《TW2835比较完整版本.docx》由会员分享，可在线阅读，更多相关《TW2835比较完整版本.docx（47页珍藏版）》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、 QCIF全称Quarter common intermediate format。QCIF是常用的标准化图像格式。在H.323协议簇中，规定了视频采集设备的标准采集分辨率。QCIF = 176144像素。 CIF是常用的标准化图像格式（Common Intermediate Format）。在H.323协议簇中，规定了视频采集设备的标准采集分辨率。CIF = 352288像素CIF格式具有如下特性： (1) 电视图像的空间分辨率为家用录像系统(Video Home System，VHS)的分辨率，即352288。 (2) 使用非隔行扫描(non-interlaced scan)。 (3) 使
2、用NTSC帧速率，电视图像的最大帧速率为30 000/100129.97幅/秒。 (4) 使用1/2的PAL水平分辨率，即288线。 (5) 对亮度和两个色差信号(Y、Cb和Cr)分量分别进行编码，它们的取值范围同ITU-R BT.601。即黑色=16，白色=235，色差的最大值等于240，最小值等于16。 下面为5种CIF 图像格式的参数说明。参数次序为“图象格式 亮度取样的象素个数(dx) 亮度取样的行数 (dy) 色度取样的象素个数(dx/2) 色度取样的行数(dy/2)”。 o sub-QCIF 12896 64 48 o QCIF 176144 88 72 o CIF 352288
3、176 144 o 4CIF 704576 352 288（即我们经常说的D1） o 16CIF 14081152 704 576目前监控行业中主要使用Qcif（176144）、CIF（352288）、HALF D1（704288）、D1（704576）等几种分辨率，CIF录像分辨率是主流分辨率，绝大部分产品都采用CIF分辨率。目前市场接受CIF分辨率，主要理由有四点：1、目前数码监控要求视频码流不能太高；2、视频传输带宽也有限制；3、使用HALF D1、D1分辨率可以提高清晰度，满足高质量的要求，但是以高码流为代价的。在现阶段，出现了众多D1的产品，但市场份额非常小；4、采用CIF分辨率，信
4、噪比在32db以上，一般用户是可以接受的，但不是理想的视频图像质量。目前业内人士正在尝试用HALF D1来寻求CIF、D1之间的平衡。但随着单块硬盘的容量达到750GB甚至1000GB，而国内的大部分DVR已经可以做到连接8块1000GB的硬盘，故D1逐渐会变成时常的主流。 DCIF分辨率是什么？经过研究发现一种更为有效的监控视频编码分辨率（DCIF），其像素为528384。DCIF分辨率的是视频图像来历是将奇、偶两个HALF D1，经反隔行变换，组成一个D1（720*576），D1作边界处理，变成4CIF（704576），4CIF经水平3/4缩小、垂直2/3缩小，转换成528384.5283
5、84的像素数正好是CIF像素数的两倍，为了与常说的2CIF（704*288）区分，我们称之为DOUBLE CIF，简称DCIF。显然，DCIF在水平和垂直两个方向上，比Half D1更加均衡。为什么选用DCIF分辨率？数字化监控行业对数字监控产品提出两项要求：首先要求数据量低，保证系统能够长时间录像和稳定实时的网络传输；其次要求回放图像清晰度高，满足对细节的要求。而DCIF分辨率在目前的软硬件平台上，能很好的满足以上两项要求。Half D1分辨率已被部分产品采用，用来解决CIF清晰度不够高和D1存储量高、价格高昂的缺点。但由于他相对于CIF只是水平分辨率的提升，图像质量提高不是特别明显，但码流
6、增加很大。经过对大量视频信号进行测试，基于目前的视频压缩算法，DCIF分辨率比Half D1能更好解决CIF清晰度不够高和D1存储量高、价格高昂的缺点，用来解决CIF和4CIF，特别是在512Kbps码率之间，能获得稳定的高质量图像，满足用户对较高图像质量的要求，为视频编码提供更好的选择。CIF清晰度不够高和D1存储量高、价格高昂的缺点分辨率，静态回放分辨率理论上最高可达360TVline的图像质量，超过模拟监控中标准VHS磁带录像机280TVline的图像水平，达到公安部安防行业视频标准二级和三级项目的清晰度要求，满足绝大部分视频监控的要求。 什么是D1？做闭路电视监控系统这一行久了，大家都
7、以为D1是硬盘录像机显示、录像、回放的分辨率，实际上不是的，D1是数字电视系统显示格式的标准，共分为以下5种规格： D1：480i格式（525i）：720480（水平480线，隔行扫描），和NTSC模拟电视清晰度相同，行频为15.25kHz，相当于我们所说的4CIF（720576） D2：480P格式（525p）：720480（水平480线，逐行扫描），较D1隔行扫描要清晰不少，和逐行扫描DVD规格相同，行频为31.5kHz D3：1080i格式（1125i）：19201080（水平1080线，隔行扫描），高清放松采用最多的一种分辨率，分辨率为19201080i/60Hz，行频为33.75kH
8、z D4：720p格式（750p）：1280720（水平720线，逐行扫描），虽然分辨率较D3要低，但是因为逐行扫描，市面上更多人感觉相对于1080I（实际逐次540线）视觉效果更加清晰。不过个人感觉来说，在最大分辨率达到19201080的情况下，D3要比D4感觉更加清晰，尤其是文字表现力上，分辨率为1280720p/60Hz，行频为45kHz D5：1080p格式（1125p）：19201080（水平1080线，逐行扫描），目前民用高清视频的最高标准，分辨率为19201080P/60Hz,行频为67.5KHZ。其中D1 和D2标准是我们一般模拟电视的最高标准，并不能称的上高清晰，D3的108
9、0i标准是高清晰电视的基本标准，它可以兼容720p格式，而D5的1080P只是专业上的标准，并不是民用级别的，上面所给出的60HZ只是理想状态下的场频，而它的行频为67.5KHZ，目前还没有如此高行频的电视问世，实际在专业领域里1080P的场频只有24HZ，25HZ和30HZ。需要指出的一点是，D端子是日本独有的特殊接口，国内电视几乎没有带这种接口的，最多的是色差接口，而色差接口最多支持到D4，理论上肯定没有HDMI（纯数字信号，支持到1080P)的最高清晰度高，但在1920：1080以下分辨率的电视机上，一般也没有很大差别。国内主流的硬盘录像机（DVR，Digital Video Recor
