對每張I -畫面,MB中的每個8x8方塊均經 過離散餘弦轉換(discrete cosine transform,DCT)而形 成8 x 8轉換係數陣列,轉換係數接著以一量化矩陣進行量 化,然後用Z字形(Zig-zag)掃瞄DCT係數的量化結果而得 到一連串的DCT係數,並且此DCT係數序列以可變長度碼 (variable length c〇de,VLC)進行編碼。p—晝面必^決 ^將每個0以1型〇或?型〇進行編碼,丨型腳的編碼係以 方式為之,而對於每個P_,則需得到該 刖畫面所做之預測,此預測係藉由—種移 及其在先刚晝面中的預估Μβ之間的 ^ 間的預測誤差則以DCT、量化、z字;;,預估Μβ與目《Β 碼。 化Ζ子形知瞄以及VLC進行編 處理Β-晝面時,必須決令 讎:!模式、F模式、Β模式^何麵模式來編碼每 身而不借助移動補償來編碼(如! I式。I模式係以MB本 的向前預測編碼,係以先前書 β —般);F模式為單向 型MB —般);反之,Β模式為以向移動補償的預估(如Ρ 來的晝面得到移動補償的^估j特向後預測編碼,係以後 向的預測編碼,運用向前的 $別的是FB模式,其係雙 做内插(interpolation)而得° < 移動補償預估兩者來 對F、B和FB模式而言,預剛亨』叛式之移動補償預估。 形掃瞄以及VLC進行編碼。、 以使用DCT、量化、z字 編碼器需要為整張晝 、取里化位階以便在給定的位07〇2-7302TWf (nl); 90P97; Lin.ptd Page 7 200421880 V. Description of the invention (3) The test mode is the best. For each I-picture, each 8x8 block in MB is subjected to a discrete cosine transform (DCT) to form an array of 8 x 8 conversion coefficients. The conversion coefficients are then quantized by a quantization matrix and then zigzag ( Zig-zag) scans the quantized results of the DCT coefficients to obtain a series of DCT coefficients, and the DCT coefficient sequence is encoded with a variable length code (VLC). p—day and night must be determined ^ Will each 0 be a type 0 or? Type 0 is coded. The coding of type 丨 is done in this way. For each P_, we need to get the prediction made by the picture. This prediction is based on- The prediction errors between the estimated Mβ are DCT, quantization, and z;;, the estimated Mβ and the head "B code. When transforming the Z-shaped image and VLC to process the B-day, you must make a decision 雠:! Mode, F mode, B mode, and any mode to encode each body without the use of motion compensation (such as! Type I. I mode uses MB-based forward prediction coding, which is based on the previous book β-like); F The mode is one-way type MB in general. On the other hand, the B mode is based on the prediction of backward motion compensation (such as the day-to-day motion compensation from P to get the ^ estimated j special backward prediction encoding, which is the backward prediction encoding. The former $ other is the FB mode, which is obtained by double interpolation. ≪ Motion compensation estimates are both for the F, B and FB modes. Estimation. Shape scan and VLC encoding. To use DCT, quantization, z-code encoder needs to be the whole day and time to get the order in the given bit.
0702-7302TWf(nl);90P97;Lin.ptd 200421880 五、發明說明(4) 元率下控制 量化位階編 中並不存在 階下,決定 器很重要的 的係彳艮有智 元數,使編 k南編碼視 卻相當複雜 理。 有鑑於 單次、即時 像視訊壓縮 及裝置。 mr直到實際編瑪後,以選擇的 碼/張旦面所用的實際位元數才會得知, 一個逆轉函數,能夠在給定希望達到的量化ς 〆張畫面所用的實際位元數, - ir 口此’視訊1編碼 一點即為位兀率的控制。位元率控制之主 慧地分配編碼每張晝面及其巾每個龍所用的位 碼過的=訊於解碼器進行解碼時,能儘可能的 訊的祝見品質。不過,傳統的位元率控制方法 ,通常需進行好幾次才能完成視訊編碼的處 此亟需一種新穎的位元率控制技術,可用於 的視訊編碼器,再者,亦期待一種適用動態影 的圖幀層級(frame-level)之位元率控制方法 【發明内容】 根據本發明,备,i而甘Ρϊ / 先前晝面的編碼心祐。位元的分S己’係基於 析的活動,,夂Π ϋ碼的目前畫面其預先分 及先前編碼晝面ί實預先分析的活動量以 種預估的目前編碼後的品質資訊,這 以f Μ地八: 編碼器是非常有用的,可 蚩面^平i的目標位元額度給每張晝面,而能達到不同 旦面間更平均的視覺品質。0702-7302TWf (nl); 90P97; Lin.ptd 200421880 V. Description of the invention (4) There is no order in the control of the quantization rank order at the element rate. The system where the determinant is important is the intelligent element number, which makes the k South coding is quite complicated. Considering single-shot, live video compression and installation. mr will not know the actual number of bits used by the selected code / Zhang Dan face until after the actual editing. A reversal function can give the desired number of quantizations. The actual number of bits used in each picture,- The ir port of this 'video 1' is a bit rate control. The master of the bit rate control wisely allocates the bit code used to encode each day face and each dragon. When the decoder decodes it, it can be as good as possible. However, traditional bit rate control methods usually need to be performed several times to complete the video encoding. A novel bit rate control technology is urgently needed, which can be used in video encoders. Furthermore, a dynamic video encoding method is also expected. Bit-rate control method for picture frame-level [Summary of the Invention] According to the present invention, the preparation of the first and the last day of the day code encoding. The bit score S 'is based on the analysis of the activity. The current picture of the 夂 Π ϋ code is pre-divided and the previous code is analyzed in advance. The amount of pre-analyzed activity is an estimate of the quality information of the current code. f Μ 地 八: The encoder is very useful, it can give the target bit quota of plane i to each day plane, and can achieve more average visual quality between different planes.
