Amount Of Bits Research Articles

3D adaptive motion vector coding integrated with block‐based adaptive interpolation filter

AbstractSince fractional pixel accuracy motion compensation precisely expresses the motion information, it is widely used for improving coding efficiency in the International Standard video coding methods for the MPEG and H.26x series. When the motion accuracy is ascertained more precisely, motion information becomes more exact, and consequently, prediction efficiency can be improved. The problem, however, is that the amount of bits for expressing large motion increases. Accordingly the adaptive interpolation filter (AIF) method, which changes the filter characteristics for half‐pixel positions on quarter‐pixel motion compensation, was proposed. This technique also changes the image at quarter‐pixel positions and makes the motion accuracy more precise. In conventional studies, the AIF method was used per frame or macroblock that was not the motion compensation block, because the overhead problem of filter indication has never been resolved. In this paper, we propose the 3D motion vector coding technique with the block‐based adaptive interpolation filter. This method integrates filter indication into motion vector coding and predicts the filter indication adaptively, so that it reduces the overhead of filter indications. In this method, the filter is changed per motion compensation block so that the prediction efficiency is improved per motion compensation block. We also demonstrate in simulations that this method achieves higher coding efficiency than the normal H.264 coding method. © 2004 Wiley Periodicals, Inc. Syst Comp Jpn, 35(14): 10–20, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.20125

A novel two-pass VBR coding algorithm for fixed-size storage application

We propose a novel variable bit rate (VBR) algorithm that adopts two-pass encoding to implement high-performance coding for fixed-size storage applications. In the first-pass encoding, we use the original frame as the reference for motion estimation. This way, we are able to obtain the exact relationship between the amount of bits generated for each frame (rate) and each possible quantization factor (Q) so that an R-Q function can be built. The corresponding computational expense is virtually the same as a typical VBR encoder. With this R-Q function, we can optimize the quantization factor for each frame throughout the entire sequence based on the constraints in total storage size and decoder buffer size before the second-pass encoding is executed. This procedure not only guarantees full use of the fixed storage size for the compressed bitstream, but also prevents the decoder buffer from underflowing. Unlike a typical constant bit rate (CBR) decoder, this VBR decoder is capable of preventing the buffer from overflowing by introducing the pausing operation at some designated time. Because the quantization factor optimization is a simple table lookup, the computational requirement is minimized. In the second-pass encoding, the estimated quantization factors for all the frames through the above procedures is used to encode the entire video sequence as a free VBR encoder. Experimental results on real video sequences show that the proposed VBR encoding can provide more consistent visual quality and an improved coding efficiency. The application of the proposed VBR coding scheme includes video streaming over the Internet, digital versatile disk, digital library, and video on demand.

A floating point vectoring algorithm based on fast rotations

In this paper we develop a fast CORDIC-based floating-point vectoring algorithm which reduces significantly the amount of microrotation steps as compared to the conventional algorithm. The overhead required to accomplish this is minimized by the introduction of an angle selection function which considers only a few of the total amount of bits used to represent the vector being rotated. At the same time, the cost of individual microrotations is kept low by the utilization of a fast rotations angle base.

Encoding and decoding apparatus causing no deterioration of sound quality even when sine-wave signal is encoded

There is provided an encoding and decoding apparatus which can provide a high quality reproduction sound. There are provided an encoding section (1, 2, 3, 4, and 5) and a decoding section (7, 8, and 9) . A block length determining section (1) of the encoding section time-sharingly divides an input signal into blocks having a block length corresponding to the degree of change of the input signal. A frequency band dividing section (2) divides the signal of each block into a plurality of frequency bands. A bit allocating section (3) obtains a masking threshold value-to-noise ratio of each frequency band from the magnitude of the power of each frequency band reflecting a psychoacoustic characteristic, and determines the amount of quantization bits to be allocated to each frequency band based on the magnitude of the masking threshold value-to-noise ratio. A quantizing section (4) generates a code string by quantizing the signal while allocating a specified amount of bits to each frequency band based on information from the bit allocating section (3). The bit allocating section (3) receives information representing the block length from the block length determining section (1), and adopts a varied psychoacoustic characteristic according to the block length.