10、ding）采用什么分辨率？怎样计算硬盘容量？国内主流的硬盘录像机采用两种分辨率：CIF和4CIF（D1），分为两种型号。硬盘录像机常见的路数有1路、2路、4路、8路、9路、12路和16路。最大可以连接8块2000GB的硬盘，总容量可高达1.6TB（可续目前市面上最大的硬盘只有1000GB），如果采用CIF分辨率，通常计算硬盘录像机的录像的硬盘容量为180MB250MB/小时，通常情况下取值200MB/小时；如果是D1的分辨率每小时录像需要的硬盘容量为720MB1000MB/小时，通常情况下为了减少硬盘的容量可以按照500MB/小时计算，帧率智能设置比25fps少一些，码流也要少一些！相信大家可
11、以计算出一台装满8块500GB的16路硬盘录像机可以录像多长时间了吧？TW2835 将6个主要程序整合进一个芯片中，其中包括：4个高质量的 NTSC/PAL 视频解码器、4个音频模拟数字转换器、一个音频复用器，双色显示控制器，双视频译码器以及一个先进的 OSD（屏幕菜单式调节方式）。在其众多功能中，TW2835 支持详尽的实时 D1 录制、在重放过程中将信道 ID 信息添加到视频流媒体中，用于自动解码与显示，同时还包含一个5层的图形覆盖功能，为 OSD（屏幕菜单式调节方式）、单盒、2D 阵列箱以及鼠标指示器显示特征/位图。TW2835 还包含一个简单的界面，使用多段连接支持多达16个信道系统。
12、此外，TW2835 还嵌入了几种特别的监视功能，其中包括：运动监测、放大以及水平与垂直缩放控制。凭借置入旨在减少交叉噪音的反锯齿过滤器和高质量的梳状过滤器，TW2835 已经成为一种针对 DVR（数码视频录像机） 与 Quad/Multiplexers 的高性能、具有成本效益的解决方案。IntroductionBlock Diagram:Functional Description:Video Input:ITU-R BT 656 = CCIR656中国电子顶级开发网-最专业的电子论坛、最专业的电子工程师博客Fi4S2I6HXITU-R BT 656: 9芯，不需要同步信号；8位数据传输；串行
13、视频传输；传输速率是601的2倍；先传Y，后传UV。656输出的是串行数据，行场同步信号嵌入在数据流中; 601是并行数据，行场同步有单独输出;The 656 use the 4:2:2 video coding protocol.The data signals are in the form of binary information coded in 8-bit words. These signals are:mU#s+?*C9X+K0fa0 video data; timing reference codes; ancillary data;(辅助数据) identification
14、 codes.中国电子顶级开发网-最专业的电子论坛、最专业的电子工程师博客h*dzeCd v0q HData format:)Om*Uz-A BT.656 data stream is a sequence of 8-bit or 10-bit bytes, transmitted at a rate of 27 Mbyte/s. Horizontal scan lines of video pixel data are delimited in the stream by 4-byte long SAV (Start of Active Video) and EAV (End of Acti
15、ve Video) code sequences. SAV codes also contain status bits indicating line position in a video field or frame. Line position in a full frame can be determined by tracking SAV status bits, allowing receivers to synchronize with an incoming stream. Individual pixels in a line are coded in YCbCr form
16、at. After a SAV code (4 bytes) is sent, the first 8 bits of Y (luma) data are sent then 8 bits of Cb (chroma U), followed by 8 bits of Y for the next pixel and then 8 bits of Cr (chroma V). To reconstruct complete Y,Cb,Cr pixel values, chroma subsampling must be usedThe 656 only support the 4:2:2。Th
17、e 656 sampling rate is 27M。 The 656 has embedded (the digital data include the coding information for Hs Vs and Field)Sync(Called Video Timing Reference Codes SAV,EAV),656 has field informations.中国电子顶级开发网-最专业的电子论坛、最专业的电子工程师博客f3Cj!V ?6uol2uW$fmAnalog video input: 支持所以的NTSC/PAL视频标准，包含有自动标准检测电路。该电路可以被覆
18、盖当写IFMTMAN和IFORMAT寄存器。如果检测到视频损耗，将被迫运行由IFORMAT寄存器知道的视频标准模式。表1显示TW2835支持的视频输入标准：Anti-aliasing Filter:带宽6MHz.Analog-to-Digital Converter(数模转换):10bit ADC.每个ADC有2个模拟开关由ANA_SW寄存器控制。ADC也可进入低功耗模式由ADC_PWDN寄存器控制。Sync Processing（同步处理）：用于在混合视频信号中检测水平和垂直的同步信号。水平同步处理器包含同步分离器，PLL，相关逻辑。同步分离器通过低通滤波，附加的逻辑来避免假检测。PLL来防
19、止抖动。垂直同步分离器检测垂直同步信号。场状态由垂直同步时间来决定。当检测到垂直信号的位置和水平同步信号一致，它暗示了帧的起始或是奇数场的开始。否则，表示偶数场。Color Decoding（色彩编码）：色度和亮度分离由自适应的梳状滤波器是实现。色度Trap Filter（陷波滤波器）来调整视觉性能：对于NTSC陷波频率为3.5MHz，PAL陷波频率为4.5MHz左右。Luminance Processing（亮度处理）： 亮度信号由自适应梳状滤波器分离得到，并输送到峰化电路。峰化电路加强亮度信号的高频分量通过Y_PEAK寄存器。图片对比度调整通过CONT寄存器，调整范围0-200%亮度的调整