0702-7302TWf(nl);90P97;Lin.ptd0702-7302TWf (nl); 90P97; Lin.ptd
Bill m 第9頁 係針對一種 將被編碼的 係在一視訊 之全活動量 活動量對複 前晝面預估 前畫面之目 間複雜度而 如晝面進行 全活動量、 量化位階, 而言,目前 動量成正比 化位階成反 面,本發明 包括一移動 及一參數更 續晝面中即 區塊匹配時 之全活動量 活動量對複 畫面之複雜 新本組連續 瞬間複雜度 訊位元率控 計算其全活 連續晝面之 續晝面中同 估計目前晝 新本組連續 的分配,係 目標位元額 别晝面編碼 實際消耗位 面其活動量 量對複雜度 晝面之實際 動態影像視 一複雜度估 估算單元接 目前晝面, 前畫面之全 續晝面中同 複雜度估計 晝面預估的 複雜度,且 位元額度給 五、發明說明(5) 本發明 首先為一即 該目前晝面 於目前晝面 編碼晝面其 ,並且以目 複雜度;目 雜度以及瞬 序列中的目 目前晝面之 晝面之平均 比’就這點 晝面之全活 以及平均量 另一方 制之裝置, 額分派器以 裡的一組連 移動估算的 於目前晝面 編碼晝面其 來估計目前 頭分派器更 複雜度以及 動態影像視 目前畫面來 序列的一組 以及這組連 雜度之比, 的複雜度更 標位元額度 定;故基於 編碼。在目 目前晝面之 計算目前晝 晝面的活動 ’而與目前 比。 還揭露一 ^ 估算單元、 新器。移動 將被編碼的 ,計算此目 以及這組連 雜度之比, 度。以目前 畫面之瞬間 ,分配目標 制之方法, 動量,其中 中;接著基 型態的先前 面之複雜度 畫面之瞬間 依預估的複 度,對視訊 之後,根據 元數和目前 對複雜度之 之比與目前 消耗位元數 訊位元率控 計器、一配 收視訊序列 用來在執行 活動量。基 型態之先前 器可因此用 複雜度,配 根據預估的 目前畫面。 200421880Bill m page 9 is for a kind of full-volume activity that is encoded in a video to estimate the complexity of the previous picture on the complex day-to-day surface. The current momentum is proportional to the reversed level. The present invention includes a movement and a more continuous parameter in the diurnal plane, that is, when the block is matched. The amount of activity is complex with the complex picture. This set of continuous instantaneous complexity signal bit rates The control of the calculation of the continuous continuous diurnal plane of the full-lived diurnal plane is the same as the estimated current diurnal continuous distribution of the group. A complexity estimation unit is connected to the current day surface, and the complexity of the same day is estimated in the full-continuous day surface of the previous picture, and the bit rate is given. 5. Description of the invention (5) The present invention is The current diurnal plane encodes the diurnal plane in the current diurnal plane, and it is based on the complexity of the mesh; the degree of mesh complexity and the average ratio of the diurnal plane of the current diurnal plane in the instantaneous sequence. A device that averages the other party's system. A set of the frontal dispatcher is connected with the motion estimation based on the current day code. The current day code is used to estimate the complexity of the current head dispatcher and the sequence of the dynamic image based on the current picture. The ratio of grouping complexity is more complex than the standard quota; therefore, it is based on coding. The current day-to-day calculation of current day-to-day activity is compared with the present. Also revealed a ^ estimation unit, new device. Move the to be coded and calculate the ratio, degree, of this item and the set of connectedness. The moment of the current picture, the method of assigning the target system, momentum, which is medium; then the moment of the complexity of the previous face of the basic type is the estimated degree of repetition. After the video, the The ratio of the bit rate to the current bit rate controller, a video sequence is used to perform the amount of activity. The basic type of pre-processor can therefore be used with complexity, based on the estimated current picture. 200421880
參數更新斋以目前書 位元數和目前晝面= = 動量、目前晝面之實際消耗 其活動量對複雜度之比7 I化位階為基礎,計算目前晝面 【實施方式】 的、特徵和優點能更明顯易懂, 並配合所附圖式,作詳細說明如 為使本發明之上述 下文特舉一較佳實施例 下:The parameter update is based on the current number of book positions and the current diurnal surface = = momentum, the actual consumption of the current diurnal surface, and the ratio of its activity to complexity. The advantages can be more obvious and easy to understand, and in conjunction with the accompanying drawings, a detailed description is provided. In order to make the above-mentioned following of the present invention a preferred embodiment is given below:
