Click here to load reader
Click here to load reader
Dec 28, 2015
Video Encoding in tools
Digital VideoStage I Sampling VideoConverting real life video into a sequence of static images - Frames
Stage II Sampling PicturesFrame rate: The video sampling speed, or the number of frames per seconds (FPS) used to represent the videoStandards Frame Rates:Cinema Film24 FPS PAL TV 25 FPS (Europe) , NTSC TV30 FPS (USA/Japan), More frames = Better quality (smoother motion)
Stage II Sampling PicturesConverting analog pictures to digital picture.
Stage II Sampling PicturesColor - Colors are represented digitally as unique numeric combination of base colors RGB (Red, Green and Blue). For example: (0,0,255) = blue (255,0,255) = purple (255,255,255) = white (0,0,0) = black
Color depth - The number of bits used to describe the color range
Stage II Sampling PicturesExamples:More colors = better quality
Stage II Sampling PicturesPixel - the smallest element in a digital image, holds the digital color representation of a specific image locationExamples:Standard TV 640 X 480 (4:3 ratio) = 307,200 pixelsFull Screen 800 X 600 (4:3 ratio) = 480,000 pixelsHDTV1920X1080 / 1280X720 (16:9 ratio) ~ 1-2M pixels3M Digital Camera2048X1536 (4:3 ratio) ~ 3M pixelsMore pixels = more details = better qualityResolution The number of pixels used in each dimension (Width X Height) to represent the picture
Digital video - Size issues IDigital Video File Size = Video duration (seconds) X Frame Rate (FPS) X Resolution (Height X Width) X Pixel size (Color depth)Example:30 seconds X 30 FPS X (320X240) X 24bit = Without including audio! 207,360,000 Bytes = 207MB !!!
Digital video Size issues IIVideo Bit Rate How much data (bits) is used to store one second of video and therefore what would be the minimum dedicated bandwidth requirement that would guarantee smooth video display in streamingVideo Bit rate = (Video file size / video duration) = Frame Rate (FPS) X Resolution (Height X Width) X Pixel size (Color depth)In our example: 30 FPS X (320X240) X 24bit =~ 55Mbit/sec (vs. 1 to 5Mbit/sec for Broadband connection)
Size issues - SummaryStorage Storing significant amounts of uncompressed digital video content is not only expensive it is unpractical. A CD-R would store less than 2 minutes Remote access Even in LAN conditions bandwidth is limited to 100Mbit/Sec. Serving uncompressed video over the internet would be out of the question Solution: Compression
Compressing videoType of compressions"Lossless" compression no data is lost, content can be restored into its original decompressed format perfectly. Mostly used on documents and other textual data (for example ZIP, but also GIF). RLE (Run Length Encoding) encodes/stores sequences of data as a single value-count pairs.This is most useful on data containing many long sequences (e.g. simple graphic images, such as icons and line drawings). For example: ABBBBBBBBBCDEEEEF = A *8B C D *4E FHuffman coding - encodes often-repeated symbols with a few bits and rare ones with more bitsLZW (Lempel-Ziv-Welch) Similar to Huffman but with group of symbols (e.g. words)
Compressing videoLossy compression some data is lost, hoping that it is insignificant and would not be noticed (quality). Mostly used on visual data / sound such as images , music and video (for example JPEG, MPEG, MP3 etc). Type of compressionsDCT (Discrete Cosine Transformation) & Vector quantization are techniques that are used to eliminate insignificant information that is not expected to be detected by the human eye
Compressing videoVideo Compression - Reduction of the size of files containing video images that are stored in digital form Encoding The process of converting data from one format to another. Video encoding is used when analog video is converted to digital video and then again when digital video is compressed Decoding The process of restoring the original format of Encoded data. Codec - Software or Hardware technology for encoding and decoding of digital video (short for compressor / decompressor or coder / decoder)
Compressing videoMicro blocks Any frame is divided into small micro blocks which are small matrixes of bits presenting part of the frame. Most of the compression techniques are working on those blocks.
Some video compression algorithm types:Motion detection - Algorithm to detect movement of micro blocks from one position on certain frame to different position on the next frame.Motion vector Key matrix of bits by which certain micro block in a current frame can be predicted by a micro block from previous frame. In some cases, the old position of the predicted micro block should be stored within the same vector.The efficiency of these algorithms is decreasing with the level of changes between frames
Decoding Workflow:Each Micro block is decompressed in the exact opposite process to the compressionI (Intra) frames are Key frames were compressed directly from a real source frameP (Predicted) frames are created from an I frame + a motion vectorB (bi-directionally interpolated) frames are artificially created to smooth the motion and increase the frame rate
Compressing Data Rate ControlQuality vs. Bit rate Higher quality needs higher bit rate, and lower bit rate produces lower quality.2 different data rate control: Bit rate control limiting the maximum bit in any given time. The encoder can not exceed this value. Pros: No big varies in bit rate during the entire videoCons: Video quality can decrease in certain sections (fast changing complex scenes) Quality limited limiting the minimum quality for each frame (percentage). The encoder can not decrease below this value. Pros: Same quality during the entire videoCons: Video bit rate can have very high peaks
Compressing Data Rate ControlCBR (Constant Bit Rate encoding) - Is not really constant ! (there is fixed bit rate encoding such as DV). The CBR means that the average bit rate of the video in a certain time interval (e.g. 5 seconds), can not exceed the requested bit rate.
Compressing Data Rate ControlVBR (Variable Bit Rate encoding) - There is a lot of confusion about this definition, but it basically means:Either the encoder needs to maintain an average bit rate on the whole video.The encoder needs to maintain a minimum quality for each frame.
Compressing Encoding passesThe number of times the encoder needs to go over the input data. 1 pass encoding The encoder encode the video compressed file during one read process of the source file. This method is fast and used mainly for real time encoding (Internet broadcast, video conferencing, etc..) 2 pass encoding The encoder read the source file twice. On the first read it gathers information about the movements and best key frames in order to optimize the compression during the 2nd read of the source.
Digital AudioStage I Sampling AudioConverting analog audio to a series of digital samples.Sample rate The number of samples per seconds (Hz). Influence the frequency range that can be compressed.More samples = Better sound Standards:8,000 Hz - speech quality (telephone) 11,025 Hz22,050 Hz radio quality (minimum for music)44,100 Hz CD quality (commonly used with audio in VCD & MP3) 48,000 Hz Digital TV 96,000 or 192,400 Hz DVD Audio
Digital AudioStage II Storing samplesSample Bit size The number of bits to represent a single sample. 16 bit is the commonly used size.
Channels The number of audio channels (1-mono, 2-stereo and more for high quality like in DVD).
Bit rate The number of bits used to store a second of audio data.
Bit rate control - Similar to video (bit rate / quality)
Video Encoding ToolsFlash Encoding: On2 Flix 8 ProAdobe Flash 8 Video EncoderSorenson Squeeze
WMV Encoding: Window Media Encoder 9
Raw Changes: Adobe Premier Pro