20、通过BRT寄存器，调整范围25IRE。Chrominance Processing（色度处理）： 为NTSC和PAL第一个正交混频来实现色度解调。混频=NTSC和PAL的子载波频率。在混频后，LPF用于除去载波的2倍频和场色度分量。若IF（中频）源失调，则，IF补偿滤波器来补偿在高频的衰减或者在色度子载波附近的不对称性。色度上边带增益由IFCOMP寄存器来控制。ACC(自动色度增益控制)用于补偿在视频信号中由于高频抑制引起的降低色度幅度。大致增益范围：-6dB到30dB。对于黑白视频或者弱噪声信号，内部的色彩killer 电路将关闭色彩。Color killing function 能够由可编
21、程的CKIL寄存器使能。色彩饱和度可由SAT寄存器来改变。Cb、Cr增益可分别由UGAIN和VGAIN寄存器来调整。Cb、Cr offset 可由U_OFF和V_OFF寄存器调整。色相（HUE）可由HUE寄存器来调整。要调节的参数相应的寄存器饱和度SAT（0x08,0x18,0x28,0x38）Cb、Cr GainUGAIN(0x48)、VGAIN(0x49)Cb、Cr OffsetU_OFF(0x4A)、V_OFF(0x4B)HUEHUE(0x07,0x17,0x27,0x37)Real-time Record Mode（实时记录模式）: 4个channel Real-time record
22、 outputs with full D1 format through the DLINK1 and MPP1/2 pins.并支持H/V/F信号。 DLINK1和MPP1/2 pins 通过 MPP_MD(1xB0)和MPP_SET(1xB1,1xB3,1xB5)寄存器控制。Digital Video Input： 用于playback支持8bit ITU-R BT.656 standard。Channel ID decoder 用于auto cropping and strobe function.Channel ID Decoder:3 kinds of channel ID: Use
23、r channel ID, Detection channel ID, and auto channel ID.The User channel ID is used for customized information like system information and date.The Detection channel ID is used for detected information of current live input such as motion, video loss, blind and night detection information.The auto c
24、hannel ID is employed for automatic identification of picture configuration which includes the channel number, analog switch, event, region enable and field/frame mode information.The TW2835 also supports both analog and digital type channel ID during VBI period. The digital channel ID has priority
25、over analog channel ID.VBI是Vertical Blanking Interval的缩写，中文意思是场消隐期，也叫场逆程。电视节目称为正程信号。在625/25的电视信号中，可以利用的电视行从第6/318行开始至第22/334行止，每场最多可插入17个数据行。VBI采集就是指对场消隐期中的数据采集。 The analog type channel ID decoding is enabled via the VBI_ENA (1x86) register. The digital type channel ID decoding is operated via VBI_C
26、ODE_EN (1x86) register. Additionally to detect properly the analog channel ID against noise such as VCR source, the channel ID LPF can be enabled via the VBI_FLT_EN (1x86) register. The decoded channel ID information is used for auto cropping / strobe function and can also be read through the host i
27、nterface. channel ID detection mode支持 automatic 和 manual（手动）。 For an automatic channel ID detection mode, the playback input should include a run-in clock. But for a manual channel ID detection mode, the playback input can include a run-in clock or not via VBI_RIC_ON (1x88) register. In a manual det
28、ection mode, the TW2835 has several related register such as the VBI_PIXEL_HOS (1x87) to define horizontal start offset, the VBI_FLD_OS (1x88) to define line offset between odd and even field, the VBI_PIXEL_HW to define pulse width for 1 bit data, the VBI_LINE_SIZE (1x89) to define channel ID line s
29、ize and the VBI_LINE_OS (1x89) to define line offset for channel ID. The VBI_MID_VAL (1x8A) register is used to define the threshold level between high and low. Even in automatic channel ID detection mode, the line size and bit width can be discriminated by reading the VBI_LINE_SIZE and VBI_PIXEL_HW