叫:汗。,預測用的卜晝面和P—晝面必須先在mpeg編石J 斋内午碼’利用計算原始晝面及解碼的晝面間的均方根 (root mean Square)誤差,可以得到解碼畫面其品質的笔 觀量度。以均方根誤差作為解碼晝面之失真度,則解碼書 面的複雜度(complex)-C定義如下: C = r X d 其中,r係編碼該晝面所需的位元率,d係解碼的晝面 之失真度。理論上’平均的量化位階(qUantizati〇n step s i z e)和失真度之間為一比例關係,故,畫面的複雜度可 定義成:Called: Khan. For prediction, the diurnal plane and P-diurnal plane must first be edited in mpeg J Zhai noon code 'by calculating the root mean square error between the original diurnal plane and the decoded diurnal plane to obtain the decoded picture. A pen-based measure of its quality. Taking the root mean square error as the distortion of the decoding day, the complexity of decoding the written complex-C is defined as follows: C = r X d where r is the bit rate required to encode the day and d is the decoding The degree of distortion of the day. In theory, there is a proportional relationship between the average quantization level (qUantizati〇n step s i z e) and the degree of distortion, so the complexity of the picture can be defined as:
0702-7302TWf(nl);90P97;Lin.ptd 第11頁 200421880 五、發明說明(7) 其中, ,將可得知 而該晝面的 晝面之 先前技術利 雜度,並且 同編碼形式 變化,則此 編碼的目標 如果視訊序 成一致的視 q係該畫面的平均量化 位元率(實際消耗位-位卩白。在晝面編碼之後 複雜度亦因此獲得。疋數)以及平均量化位階, 複雜度端视晝面的本⑽ 用最近編螞蚩而夕、〃貝及其編碼的形式而定。 為I—、及B-晝面八度來表示目前畫面之複 的影響,倘若視訊^別保持其複雜度以減輕不 方式可達5|丨^ ^ τ 列的晝面其内容為平順的 万式」運到相當不錯 ^ ^ ^ 位元配額與直實^ $ a/文果。然而,由於晝面 u认咖…丄 询耗位元數之間不相符合, 上迷方式並不適合用來達 覚口口貝。 根據本發明’預先分析目前查;Α 曰, .. U 口月U畺面的活動量(a C t i ν i t y ) 以及先前編,畫面的實際複雜度之間的關係,目前晝面之 複雜度將能藉此預估。活動量係一種在畫面編碼後其位元 率和品質的量度嘗試’假設一整張晝面的全活動量A正比 於複雜度C,則 A = k X C ^ X r X q = k^xr x a 2 丄 ^ 其中,k係比例常數。若k’為活動量對複雜度之比 (activity-t〇 - complexity ratio,ACR),則全活動里A 還 可以下列式子表示: A = ACRxr^<g 根據本發明,可以從相同型態的先前編碼畫面來預估ACR 。觀念上,先在編碼/張晝面前計算其全活動量A ’然後0702-7302TWf (nl); 90P97; Lin.ptd page 11 200421880 V. Description of the invention (7) Among them, will be able to know the degree of the prior art of the day surface and the degree of heterogeneity of the day surface, and the same as the encoding form, Then, if the video is sequenced into a consistent view q, the average quantization bit rate (actual consumption bits-bits of white) of the picture. The complexity is also obtained after day-time encoding. The number of digits and the average quantization level The complexity depends on the local surface of the day, using the recently edited version, encoding and encoding. I-, and B- octave of octave to indicate the effect of the current picture. If the video ^ don't keep its complexity to reduce the number of ways that can reach 5 | 丨 ^ ^ τ, the content of the day is smooth. "Shi" arrived pretty good ^ ^ ^ bit quota and straight real ^ $ a / wenguo. However, due to the inconsistency between the number of digits consumed by the day and time, the method of addiction is not suitable for reaching scallops. According to the present invention, 'analyze the current investigation in advance; A said: .. U mouth month U 畺 surface activity (a C ti ν ity) and the relationship between the actual complexity of the screen, the current day complexity Will be able to take this estimate. Activity amount is a measure of its bit rate and quality after the picture is encoded. 'Assuming that the full amount of activity A of a whole day is proportional to the complexity C, then A = k XC ^ X r X q = k ^ xr xa 2 丄 ^ where k is a proportional constant. If k 'is the activity-to-complexity ratio (ACR), then A in the whole activity can also be expressed by the following formula: A = ACRxr ^ < g According to the present invention, the same type State of the previously encoded picture to estimate ACR. Conceptually, first calculate its full activity A in front of the code / Zhang Tian, and then
0702-7302TWf(nl);90P97;Lin.ptd 第12頁 2004218800702-7302TWf (nl); 90P97; Lin.ptd Page 12 200421880
0702-7302TWf(nl);90P97;Lin.ptd 第13頁 200421880 五、發明說明(9) 訊緩衝器檢驗者(Video Buffer Verifier,VBV)規則,分 配的目標額度必須受限以使VBV緩衝器不會滿溢 (overflow)或匱乏(underflow)。當目前畫面之目標位元 額度決定好時,便可利用許多不同型態的影像編碼器,像 是MPEG-2標準描述的測試原型5(Test Model 5),根據目 標位元額度來完成晝面的壓縮。 目前晝面在完成編碼之後,其實際的複雜度將可得 知’而目前畫面的全活動量和實際複雜度之間的關係, ACR⑴可從以下計算得至q ACRU) =-ii(〇 AQ(l) χ [/3(0 其中,ACR⑴與全活動量Α(υ成正比,而與目前畫面的實際 消耗位元數UB(1)以及平均量化位階糾⑴成反比。此ACR⑴可 拿來預估同裂態的下張晝面之複雜度。ACR⑴可以和乩…⑴ 做線性結合來,免受到那些富含雜訊晝面的影響。 士發明的單次(Singie_pass)視訊編碼可藉由第3圖之 較‘:施例並配合第4圖之操作流程圖來 :雜碼器300包括-移動估算單湖-以及-麥像;DO配頜分派器3 30、—參數更新器340 移動估算的區塊前ί,’ Α(1)(步驟S410)。基於全活動;m之全活動量 切里Λ 以及XI組連續晝面中同 第14頁 〇7〇2-73〇2TWf(ni);9〇p97;Lin.ptd 200421880 五、發明說明(10) 型態之j丽編碼晝面其’複雜度估計器32〇可用來 估計目前畫面之複雜度CS (步驟S420 )。配額分派器33〇以 預估的複雜度ct)更新本組連續晝面其瞬間複雜度刪—c 二二且根據“及inst-c分配目標位元額度— Z - Μ ^ Β- ! Γ ^ ^S44〇) 〇 ^ ^ E'350 ^ ^ § ^ 位兀減TB而決定目前晝❹⑴所使用之量化位階,這些 量化位”作為目前畫面P⑴令的每一空間區域編碼之用: 以EG = σ此工間區域係指大區塊;根據以μ⑴所決定 的这些里化位階,影像編碼器3 5 〇編碼目前晝面ρ⑴( S450),且回報實際消耗位元數_⑴以及 _ 給參數f新器340。