Coded data generation or conversion – Digital code to digital code converters – Adaptive coding
Patent
1996-01-26
1998-11-10
Hoff, Marc S.
Coded data generation or conversion
Digital code to digital code converters
Adaptive coding
341106, H03M 703, G11B 2010
Patent
active
058350306
DESCRIPTION:
BRIEF SUMMARY
DESCRIPTION
1. Technical Field
This invention relates to signal encoding and decoding and, in particular, to encoding and decoding efficiently and to achieve a desired bit rate.
2. Background Art
Transform encoding using "spectrum transform processing" is known. Spectrum transform processing is an efficient encoding for carrying out bit compression of time series sample data signals (such as digital audio signals, etc.) with high efficiency.
Specifically, Spectrum transform processing includes dividing an input signal into blocks, to implement spectrum transform processing in respective block units. Discrete Cosine Transform (DCT) processing is the representative spectrum transform processing of the transform encoding.
In such a transform encoding, there is a problem with block distortion such that discontinuous connecting (joining) portions between blocks are perceived as noise. In order to lessen such a block distortion, a method of allowing end portions of a block to overlap with adjacent blocks is generally carried out.
That is, MDCT (Modified DCT or Improved DCT) is such that double transmission with respect to samples at overlapping portions is not carried out. An arbitrary block is allowed to overlap with blocks adjoining in both directions thereof respectively by halves (half blocks). The arbitrary block is therefore suitable for efficient encoding.
Encoding and decoding using such MDCT (and IMDCT, which is the inverse transform processing thereof) is disclosed in, e.g., Mochizuki, Yano, Nishitani "Filter Constraints of Plural Block Size MDCT", Technical Research Report of Institute of Electronics and Communication Engineers of Japan, CAS90-10, DSP 90-14, pp. 55-60, and Hazu, Sugiyama, Iwatare, Nishitani "Adaptive Block Length Adaptive Transform Coding using MDCT (ATC-ABS)", Society of Electronics and Information Communication Engineers of Japan Spring National Meeting Lecture Collection of 1990, A-197.
The processing procedure using MDCT and IMDCT discussed above will be briefly described below with reference to FIG. 12. In FIG. 12, an arbitrary block of time series sample data, e.g., the J-th block has a 50% portion overlapping with the (J-1)-th block and a 50% portion overlapping with the (J+1)-th block. When the number of samples of the J-th block is assumed to be N (N is natural number), the J-th block has overlap of N/2 samples between the J-th block and the (J-1)-th block, and also has overlap of N/2 samples between the J-th block and the (J+1)-th block. Pre-processing filter window Wh is applied to these respective blocks, e.g., an arbitrary J-th block input time series sample dO1 to obtain N time series data dO2.
A characteristic of the pre-processing filter transform window Wh, is selected in correspondence with the statistical property of an input signal such that the degree of power concentration of transform data is maximum. By implementing linear transform processing of MDCT an the time series data dO2 of N samples, N/2 (number of), which is one half of the number of input samples, independent spectrum data dO3 indicating levels of spectrum components are obtained on the frequency base. By implementing processing of linear inverse transform of IMDCT to the N/2 (number of) spectrum data dO3, N (number of) time series data dO4 are obtained. Synthesis filter (window Wf for inverse transform) is applied to the time series data dO4 to obtain time series data dO5, which is thereafter added to output results of blocks before and after to restore (reconstruct) original input time series sample data.
In the conventional efficient encoding, there has been adopted a method of dividing spectrum data dO3 obtained in a manner as described above into several units, every band, to carry out normalization. Quantization is conducted by taking the characteristic from a viewpoint of the hearing sense into consideration to carry out recording or transmission along with normalization coefficients of respective units.
In addition, as in the ISO11172-3 of the ISO standard, entropy encoding is implemented su
REFERENCES:
patent: 4485337 (1984-11-01), Sandusky
patent: 4669120 (1987-05-01), Ono
patent: 4908810 (1990-03-01), Oie
patent: 4972484 (1990-11-01), Theile
patent: 5040217 (1991-08-01), Brandenburg
patent: 5045812 (1991-09-01), Tateishi
patent: 5063550 (1991-11-01), Watari et al.
patent: 5109417 (1992-04-01), Fielder
patent: 5204677 (1993-04-01), Akagiri et al.
patent: 5223926 (1993-06-01), Stone et al.
patent: 5301205 (1994-04-01), Tsutsui et al.
patent: 5311561 (1994-05-01), Akagiri
patent: 5365530 (1994-11-01), Yoshida
patent: 5440404 (1995-08-01), Okamoto
T. Mochizuki et al., "Constraint Conditions for Multiple-Blocksize Modified-DCT," Technical Research Report of the Institute of Electronics and Communication Engineers of Japan, CAS90-10, DSP 90-14, pp. 55-60.
F. Hazu et al., "Adaptive Block Length Adaptive Transform Coding Using MDCT (ATC-ABS)," Society of Electronics and Information Communication Engineers of Japan, Spring National Meeting Lecture Collection of 1990, A-197.
Shimoyoshi Osamu
Sonohara Mito
Tsutsui Kyoya
Hoff Marc S.
Jean-Pierre Peguy
Sony Corporation
LandOfFree
Signal encoding method and apparatus using selected predetermine does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Signal encoding method and apparatus using selected predetermine, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Signal encoding method and apparatus using selected predetermine will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1520689