30、 (1xCB) register. The Fig 10 shows the relationship between channel ID and register setting.This channel ID information can be read through the CHID_TYPE or CHID_VALID (1x8B), AUTO_CHID 0/1/2/3 (1x8C 1x8F), DET_CHID 0/1/2/3/4/5/6/7 (1x981x9F), and USER_CHID 0/1/2/3/4/5/6/7 (1x901x97) registers. The
31、CHID_TYPE register discriminates between the Auto channel ID (CHID_TYPE = “1”) and User channel ID (CHID_TYPE = “0”). The CHID_VALID register indicates whether the detected channel ID type is valid or not. The AUTO_CHID, DET_CHID and USER_CHID registers are used to check the decoded channel ID data
32、when the VBI_RD_CTL (1x88) register value is “1”.Basically the channel ID is located in VBI period and auto strobe and cropping is executed after channel ID decoding. But for some case, the channel ID can be placed in vertical active period instead of VBI period. For this mode, the TW2835 also suppo
33、rts the channel ID decoding function within vertical active period via the VAV_CHK (1x89) register and manual cropping function via the MAN_PBCROP (0xC0) register with proper VDELAY value.Cropping and Scaling Function Scaling function 提供在低分辨率下的视频图像当cropping function提高一部分的视频图像。 TW2835也支持自动cropping fu
34、nction 为playback input with channel ID decoding. TW2835有个free scaler 用于variable image size in display path，but has a limitation of image size in record path such as Full / QUAD / CIF format.Cropping Function for LiveThe cropping function allows only subsection of a video image to be output. The acti
35、ve video region is determined by the HDELAY, HACTIVE, VDELAY and VACTIVE (0x02 0x06, 0x12 0x16, 0x22 0x26, 0x32 0x36) register.The first active line is defined by the VDELAY register and the first active pixel is defined by the HDELAY register. The VACTIVE register can be programmed to define the nu
36、mber of active lines in a video fieldthe HACTIVE register can be programmed to define the number of active pixels in a video line. This function is used to implement for panning （平移）and tilt（倾斜）.HDELAY + HACTIVE Total number of pixels per lineWhere the total number of pixels per line is 858 for NTSC
37、 and 864 for PAL全屏大小：the HDELAY should be set to 32 and HACTIVE set to 720 for both NTSC and PAL system.VDELAY + VACTIVE Total number of lines per fieldWhere the total number of lines per field is 262 for NTSC and 312 for PAL全屏大小：NTSCPALVDELAY65VACTIVE240288the VDELAY should be set to 6 and VACTIVE
38、set to 240 for NTSC and the VDELAY should be also set to 5 and VACTIVE set to 288 for PAL.Scaling Function for LiveScaler用于display path.不能用于record path. Because channel size definition for record path has a limitation such as Full / QUAD / CIF (Please refer to “Record Path Control” section, page 64)