以全活動量Α⑴、實際消耗位元數:二 和平均—里化位階AQ。)為基礎’參數更新器34〇為目前晝面 P(〇計算^其活動量對複雜度之比ACR⑴(步驟S46〇)。— 在等效上,第3圖之較佳實施例可考以 是軟體來實現。根據本發明,移動估算單元31 3及圖= =1元二可,管線(P1PeUne)模式運作,在複編 益320開始叶舁目前晝面的複雜度之前,㈣估算單元川 :Ϊ ΐ Ϊ成即將被編碼的目前晝面其全活動量之計算以及 之估算,並且’當移動估算單元310為下張書面 =備艾複雜度估計器3 2 0和其他元件仍正忙於完:所 有人目則旦面相關的運作。接下來將 例中的每個元件。 I罕乂佳只施 移動估算單元31〇的主要目的之一為決定用何種預測 第15頁 0702-7302TWf(nl);90P97;Lin.ptd 200421880 五、發明說明(11) 模式來編碼一張晝面裡的每個Μ B,如果必要的話,亦進行 向前和向後的移動預測,還可從區塊匹配運算中提取晝面 活動量的資訊。首先,計算每個MB之自身活動量(intra-activity) , 將一個〇中4 個8x8 亮度值區塊之像素 強度以 丫111,11’11[1::=〇,...,7,]1=::0,...,7來表示,並且每個8又8區塊其 平均值為Γ ,則每個μ B之内活動量I n t r a A c t : 30702-7302TWf (nl); 90P97; Lin.ptd Page 13 200421880 V. Description of the invention (9) The rule of the Video Buffer Verifier (Video Buffer Verifier, VBV), the allocated target quota must be limited so that the VBV buffer does not Will overflow or underflow. When the target bit quota of the current picture is determined, many different types of image encoders, such as Test Model 5 described in the MPEG-2 standard, can be used to complete the daytime surface according to the target bit quota. Compression. At present, after the encoding of the daytime plane is completed, its actual complexity will be known ', and the relationship between the full activity of the current picture and the actual complexity, ACR⑴ can be calculated from the following to q ACRU) = -ii (〇AQ (l) χ [/ 3 (0, where ACR⑴ is proportional to the full activity A (υ, but inversely proportional to the actual number of bits UB (1) and the average quantization level correction of the current picture. This ACR⑴ can be taken Estimate the complexity of the next diurnal surface in the same split state. ACR⑴ can be linearly combined with 乩 ... ⑴ to avoid the influence of the noisy rich diurnal surface. The single-pass (Singie_pass) video coding invented by the scholar can be obtained by Comparison of FIG. 3: An example and the operation flowchart of FIG. 4 are used to illustrate: The cipher device 300 includes-motion estimation single lake-and-wheat image; DO with jaw dispatcher 3 30,-parameter updater 340 movement In front of the estimated block, Α (1) (step S410). Based on the full activity; the full activity volume of m, Chery Λ, and the continuous diurnal plane of group XI are the same as on page 14 〇07〇2-73〇2TWf (ni ); 9〇p97; Lin.ptd 200421880 V. Description of the invention (10) The J-coded daytime surface and its' complexity estimator 32 can be used to estimate the target The complexity CS of the picture (step S420). The quota dispatcher 33 uses the estimated complexity ct to update the instantaneous complexity of this group of day-to-day planes—c 22 and allocate the target bit quota according to "and inst-c" — Z-Μ ^ Β-! Γ ^ ^ S44〇) 〇 ^ ^ E'350 ^ ^ § ^ Decrease the TB level to determine the quantization level currently used in the day time. These quantization bits are used as each The use of a spatial area coding: EG = σ This inter-work area refers to a large block; according to these internalization levels determined by μ⑴, the image encoder 3 5 0 encodes the current day surface ρ⑴ (S450), and returns the actual The number of consumed bits _⑴ and _ give the parameter f a newer 340. Based on the full activity Α⑴, the actual number of consumed bits: two and the average-Lihua rank AQ.) 'Parameter updater 340 is the current day surface P (〇 Calculate ^ its activity to complexity ratio ACR⑴ (step S46). — Equivalently, the preferred embodiment of FIG. 3 can be implemented by software. According to the present invention, the mobile estimation unit 31 3 And the figure = = 1 yuan and two can, the pipeline (P1PeUne) mode operation, starting in the copy of Yi 320, Ye Ye currently face Prior to the complexity, ㈣Estimation Unit Chuan: Ϊ ΐ Ϊ into the calculation and estimation of the current day's full activity that will be coded, and 'when the mobile estimation unit 310 is the next sheet = Bei Ai Complexity Estimator 3 2 0 and other components are still busy: everyone is concerned about the operation. Next, each component in the example will be used. One of the main purposes of the mobile estimation unit 31 is to determine what prediction to use. Page 15 0702-7302TWf (nl); 90P97; Lin.ptd 200421880 V. Description of the invention (11) Mode to encode each BM in a daytime plane, and if necessary, forward and backward movement prediction , Can also extract the information of day-to-day activity from the block matching operation. First, calculate the intra-activity of each MB, and divide the pixel intensities of 4 8x8 luminance blocks in a 〇 to 111,11'11 [1 :: = 〇, ..., 7, ] 1 = :: 0, ..., 7, and the average value of each 8 and 8 blocks is Γ, then the activity in each μ B I ntra A ct: 3
