Introduction to Video Transcoding Of MCLAB Seminar Series

Introduction to Video Transcoding Of MCLAB Seminar Series By Felix

Content w Introduction w MPEG Coding Basics w Transcoding Algorithms w Summary

Introduction

Introduction w Heterogeneous Networks in Vo. D n Different Network Bandwidths w Heterogeneous Client Configurations n n Different Screen Resolutions, CPU speed, etc. Different Decoders (eg. MPEG 1, 2, H. 263, Quicktime… etc. )

Introduction w Solutions n Multiple Copies Simple and Fast l Waste Storage l n Re-encoding Only one high-quality video is stored l Complex motion estimation is required l

Introduction Cont’ n n Use video formats with bit-rate scalability (e. g. MPEG 4) Transcoding l Convert a compressed video into another compressed video with lower bit rate or in other formats.

Introduction w Transcoding Drop away data selectively Input Movie n n n Partial Decode Transcode Partial Encode Output Movie Only one high quality compressed video is stored No/Much less computations on motion estimation Can produce comparable video quality with direct encoding

MPEG Coding Basics w A MPEG stream consists of 3 types of frames n n n I (Intra-coded) P (Predictive-coded) B (Bidirectional-predictive coded)

MPEG Coding Basics w I Frames n The original picture represented can be reconstructed without reference to other frames w P Frames n The original picture has to be reconstructed with reference to the past constructed I or P frames w B Frames n The original picture has to be reconstructed with reference to the past and future constructed I or P frames

MPEG Coding Basics w Macroblocks Frame Macroblock 16 16 Y = 8 8 8 0 2 1 3 8 Cr 4 Cb 5 8 8

MPEG Coding Basics w Coding of Macroblocks n 5 types of coding mode Intra-coded l Forward Predictive-coded l Backward Predictive-coded l Bidirectional Predictive-coded l Skipped l

MPEG Coding Basics n Types of macroblocks in I, P and B frames Intracoded Forward Predictive-coded I X P X X Backward Predictivecoded Bidirectional Predictive-coded Skipped X X

MPEG Coding Basics n Predictive-coding Reference frame Current frame Motion Vector Actual MB Ref. MB - Prediction Residue =

MPEG Coding Basics n Encoding of a 8 x 8 block DCT Quantization Run Length Coding Huffman Coding

Transcoding Algorithms w Requantization n n Some DCT coefficients become zero after coarser quantization Open-loop Requantization Input Stream Q 1 -1 Q 2 Output Stream

Transcoding Algorithms n Drifting effect of open-loop requantization Requantization error accumulates along the frames l Eg. l Prediction Residue Original: Reference MB Actual MB + = After Requantization:

Transcoding Algorithms n PSNR drops along the frames

Transcoding Algorithms n Spatial-domain drift-free Requantization Extracted from: “A Frequency-Domain video Transcoder for Dynamic Bit-Rate Reduction of MPEG-2 Bit Streams”

Transcoding Algorithms w Frequency-domain drift-free Requantization Accumulated error Extracted from: “A Frequency-Domain video Transcoder for Dynamic Bit-Rate Reduction of MPEG-2 Bit Streams”

Transcoding Algorithms w MC-DCT (Motion Compensation in DCT) A 0 h 8 -h A 1 w 8 -w A 2 Aref Motion Vector Current Block A 3 In the reference frame (DCT) In the current frame (DCT)

Transcoding Algorithms w Limitations of requantization n n Limited bit rate reduction The clients may have lower screen resolution than the original video, thus the video bit rate can be further reduced by reducing the video resolution

Transcoding Algorithms w Spatial downscaling n n n Compressed movie is downscaled spatially in a multiple of 2 in both horizontal and vertical directions Four macroblocks are combined and downscaled to the size of one macroblock Much more complicated than requantization

Transcoding Algorithms General Block Diagram of Spatial Downscaling Buffer Frames Reconstruction Downscaling Reconstruction of Motion Vectors Original frame Reconstruction of Prediction Residual downscaled frame Reconstruction of macroblocks

