Yutaka Yokoyama

A 1.5-W single-chip MPEG-2 MP@ML video encoder with low power motion estimation and clocking

A 1.5-W single-chip MPEG-2 MP@ML real-time video encoder large scale integrated circuit (LSI) has been developed. To form an MPEG-2 encoder system, we employ two 16-Mb synchronous DRAMs, a microprocessor unit (MPU), and an audio encoder LSI. Owing to a two-step hierarchical search scheme and a novel adaptive search window scheme, the search range of motion estimation is -48/+47 horizontal and -96/+15.5 vertical, and the pseudo search range, which is the size when the location of the search window is adaptively shifted, is -96/+95 horizontal and -32/+31.5 vertical. We have also developed low-power clocking techniques, i.e., demand-clock controller, local-clock controller, and low-power flip-flops, which can eliminate waste of power in clocking. We have successfully fabricated these new designs as a low-power single-chip MPEG-2 encoder LSI. The operating frequency except for a synchronous DRAM interface unit and a video in/out unit is 54 MHz. The supply voltage to the first and second search engines in a motion estimation unit can be successfully lowered to 2.5 V and the others are 3.3 V. Into a 12.45/spl times/12.45 mm/sup 2/ chip with 0.35-/spl mu/m CMOS and triple-metal layer technology are integrated 3.1 M transistors.

Very low bit rate video coding using arbitrarily shaped region-based motion compensation

This paper presents a novel region-based video coding algorithm for very low bit-rate transmission. A huge amount of region shape information in a conventional region-based coding does not need to be transmitted in the new approach, because it can be synchronously obtained from a reconstructed picture in both the encoder and decoder by image segmentation. Segmented region updating, which can obtain appropriate regions for motion compensated prediction, and a new bit assignment algorithm in prediction error coding are also developed. Computer simulations are carried out to evaluate the coding efficiency at 9.6 kb/s. The result indicates that the objective and subjective qualities for decoded pictures are improved compared to those for the conventional block-based method. This algorithm is significantly effective for mobile communication.

Adaptive GOP structure selection for real-time MPEG-2 video encoding

This paper proposes an adaptive real-time group of pictures (GOP) structure selection method for real-time MPEG-2 video coding. This method changes the interval of P pictures dynamically in the the 3-picture period by using coding complexity with single pass encoding. Experimental results show that the prediction error is reduced to 88% /spl sim/ 95% compared with the fixed GOP structure and the quality of inter-frame prediction pictures is improved. The proposed method selects the same GOP structure as that of an optimal method (multiple-pass encoding) in the time periods by over 93%. The method gives a close performance to the optimal encoding. It can be applied to real-time video recoding.

A scene-adaptive one-pass variable bit rate video coding method for storage media

This paper proposes a one-pass VBR (Variable Bit Rate) video coding method that maintains the stability of its quantization scale by using a large size virtual buffer to smooth out the difference between a generated bit count and a target bit count. To improve video quality, the quantization scale is adjusted according to the coding complexity of scenes. Experimental results show that the standard deviations of quantization scale distributions are only 30/spl sim/40% of those produced by CBR (Constant Bit Rate) coding, where averaged quantization scale values are roughly the same as those of CBR. The worst quality periods defined as the duration that the quantization scale exceeds a certain level are reduced from 28.7% to 7.8% in total coding time at 4 Mbps coding. The method can be applied to real-time video recording applications.

An MPEG-2 encoder architecture based on a single-chip dedicated LSI with a control MPU

This paper describes an MPEG-2 encoder architecture based on a hard-wired LSI with a control MPU. All basic functions of MPEG-2 MP@ML video compression are integrated in the dedicated LSI. For the motion estimation, a horizontally subsampled, diamond search was employed as a simplified first search step. It can reduce operations to 20% of the full-search, with an estimated SNR degradation of only -0.1 dB. To help achieve a single-memory interface, a pair of 81 MHz, 16 Mb SDRAMs are used as a frame buffer and a code buffer. Data bandwidth between the SDRAMs and the LSI is kept to less than 94% of the maximum data rate. Jobs assigned to the control MPU need be executed less frequently than those of the macroblock coding, which helps reduce the requirements for MPU performance to about 7 MIPS.

Very low bit-rate video coding with object-based motion compensation and orthogonal transform

This paper presents a new object-based video coding algorithm for very low bit-rate video transmission. The algorithm employs an object-based approach for motion compensation, so that block distortion does not appear. The moving objects are detected from local decoded pictures, so information regarding object contours is not sent to a decoder. This technique contributes to drastically reducing the code amount to be transmitted. Motion compensated prediction errors are encoded by transform coding, but only large error regions are encoded. Computer simulation results at 16 kbps show that the subjective quality for a decoded picture by this algorithm is substantially better than that of a conventional algorithm.

System for coding and decoding moving pictures based on the result of speech analysis

A speech and moving picture coding and decoding system wherein pictures which are close to real pictures and look more natural than ever are displayed by a simple method without the necessity of complicated processing for picture analysis or picture synthesis and pictures which look like pictures in a real speaking condition can be decoded and reproduced with high picture quality even where the transmission capacity is restricted low is disclosed. On the transmission side, input speech is coded while it is simultaneously analyzed by a speech analysis section, and an input picture corresponding to a result of the speech recognition is coded by a picture coding section and transmitted. On the reception side, the coded picture is decoded and stored into a picture memory. Then, the coded speech received is decoded by a speech decoding section, and one of stored pictures corresponding to the received result of speech recognition is read out from the picture memory.

A rate control method with preanalysis for real-time MPEG-2 video coding

This paper proposes a bit rate control method with a preanalysis for real-time MPEG-2 video coding. The rate control method allocates bits to pictures according to picture coding complexities obtained from the preanalysis. The bit allocation is adjusted to satisfy the VBV constraints. The method is evaluated by MPEG-2 video encoding experiments for sequences whose characteristics significantly change, such as scene changes or fade scenes. The experimental results show that the picture quality is improved by the proposed control method. The PSNR is 2 to 3 dB higher than the conventional method without a preanalysis in difficult scene periods. The method is very effective when a certain coding delay is allowed.