IntraAct = 其中IntraAct = where
倘若需要較低的計算複雜度,可以相對f的絕對差值代 替:If a lower computational complexity is required, it can be replaced by the absolute difference of f:
IntraAct = lAk 其中 = Σ Σ ; - 7 m=0?2=0 即為I-晝面的‘個活二2式.I核式,因此htraAct 估算。最常使 對非自身編碼 用 如果晝面為P-或B-晝面,則需進行 來發現最佳移動向量之技術係區塊匹配。IntraAct = lAk where = Σ Σ;-7 m = 0? 2 = 0 is the ‘live 2 form. I kernel form of the I-day surface, so htraAct is estimated. It is most commonly used for non-self-encoding. If the diurnal plane is P- or B-diurnal plane, the technique to find the best motion vector is block matching.
200421880 五、發明說明(12) (nonintra coding)而 士 準如變異數或失真絕對。值:和^晝面),利用將失」票 後、雙向的預測或不需最低’來選擇向前J MB之MB模式決定,在每個^貝。一旦P_或卜晝面中母個 之變異數亦可求得,^ 2補償的差讎中4働8區塊 Μ間像素對像素之差= = 失真絕對值和相加以i;匕二個8x8區塊之變… 曰 K付非内編碼晝面中每個MB之相互活 + ^ aetlVlty) ’ InterAct,然後把非内編碼晝面 中母個MB之IntraAct及其InterAct拿來作比較,以判斷 i^erAct是否較小,若是,則以InterAct作為該Μβ之活動 里,亚以相互模式(lnter_m〇de)編碼該MB ;否則以 作為該MB之活動量,並以自身模式(intra_m〇de) 編。最後,對目前的Ϊ一、卜或^晝面,將所柳之 活動置相加而得全活動量A⑴’移動估算單元31〇再把全活 動量A(n傳送給複雜度估計器32〇及參數更新器34〇。 接著,複雜度估計器32 0為某型態的目前晝面^^估筲 其複雜=,且依照三種晝面型態,弓i進加權係數至預估^ 複雜度巧丨。因為絕不能以Β-畫面預估其他的晝面,故可減 少Β-晝面之加權係數以分配較少的位元給^晝面而保留較 多的位元給I-和Ρ-晝面;一般而言,編碼卜晝面會產生 多的位元,因此Ρ-晝面之加權係數又小於卜畫面之加權係 數。複雜度估計器3 2 0根據目前晝面的型態更新複雜度200421880 V. Description of the invention (12) (nonintra coding) The standard is the number of variations or absolute distortion. Values: and ^ day face), using the forward-to-backward MB prediction, the two-way prediction or the minimum required 'option to select the MB mode forward J MB, at each time. Once the number of variations in P_ or Bu day surface can also be calculated, the difference between the pixel-to-pixel difference between the 4 and 8 blocks in the 2 ^ 2 compensation difference = = the absolute value of the distortion and the addition i; The change of the 8x8 block ... It means that K pays for each MB in the non-intra-coded daytime plane + ^ aetlVlty) 'InterAct, and then compares the IntraAct and InterAct of the parent MBs in the non-intra-coded daytime plane to compare Determine whether i ^ erAct is small. If it is, use InterAct as the Mβ activity to encode the MB in mutual mode (lnter_m〇de); otherwise, use it as the MB's activity and use its own mode (intra_m〇de). ). Finally, for the current day, day, or day, the activities of all willows are added together to obtain a full activity A⑴ ', the mobile estimation unit 31, and then the full activity A (n) is transmitted to the complexity estimator 32. And the parameter updater 34. Next, the complexity estimator 32 0 estimates the complexity of the current day surface ^^ for a certain type, and according to the three day surface types, the weighting coefficient is advanced to the estimated ^ complexity. Coincidentally, since it is impossible to estimate other daytime planes with the B-picture, the weighting coefficient of the B-daytime plane can be reduced to allocate fewer bits to the daytime plane and retain more bits to I- and P. -Diurnal plane; generally speaking, encoding the diurnal plane will produce many bits, so the weighting coefficient of P-diurnal plane is smaller than the weighting coefficient of the diurnal plane. The complexity estimator 3 2 0 is updated according to the current diurnal plane type. the complexity
200421880 五、發明說明(13) CP或CB三者其中之一,目前晝面P(1)之複雜度^〗丨以如下方式 估算: if (I-晝面) ^(0 匕esf A(〇 Cr = KrX -γ—ΓΤ 1 1 ACRf~l) else if (P-晝面)200421880 V. Description of the invention (13) One of CP or CB. The complexity of the current day surface P (1) ^〗 丨 Estimated as follows: if (I-day surface) ^ (0 0esf A (〇 Cr = KrX -γ—ΓΤ 1 1 ACRf ~ l) else if (P- 天 面)
