Multimedia Software Systems CS4551
Video Compression – ISO Standards (MPEG)
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO Standards
Copyright By PowCoder代写 加微信 powcoder
• MPEG-1(moviesonCD-ROM)
• MPEG-2(digitaltelevision,moviesonDVD)
• MPEG-4(stillanevolvingstandardandisdividedinto a number of parts)
– MPEG-4 Part2 used by DivX, Xvid, , 3ivx and Quick Time
– MPEG-4 Part10 (MPEG-4 AVC/H.264) used by QuickTime 7, HD Video media like Blue-ray Disc.
– MPEG-H Part 2 (MPEG-H HEVC/H.265)
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-1 Video
• MPEG-1: true multimedia standard with specifications for coding and transmission of audio, video and data streams in a series of synchronized, mixed packets
• Driving focus: storage of multimedia content on digital storage media (e.g. CDs or VCDs)
• Quality: VHS VCR-like video quality and CD audio quality
• Picture format: SIF (Source Input Format)
– 352×288, 25 fps for PAL (Phase Alternating Line) video
– 354×240, 30 fps for NTSC (National Television System Committee) video
– 4:2:0 Chroma sampling
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
Motion Compensation in MPEG-1
• Coding mechanism similar to H.26x
• Three types of frames:
– I-frames (coded in intra-frame mode)
– P-frames (coded with motion compensation using as reference a previous I or P frame)
– B-frames (coded with bidirectional motion compensation based on a previous and a later I or P frame)
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
Need for Bidirectional Search
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
Motion Compensation in MPEG-1
• For B-frame encoding/decoding, it needs succeeding P- or I- frame, the actual coding and transmission order is different from the display order. => Inevitable delay and need for buffering.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
Display order
Coding and transmission order
Acquisition/Transmission Order
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
Example Question
• Suppose we encode a sequence of 23 frames using MPEG. Each frame is encoded in I (intra frame), P (forward prediction) or B (bi-directional prediction) mode according to the following order: IPPPBBPIPPPIPPBBBBPIPIP
– Derive the correct transmission and decoding order. Answer) 1 2 3 4 7 5 6 8 9 10 11 12 13 14 19 15 16 17 18 20 21 22 23 with respect to the indices of the original sequence starting from 1.
– What is the minimum delay (in terms of frame period) due to bi-directional encoding in this case?
Answer) 4 frames. The 15th frame in the sequence (a B- frame) cannot be encoded until the 19th frame in the sequence has been encoded.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-1 Video
• Differences from H.261
– Quantization (allow non-constant quantization interval for the block)
– Half pixel accuracy
– Supports larger gaps between I and P frames
– Bitstream format allows random access
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-1 Video • QuantizationinMPEG-1
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-1 Video • MPEG-1 Video Bitstream
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-2
• MPEG-2 is for high-quality video at a bit rate of more than 4Mbps. It was originally designed as a standard for digital broadcast TV.
• Overtime,theMPEG-2videostandardwasexpanded to include high-resolution video (such as HDTV), as well as hierarchical (or scalable or layered) video coding.
• MPEG-2hasgainedwideacceptancebeyond broadcasting digital TV over satellite or cable networks.
• It is also adopted for DVDs. CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-2
• MPEG-2defined7profilesaimedatdifferent applications such as low-delay video conferencing, scalable video, or HDTV.
• 7 profiles: Simple, Main, SNR Scalable, Spatially Scalable, High, 4:2:2 and Multiview.
• Within each profile, up to four levels are defined.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
ISO MPEG-2
• DVD video specification uses only a restricted form of the MPEG-2 Main profile at the Main and Low levels
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-1 vs. MPEG-2
• Compression based on DCT-based transform coding and motion- prediction.
• Differences from MPEG-1
– Support of interlaced video whereas MPEG-1 supports only non-interlaced (progressive) video. MPEG-2 is adopted by digital broadcast TV such as HDTV, it must support interlaced video.
– Scalability – make a single video stream scalable to various bitrates.
– Support of 4:2:2 and 4:4:4 chroma sampling
– More flexible video formats –…
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Scalability
• The coded representation (bit-stream) is generated in such a way that decoders of various complexities are able to decode video of different resolution/quality from the same bit stream
– If the bitstream is truly scalable, decoders of different complexities can coexist: inexpensive decoders would be expected to decode only small portions of the same bitstream producing basic quality pictures, while more sophisticated and expensive decoders will produce higher quality pictures
• For applications in networks with multi-quality video services and windowed video on computer workstations.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Standard Non-Scalable Encoder
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Scalabilities
• MPEG-2scalablecodingisalsoknownaslayered coding.
• Input video goes through a pre-processor producing a base layer signal and one or more enhancement layers.
– Base layer is decoded independently by a standard MPEG-2 non-scalable video encoder with a coarse quantization, which result in fewer bits and a relatively low-quality video.
– Enhancement layer is encoded with respect to the base layer by a MPEG-2 enhancement encoder.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
2 Layer Scalable MPEG-2 Video Encoder
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Scalability
• MPEG-2 supports the following scalabilities
– SNR scalability
– Spatial Scalability (like pyramid)
– Temporal Scalability (higher frame rate)
– Hybrid scalability (combines any two of the above three)
– Data partitioning (like Spectral selection)
• Scalability is very useful for
– Very different bitrates: different network link speeds (fast/slow)
– Variable-bit rate channel: bitrate of the channel deteriorates sometime
– Noisy connections: send the base layer via a less-noisy channel
– Progressive transmission: coarse to finer image
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – SNR Scalability
• SNRscalableencodergeneratestwolayers – Base layer bit stream called Bits_base
• Coarse quantization of the DCT coefficients is employed. -> Fewer bits
– Enhancement layer bit stream called Bits_enhance
• Generate a DCT coefficient refinement by computing quantization
• Quantization error = original DCT – de-quantization (quantized DCT in the base layer)
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
CSULA CS451 Multimedia Software Systems by Eun-Young Kang
SNR Scalability
CSULA CS451 Multimedia Software Systems by Eun-Young Kang
SNR Scalability
MPEG-2 – Spatial Scalability
• SpatialScalability
– Base layer
• Generateabitstreamofreduced-resolutionframes
– Enhancement layer(s)
• Generate a bitstream of the details of higher resolution frames
– Frames at the original resolution is generated by combining two layers.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Spatial Scalability
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Temporal Scalability
• Temporal Scalability (a) – Both the base layer and enhancement layer are generated at a reduced frame rate with the same spatial resolution as in the input video. To restore the original video, it combines two layers.
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Temporal Scalability
• Temporal Scalability (b) – The prediction of matching macroblocks at the
enhancement layer by interlayer motion-compensated prediction
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Temporal Scalability
• Temporal Scalability (c) – The prediction of matching macroblocks at the enhancement layer by combined motion-compensation prediction and interlayer motion compensation prediction
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
MPEG-2 – Other Scalability
• Hybrid scalability (combines any two of the above three)
– SNR and Spatial
– Spatial and temporal – SNR and temporal
• Datapartitioning(similartoSpectralselection)
– Base layer contains lower frequency DCT coefficients
– Enhancement layers contain higher frequency DCT coefficients
CSULA CS4551 Multimedia Software Systems by Eun-Young Kang
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com