39、.亮度水平scaler包括anti-aliasing filter to reduce image artifacts in the resized image via the HSFLT (0x80/90/A0/B0, 0x85/95/A5/B5 and 0x8A/9A/AA/BA) register and a 32 poly-phase filter to accurately interpolate the value of a pixel.垂直scaler also contains an anti-aliasing filter controlled via the VSFLT (
40、0x80/90/A0/B0, 0x85/95/A5/B5 and 0x8A/9A/AA/BA) register and 16 poly-phase filters for down scaling. Down scaling is achieved by programming the scaling register HSCALE and VSCALE (0x81 0x84, 0x91 0x94, 0xA1 0xA4, 0xB1 0xB4) register. When no scaled video image, the TW2835 will output the number of
41、pixels as specified by the HACTIVE and VACTIVE (0x02 0x06, 0x12 0x16, 0x22 0x26, 0x32 0x36) register. If the number of output pixels required is smaller than the number specified by the HACTIVE/VACTIVE register, the 16bit HSCALE/ VSCALE register is used to reduce the output pixels to the desired num
42、ber. The following equation is used to determine the horizontal scaling ratio to be written into the 16bit HSCALE register.HSCALE = Npixel_desired/ HACTIVE * (216 1)(例子在32页)Where Npixel_desired is the desired number of active pixels per lineThe following equation is used to determine the vertical sc
43、aling ratio to be written into the 16bit VSCALE register.VSCALE = Nline_desired / VACTIVE * (216 - 1) (例子在32页)Where Nline_desired is the desired number of active lines per fieldCropping and Scaling Function for Playbackan auto cropping function with channel ID decoding for playback input.If the PB_A
44、UTO_EN (1x16) = “0”, the TW2835 is set to a manual cropping mode so that user can control cropping with VDELAY_PB and HDELAY_PB (0x8B0x8F, 0x9B9F, 0xABAF and 0xBBBF) register. If the PB_AUTO_EN = “1”, the TW2835 is set into an auto cropping mode. In this mode, the desired channel can be chosen by PB
45、_CH_NUM register (1x16, 1x1E, 1x26, 1x2E) and it will be cropped automatically to horizontal and vertical direction in playback input. The TW2835 has several related registers for this mode such as PB_CROP_MD, PB_ACT_MD and MAN_PBCROP (0xC0). The PB_CROP_MD defines the record mode of the playback in
46、put such as normal record mode or DVR record mode (Please refer to “Record Path Control” section, page 64). The PB_ACT_MD defines an active pixel size of horizontal direction such as 720 / 704 / 640 pixels. The MAN_PBCROP controls the horizontal and vertical starting offset in the auto cropping mode
47、 with HDELAY_PB and VDELAY_PB registers. It is useful in case that the encoded channel ID is located at vertical active area in ITU-R BT.656 data stream.The TW2835 includes four additional free down scalars for playback path so that the video image from playback input can be downscaled to an arbitra
48、ry size in both horizontal and vertical direction. Therefore, using this cropping and scaling function, the TW2835 supports free size and positioning function for both live and playback input in display path.Motion Detection The TW2835 supports motion detector individually for 4 analog video inputs.