_ y 一 Up 一厶P_ y one Up one 厶 P
XACRfTJ else if (B-晝面) 0卜 Cs = Ksx A(i) ACR^l) 其中,及‘4C/?fu係一組連續晝面中,I、P 和B型態的先前編碼晝面各自的活動量對複雜度之比。這 些H广υ 和乂 α?Γ;存放在參數更新器340之中,而 複雜度估計器3 2 0會為適當型態的先前晝面讀取對應的活 動量對複雜度之比。I、ΚΡ及ΚΒ分別係I -、Ρ -和Β -晝面之加 權係數,其範圍一般是在0到1. 0之間。至於I -晝面,較佳 實施例可以採用I = 1 · 0。 當收到β之時,配額分派器3 3 0為該組連續晝面更新 其瞬間複雜度INST_C且分配目標位元額度TB(1)給目前晝面 P(1),瞬間複雜度INST_C之更新如下:XACRfTJ else if (B-day surface) 0 b Cs = Ksx A (i) ACR ^ l) where, and '4C /? Fu is a previously coded day surface of type I, P, and B in a set of continuous day surfaces The ratio of respective activities to complexity. These H υ and 乂 α? Γ; are stored in the parameter updater 340, and the complexity estimator 3 2 0 will read the corresponding amount of activity to complexity for the previous type of the previous day. I, Κ and Β are weighting coefficients of I-, P-, and B-, respectively, and their ranges are generally between 0 and 1.0. As for the I-day surface, the preferred embodiment can use I = 1 · 0. When β is received, the quota dispatcher 3 3 0 updates the instantaneous complexity INST_C for the group of continuous daytime planes and allocates the target bit quota TB (1) to the current daytime plane P (1), and updates the instantaneous complexity INST_C. as follows:
0702-7302TWf(ni);90P97;Lin.ptd 第18頁 2004218800702-7302TWf (ni); 90P97; Lin.ptd page 18 200421880
五、發明說明(14) INST _C = NIxCI + NpxCp +NBxC3 其中 數目 、NP和NB分別是本組連續晝面中I -、p -及b 再者,目前畫面P(1)之目標位元額度TB(i)為 晝面之 TB(i)=V. Description of the invention (14) INST _C = NIxCI + NpxCp + NBxC3 where the number, NP and NB are I-, p-and b in the continuous diurnal plane of the group. Furthermore, the target bit quota of the current picture P (1) TB (i) is TB (i) =
INST C 其中’ n係本組連續晝面之晝面數,f係每秒晝面數,即· 圖幀率(f rame rate ),R則係每組連續晝面之期望早 一十、 , τ太勻位 疋率。然而,實際的消耗位元數並不會和期望的#;〜 丨儿凡領度 剛好相等,因此發展一種回授策略來使晝面實際消耗的 元數接近於目標位元額度,在較佳實施例中,編碼INST C where 'n is the number of continuous diurnal planes in this group, f is the number of diurnal planes per second, that is, the frame rate (frame rate), and R is the expectation of continuous diurnal planes in each group ten, τ is too uniform. However, the actual number of bits consumed will not be exactly equal to the expected number of children; therefore, a feedback strategy is developed to make the actual number of bits consumed by the day close to the target bit amount. In the embodiment, the encoding
^ Θ 為 |J 為止的超用位元量,可由正在編碼的目前晝面攤還一 份:^ Θ is the amount of overused bits up to J, which can be amortized by the current daytime being coded:
TB(i) =TB(i) -JLRxEB 其中,EB係參數更新器340傳來的超用位元量,AR則係既 定之攤還率,其範圍一般是在〇 · 〇 5到〇 · 2之間。配額分派 器3 3 0須調整目標位元額度TB⑴以符合VBV規範,所以還定 下額度的上限(U-boiind)及下限(L-b〇und)。就固定位元^ (constant bitrate,CBR)的操作而言,分配給一張晝面 的目標位元額度須使VBV緩衝器不會滿溢或匱乏,因此目 標位元額度TB⑴限制在上、下限範圍内: i f (TB(i)>U —bound) then TB⑴—bound if(TB⑴<L —bound) then TB⑴=L —bound 若是可變位元率(variable bitrate,VBR)的操作,則只TB (i) = TB (i)-JLRxEB where EB is the over-bit amount transmitted by the parameter updater 340, and AR is the predetermined amortization rate, and its range is generally from 0.5 to 0.2 between. The quota dispatcher 3 3 0 must adjust the target bit quota TB⑴ to comply with the VBV specification, so it also sets the upper limit (U-boiind) and lower limit (L-bund) of the quota. For constant bitrate (CBR) operations, the target bit quota allocated to a daytime plane must be such that the VBV buffer will not overflow or be scarce, so the target bit quota TB⑴ is limited to the upper and lower limits. In the range: if (TB (i) > U —bound) then TB⑴—bound if (TB⑴ < L —bound) then TB⑴ = L —bound If it is a variable bitrate (VBR) operation, only