Joint video coding of MPEG-2 video programs for digital broadcasting services

In digital broadcasting services such as digital satellite TV, cable TV, and digital terrestrial TV, several video programs are compressed by MPEG-2 and then simultaneously transmitted over a conventional CBR (constant bit rate) broadcasting channel. In this environment, the picture quality of the aggregated video programs should be kept as equal as possible to provide fair video services for all the viewers. We propose a joint video coding scheme for ensuring that the picture quality of all the programs are nearly the same. This objective is achieved by simultaneously controlling the video encoders to generate the VBR (variable bit rate) compressed video streams. This paper contributes in two ways. First, we propose a rate-distortion estimation method for MPEG-2 video, which enables us to predict the amount of bits and the distortion generated from an encoded picture at a given quantization step size and vice versa. The most attractive features of the proposed estimation method are its accuracy and a computational complexity low enough to be applied to real-time video coding applications. Second, this paper presents an efficient and accurate joint video coding scheme using the rate-distortion estimation results. The experimental results show that our coding scheme gives a higher and more stable picture quality and a more efficient channel utilization than an independent coding scheme that encodes each program independently.

Bit-rate control for MPEG encoders

Bit-rate control is a central problem in designing image sequence compression systems. In this paper we describe a new approach to bit-rate control for inter-frame encoders such as MPEG encoders. This approach uses concepts from control theory. Its central feature is a surprisingly simple but effective model for the encoder, which consists of a gain element, a delay element and additive noise. In our system we control the bit-rate with a PI-controller which is set to achieve two objectives: (1) we want the picture quality to be as uniform as possible, and (2) we want to use as closely as possible the available amount of bits. It is demonstrated in the paper that these two objectives, when considered separately, lead to contradictory settings of the controller. This dilemma can be solved by using Bit Usage Profiles that indicate how the bits have to be spread over the pictures. The effectiveness of the approach is demonstrated by designing a bit-rate control for an MPEG encoder that has a nearly constant bit-rate per group of pictures (GOP). Such a bit-rate control is of high value for applications like magnetic recording, where a constant bit-rate per GOP is required in order to realize playback trick modes, e.g. the fast forward mode.

Classification-Based Motion Estimation for Video Coding

To reach even still higher compression ratios in video sequence coding the demands posed on the motion estimation module become higher and higher. In this paper we will present a motion estimation scheme which gives a motion field defined on every block (typical 8 by 8 or 16 by 16) and resembles the real motion in the scene very well. The proposed algorithm is a "one-pass" algorithm and is not based on any statistical model. A classification procedure on a predefined number of motion vector candidates defines the final motion field. Applications for such a type of motion estimation algorithms can be found in very low bitrate coding where the amount of bits to code the motion field is important compared to the total bitrate, in codecs which want to exploit some features of the human visual system (classification for optimal bit allocation, use of motion masking,…) or in object-based coding where the motion estimation algorithm interacts with the segmentation procedure.

A study on colorimetric errors caused by quantizing color information

When three-dimensional analog signals carrying color information are quantized, quantization errors themselves to do not reflect errors in psychometric color differences because of nonlinear characteristics in the sensitivity of human eyes for colors. The authors convert the quantization errors in the input analog domain to the errors in the uniform color space, the CIE 1976 L*a*b*, so as to evaluate corresponding psychometric quantization errors. The study was conducted for analog signals expressed in the tristimulus values X, Y, Z, and in the primary color signals, R, G, B. The results reveal necessary bits for ADCs as a function of the maximum perceptual color differences between the original and the quantized colors. They also reveal the bit-gain of quantization in the primary color domain against the tristimulus value domain. The study shows that the required minimum amounts of bits for quantization are 12-bit in RGB space and 13-bit in XYZ space to attain quantization color difference of 1.0. >