Transcoding Algorithms w Problems to be solved in spatial downscaling n n n Macroblock coding mode decision Motion vector reconstruction Downscaling

Transcoding Algorithms w Macroblock Coding Mode Decision Forward Predictive Coded Intra Coded ? + Backward Predictive Coded Bidirectional Predictive Coded

Transcoding Algorithms Cont’ n n Very few researches worked specifically on this problem Considerations l Target bit rate w Intra-coding mode requires more bits l Speed w Bidirectional predictive coding is the most computational expensive w Intra coding is the least computational expensive l Quality w Intra coding introduces no drift

Transcoding Algorithms w Motion Vector Reconstruction

Transcoding Algorithms w Align To Average n n One common method is align-to-average(AAW) AAW gives poor results when the four motion vectors are not wellaligned

Transcoding Algorithms w Adaptive Motion Vector Resampling (AMVR) n Ai is the weighting associated with motion vector Vi Extracted from: “Adaptive Motion-Vector Resampling for Compressed Video Downscaling” By Bo Shen, Ishwar K. Sethi, and Bhaskaran Vasudev

Transcoding Algorithms Cont’ n n The resulting motion vector should skew closer to the motion vector that yielded poor prediction (larger Ai), this is so called align-to-worst (AWW) Ai is calculated as the number of nonzero AC coefficients (DCT coefficients at positions other then (0, 0))

Transcoding Algorithms w Predictive Motion Estimation (PME) n n Similar to full-search motion estimation, however, it just compares the Mean Absolute Difference (MAD) obtained from 4 candidate motion vectors (Vi/2). If none of the MAD obtained is zero, then a new candidate motion vector is computed with the equation proposed by AMVR where Ai=1/MADi Extracted from: “Predictive Motion Estimation for Reduced-Resolution Video from High-resolution Compressed Video” By Justy W. C. Wong, Oscar C. Au, Peter H. W. Wong and A. Tourapis

Transcoding Algorithms Cont’ n n Then choose the candidate vector with the minimum MAD value as the new motion vector PME produces higher quality than AMVR but requires much more computations

Transcoding Algorithms w Downscaling n n Given a picture in DCT domain, how to downscale the spatial resolution of it? Downscale in pixel domain Original DCT Picture IDCT Original Pixel Picture Averaging Every 4 Neighboring pixels Downscaled Pixel Picture DCT Downscaled DCT Picture

Transcoding Algorithms n Downscaling in DCT domain S T X A 1 A 2 A 3 A 4 As 0 0 0 ST X TT Extracted from: “A fast approximate algorithm for scaling down digital images in the DCT domain” by B. Natarajan and V. Bhaskaran

Transcoding Algorithms w Temporal Downscaling n n In typical movies, the difference between consecutive pictures is usually small, so some frames can be safely skipped without any significant quality degradation The decoders have to repeat the previous nonskipped picture to maintain the same playback speed

Transcoding Algorithms w B frames can be skipped without affect other frames w Skipping P or I frames is much more complicated, as this affects the subsequent P and B frames

Transcoding Algorithms w Finding the new best-matched MB ? Ft-2 Ft-1 (Skipped) Ft

Transcoding Algorithms w Motion Vector Reconstruction for temporal downscaling =? MVt’ = MVt + MVt-1

Transcoding Algorithms Cont’ n n MVt-1 can computed by adopting the AMVR equation with Ai being the overlapping area of MBt-1 on MBi After finding the MBt-2 , we can reconstruct the prediction residue of MBt

Transcoding Algorithms w Which method is the best? n No absolute answer because l l Image complexity and motion activity varies along the video streams Depends on target bit rate w Content-based hybrid transcoding n Use different transcoding algorithms in different parts of a video stream according to the content

Summary w Video transcoding is a good solution to provide an adaptive Vo. D service in a heterogeneous environment because of its n n n Low storage requirement Low computational complexity Comparable quality with direct encoding w Requantization, spatial and temporal downscaling are the most common transcoding algorithms w Content-based Hybrid Transcoding combines different algorithms to improve the overall transcoding performance

Thank You