200421880 五、發明說明(15) 一 " --------:— 要防止VBV匱毛即可,故: 梦德if(JB:〉u-bound) then TB(1)=u』_d …、後,目軚位元額度TB(n會被傳送到泉 像編碼器3 5 0。 數更新為3 4 0以及影 收佥==器35()透過訊號線312從移動估算單元310接 貝枓U及每個MB之移動向量和MB模式。為了儘 最m像編碼器3 50依據目標位元額 "^來決疋編碼目丽畫面的每個mb所用之量化位階,以 這些資料為基礎,影像編碼器350開始對目前晝面?⑴進行 編碼且,出編碼後的資料CD。在目前晝面p⑴編碼之後,影 像編碼器3 5 0將量化位階做平均並且計算實際消耗位元數/ UB(1),再回報目前畫面的υβ⑴以及平均量化位階⑴給夾 數更新器3 4 0。 少 晝面Ρ⑴之目標位元額度ΤΒ⑴以及實際消耗位元數⑽⑴ 間的差距,會由參數更新器340做累計以便在晝面ρ⑴編碼 之後得到超用位元量ΕΒ : ΕΒ - EBx(l-AR)+UB(i)-TB(i) 其中,AR係既定之攤還率。因此,活動量和實際複雜度之 間的關係,ACR⑴,可從以下計算得到 ACR(i)=: A(i) AQU) x UB{i) 之複雜度。為使 較佳實施例利用了 此ACR(1)可用來預估同型態的下張晝面 ACR(l)不會受到富含雜訊畫面的影響,200421880 V. Description of the invention (15) One " --------:-To prevent VBV from being short of hair, so: Mengde if (JB:> u-bound) then TB (1) = u ' _d…, after that, the target bit quota TB (n will be transmitted to the spring image encoder 3 5 0. The number is updated to 3 4 0 and the video is received == 35 () from the mobile estimation unit 310 through the signal line 312 Connect U and each MB's motion vector and MB mode. In order to maximize the image, the encoder 3 50 determines the quantization level used to encode each mb of the Muli picture according to the target bit amount " ^. Based on the data, the image encoder 350 starts to encode the current diurnal plane, and outputs an encoded data CD. After the current diurnal plane, the image encoder 350 averages the quantization levels and calculates the actual consumed bits. Unit number / UB (1), and then return the current picture υβ⑴ and the average quantization level ⑴ to the clip number updater 3 4 0. The difference between the target bit quota TB and the actual number of consumed bits of the diurnal plane P⑴ will be Accumulation is performed by the parameter updater 340 so as to obtain the super-used bit amount EB after daytime ρ⑴ encoding: EB-EBx (l-AR) + UB (i)- TB (i) Among them, AR is a predetermined amortization rate. Therefore, the relationship between the amount of activity and the actual complexity, ACR⑴, can be calculated from the following ACR (i) =: A (i) AQU) x UB {i ) Complexity. In order to make use of the preferred embodiment, this ACR (1) can be used to predict the same type of next diurnal surface ACR (l) will not be affected by noise-rich pictures.
200421880 五、發明說明(16) ACR(1)和ACRd-1)的線性組合,例如·· if (I-畫面)200421880 V. Description of the invention (16) Linear combination of ACR (1) and ACRd-1), such as if (I-picture)
= ACRa-; χ cwy ACr (i) X CW else if (P~ 晝面)= ACRa-; χ cwy ACr (i) X CW else if (P ~ day surface)
= ACR(;~l) X (l - cw)+ ACRa) X CW else if (B-晝面) "4〇βί; ACR (J'l) (l - CW)+ ACR(i) x CW /、中CW仏線性組合既定之加權係數。更新過的、 型態的書面做;ΐ傳;::複:ΪΓ計器320來為下一張適當 關作一為位元分:,之此用外’超用位元量⑽則送至配額分派 雖然本發明已以一具體實施例揭露如上,缺立 易於,明本發明之技術内容,而並非將本發明狹義地I 於该貫施你J,任何熟習此技藝者,在不脫離本發明之J: 和範圍内’當可作些許之更動與潤飾 ::申 範圍當視後附之申請專利範圍所界本U之保護 第21頁 0702-7302TWf(nl);90P97;Lin.ptd 200421880 圖式簡單說明 第1圖係以顯示的順序呈現典型的畫面群組(GOP); 第2圖係MPEG之大區塊; 第3圖係本發明較佳實施例的位元率控制裝置之方塊 不意圖,以及 第4圖係本發明之操作流程圖。 符號說明 3 0 0〜動態影像視訊編碼器; 310〜移動估算單元; 3 1 2〜訊號線; 3 2 0〜複雜度估計器; 3 3 0〜配額分派器; 3 4 0〜參數更新器; 3 5 0〜影像編碼器; P(i)〜晝面; A⑴〜全活動量; 〜預估複雜度; ACR(i)〜活動量對複雜度之比; EB〜超用位元量; TBm〜目標位元額度; AQ(1)〜平均量化位階; UB(i)〜實際消耗位元數; C D〜編碼後的資料。= ACR (; ~ l) X (l-cw) + ACRa) X CW else if (B- 天 面) "4〇βί; ACR (J'l) (l-CW) + ACR (i) x CW /, The predetermined weighting coefficient of the CW 仏 linear combination. Updated and typed written work; Biography; :: Complex: ΪΓ Counter 320 to make a bit for the next one :, the extra amount of extra bits will be sent to the quota Distributing Although the present invention has been disclosed as above with a specific embodiment, it is easy to lack and clarify the technical content of the present invention, but not to narrowly limit the present invention to you. Anyone skilled in this art will not depart from the present invention. J: and within the scope, when you can make a few changes and retouching :: Application scope Dang depending on the protection of the scope of the patent application attached at the end of page 21 0702-7302TWf (nl); 90P97; Lin.ptd 200421880 Figure Brief description of the formulas Figure 1 shows a typical group of pictures (GOP) in the order of display; Figure 2 is a large block of MPEG; Figure 3 is a block diagram of a bit rate control device of a preferred embodiment of the present invention The intention, and FIG. 4 are flowcharts of the operation of the present invention. Symbol description 3 0 0 ~ Motion video encoder; 310 ~ Motion estimation unit; 3 1 2 ~ Signal line; 3 2 0 ~ Complex estimator; 3 3 0 ~ Quota dispatcher; 3 4 0 ~ Parameter updater; 3 5 0 ~ Image encoder; P (i) ~ Day surface; A⑴ ~ Full activity amount; ~ Estimated complexity; ACR (i) ~ Ratio of activity amount to complexity; EB ~ Over-use bit amount; TBm ~ Target bit quota; AQ (1) ~ average quantization level; UB (i) ~ actual number of bits consumed; CD ~ encoded data.
