Data processing: speech signal processing – linguistics – language – Speech signal processing – Psychoacoustic
Reexamination Certificate
1999-06-03
2002-07-02
{haeck over (S)}mits, T{overscore (a)}livaldis Ivars (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Psychoacoustic
C704S205000, C704S206000
Reexamination Certificate
active
06415251
ABSTRACT:
TECHNICAL FIELD
This invention relates to an information decoding method and device, an information coding method and device, and a providing medium. It particularly relates to an information decoding method and device, an information coding method and device, and a providing medium for restraining output of an unpleasant sound by erasing an aliasing component with respect to a code string formed by coding only a signal of a partial frequency band of acoustic waveform signals.
BACKGROUND ART
Conventionally, there are various methods and devices for high efficiency coding of audio or sound signals. For example, such methods and devices can be exemplified by a transform coding system which is adapted for forming frames of signals in the time domain, then converting (spectral conversion) each frame of signals in the time domain to signals in the frequency domain, splitting the signals into a plurality of frequency bands and coding each band of signals, and a so-called subband coding (SBC) system which is adapted for splitting audio signals in the time domain into a plurality of frequency bands and coding each band of signals, without forming frames of audio signals. Also, there is considered a method and device for high efficiency coding using the above-described subband coding in combination with transform coding. In this case, after band splitting is carried out by the subband coding system, each band of signals are spectrally converted to signals in the frequency domain, and coding is carried out on each spectrally converted band.
As a band splitting filter used in the above-described subband coding system, there is employed, for example, a polyphase quadrature filter (PQF), which is described in Joseph H. Rothweiler, “Polyphase Quadrature Filters—A new subband coding technique,” ICASSP 83, BOSTON. This PQF can split a signal into a plurality of bands of equal widths at a time and is characterized in that so-called aliasing is not generated in synthesizing the split bands later.
As the above-described spectral conversion, there is employed spectral conversion for forming frames of input audio signals of predetermined duration and carrying out a discrete Fourier transform (DFT), discrete cosine transform (DCT) or modified discrete cosine transform (MDCT) on each frame so as to convert the time domain to the frequency domain, MDCT is described in J. P. Princen and A. B. Bradley, “Subband/Transform Coding Using Filter Bank Designs Based on Time Domain Aliasing Cancellation,” Univ. of Surrey Royal Melbourne Inst. of Tech. ICASSP 1987.
By thus using the filter and spectral conversion to quantize the signals split for each band, the band in which quantization noise is generated can be controlled and coding can be carried out at an auditorily higher efficiency using characteristics such as a so-called masking effect. Also, by normalizing the maximum value of absolute values of signal components for each band before carrying out quantization, coding can be carried out at a higher efficiency.
As a frequency splitting width in quantizing each frequency component (hereinafter referred to as a spectral component) split into frequency bands determined by human auditory characteristics is often employed. Specifically, critical bands whose bandwidths increase as the frequency becomes higher are used to split audio signals into a plurality of bands (for example, 25 bands). In coding each band of data at this point, coding is carried out by using predetermined bit distribution for each band or adaptive bit allocation for each band. For example, in coding coefficient data obtained by MDCT processing by using the foregoing bit allocation, coding is carried out by using an adaptive number of allocated bits with respect to each band of MDCT coefficient data obtained by MDCT processing on each frame.
The following two methods are known as the bit allocation method.
For example, in R. Zelinski and P. Noll, “Adaptive Transform Coding of Speech Signals,” IEEE Transactions of Acoustics, Speech, and Signal Processing, Vol.ASSP-25, No.4, August 1977, bit allocation is carried out on the basis of the magnitude of signals of each band. In this method, the quantization noise spectrum becomes flat and the noise energy is minimized. However, since the auditory masking effect is not used, the actual auditory perception of noise is not optimum.
On the other hand, for example, in M. A. Kransner, “The critical band coder—digital encoding of the perceptual requirements of the auditory system,” MIT, ICASSP 1980, there is described a method for obtaining a signal-to-noise ratio required for each band by utilizing auditory masking so as to carry out fixed bit allocation. However, in this method, since bit allocation is fixed, a satisfactory characteristic value cannot be obtained even in the case where characteristics are measured by using a sine wave input.
To solve these problems, a high efficiency coding device is proposed in which all the bits that can be used for bit allocation are divided into bits for a fixed allocation pattern predetermined for each band or each block obtained by subdividing each band and bits for bit distribution dependent on the magnitude of signal frequency components in each subband, and in which the division ratio is caused to depend on signals related to input signals to that the division ratio for the fixed bit allocation pattern is increased as the spectral distribution of the signals becomes smoother.
According to this method, in the case where the energy is concentrated on a specified spectral component as in the case of a sine wave input, the overall signal-to-noise characteristic can be significantly improved by allocating a greater number of bits to a block including that spectral component. In general, the human auditory sense is extremely acute with respect to signals having steep spectral distribution. Therefore, improvement of the signal-to-noise characteristic by using such method is effective not only for improving the numerical value of measurement but also for improving the auditorily perceived sound quality.
In addition to the foregoing method, various other methods are proposed as the bit allocation method. If a model related to the auditory sense is made fine to improve the capability of the information coding device, coding can be carried out at an auditorily higher efficiency.
In the case where the above-described DFT or DCT is used as the method for spectral conversion of waveform signals consisting of waveform elements (sample data) such as digital audio signals of the time domain, blocks are formed by every M units of sample data and spectral conversion of DFT or DCT is carried out on each block. By carrying out spectral conversion on such blocks, M units of independent real number data (DFT coefficient data or DCT coefficient data) are obtained. The M units of real number data thus obtained are quantized and coded, thus generating coded data.
In decoding the coded data to reproduce regenerative waveform signals, the coded data are decoded and inversely quantized, and inverse spectral conversion by inverse DFT or inverse DCT is carried out on each block of the resultant real number data corresponding to the block at the time of coding, thus generating waveform element signals. Then, blocks consisting of the waveform element signals are connected to reproduce waveform signals.
In the regenerative waveform signals thus obtained, a connection distortion in connecting the blocks remains, which is less desirable in terms of the auditory sense. Thus, in order to reduce the connection distortion between the blocks, in carrying out spectral conversion using DFT or DCT in actual coding, M
1
units of sample data each of the adjacent blocks are caused to overlap each other for spectral conversion.
However, in the case where M
1
units of sample data each of the adjacent blocks are caused to overlap each other for spectral conversion, M units of real number data are obtained with respect to (M−M
1
) units of sample data on the average, and the number of real number da
Akagiri Kenzo
Hatanaka Mitsuyuki
Oikawa Yoshiaki
Sonnenschein Nath & Rosenthal
{haeck over (S)}mits T{overscore (a)}livaldis Ivars
LandOfFree
Subband coder or decoder band-limiting the overlap region... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Subband coder or decoder band-limiting the overlap region..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Subband coder or decoder band-limiting the overlap region... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2864564