0702-7302TWiInl);90P97;Lin.ptd 第22頁0702-7302TWiInl); 90P97; Lin.ptd Page 22
Claims (1)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW92107989A TWI221741B (en) | 2003-04-08 | 2003-04-08 | Method and apparatus for rate control in moving picture video compression |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW92107989A TWI221741B (en) | 2003-04-08 | 2003-04-08 | Method and apparatus for rate control in moving picture video compression |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TWI221741B TWI221741B (en) | 2004-10-01 |
| TW200421880A true TW200421880A (en) | 2004-10-16 |
Family
ID=34389011
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW92107989A TWI221741B (en) | 2003-04-08 | 2003-04-08 | Method and apparatus for rate control in moving picture video compression |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TWI221741B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013021824A1 (en) | 2011-08-05 | 2013-02-14 | ソニー株式会社 | Reception device, reception method, program, and information processing system |
-
2003
- 2003-04-08 TW TW92107989A patent/TWI221741B/en not_active IP Right Cessation
Also Published As
| Publication number | Publication date |
|---|---|
| TWI221741B (en) | 2004-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10013746B2 (en) | High dynamic range video tone mapping | |
| CN102326391B (en) | Multi-view image coding device, multi-view image decoding method, multi-view image decoding device, multi-view image decoding method | |
| US8406297B2 (en) | System and method for bit-allocation in video coding | |
| CN101729891B (en) | Method for encoding multi-view depth video | |
| US10469851B2 (en) | Advanced video coding method, system, apparatus, and storage medium | |
| CN114786019B (en) | Image prediction method, encoder, decoder, and storage medium | |
| CN108200431B (en) | Bit allocation method for video coding code rate control frame layer | |
| CN103634601B (en) | Structural similarity-based efficient video code perceiving code rate control optimizing method | |
| CN105432077A (en) | Adjusting quantization/scaling and inverse quantization/scaling when switching color spaces | |
| CN106210721B (en) | A kind of quick code check code-transferring methods of HEVC | |
| US20040202247A1 (en) | Method and apparatus for moving picture compression rate control using bit allocation with initial quantization step size estimation at picture level | |
| US20120275509A1 (en) | Region-of-interest encoding enhancements for variable-bitrate mezzanine compression | |
| TW200533206A (en) | A video encoder and method of video encoding therefor | |
| CN104994387B (en) | A kind of bit rate control method of fused images feature | |
| CN101404765A (en) | Interactive multi-view point video encoding method | |
| US9398314B2 (en) | Entropy coding efficiency enhancement utilizing energy distribution remapping | |
| Flack et al. | Low-bandwidth stereoscopic image encoding and transmission | |
| TW200421880A (en) | Method and apparatus for rate control in moving picture video compression | |
| CN101969559A (en) | Transpression method of video code stream and system of same | |
| Guo et al. | Quality-progressive coding for high bit-rate background frames on surveillance videos | |
| TWI232684B (en) | Apparatus for variable bit rate control in video compression and target bit allocator thereof | |
| TWI246862B (en) | Video coding/decoding buffering apparatus and buffering method thereof | |
| TWI221739B (en) | Method and apparatus for moving picture compression rate control using bit allocation with initial quantization step size estimation at picture level | |
| TWI226804B (en) | Apparatus for constant quality rate control in video compression and target bit allocator thereof | |
| KR100327102B1 (en) | A conversion method of compressed moving video on the video communication system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| MM4A | Annulment or lapse of patent due to non-payment of fees |