EP0566131A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Method and device for discriminating voiced and unvoiced sounds
Assignee: Sony Corp
•
Filing: 1993-04-15
•
Granted: —
•
Expires: 2013-04-15
Abstract
Abstract A method and a device for discriminating a voiced sound from an unvoiced sound or background noise in speech signals are disclosed. Each block or frame of input speech signals is divided into plural sub-blocks and the standard deviation, effective value or the peak value is detected in a detection unit for detecting statistical characteristics from one sub-block to another. A bias detection unit (17) detects a bias on the time scale of the standard deviation, effective value or the peak value to decide whether the speech signals are voiced or unvoiced from one block to another.
Inventors
- 👤 Masayuki C/O Sony Corporation Nishiguchi
- 👤 Jun C/O Sony Corporation Matsumoto
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-02 17:08
CPC Classifications
G; G10; G10L; G10L25/00; G10L25/93; G10L25/78; G10L2025/783; G10L2025/932
Claims Analysis
Total Claims: 41
Claim 1
Independent Claim
103 words
A method for discriminating a digital speech sound comprising dividing digital speech signals into blocks each consisting of a predetermined number of samples, and making a decision for each of said blocks as to whether or not the speech sound is voiced, said method further comprising the steps of dividing signals of said block into plural sub-blocks, finding statistical characteristics of signals of each of said sub-blocks, finding a bias of said statistical characteristics of said signals on the time domain, and deciding whether said signals of said blocks are voiced based on said bias of said statistical characteristics on the time domain.
Claim 2
Dependent Claim
24 words
The method as claimed in claim 1 wherein said statistical characteristics are found based on the standard deviation of said signals constituting said sub-blocks.
Claim 3
Dependent Claim
24 words
The method as claimed in claim 1 wherein said statistical characteristics are found based on the effective values of said signals constituting said sub-blocks.
Claim 4
Dependent Claim
35 words
The method as claimed in claim 1 wherein said bias of said statistical characteristics of said signals on the time domain is found based on the arithmetical mean and geometrical mean of said statistical characteristics.
Claim 5
Dependent Claim
38 words
The method as claimed in claim 4 wherein said dispersion of said statistical characteristics of said signals on the time domain is found by finding the ratio between the arithmetical mean and geometrical mean of said statistical characteristics.
Claim 6
Dependent Claim
24 words
The method as claimed in claim 1 wherein said statistical characteristics are found based on the peak values of said signals constituting said sub-blocks.
Claim 7
Dependent Claim
70 words
The method as claimed in claim 6 wherein said statistical characteristics are found by the step of finding the standard deviation of said signals of said blocks and the step of finding a mean peak value from peak values of signals of said sub-blocks and wherein the bias of said statistical characteristics on the time domain is found from the ratio between said standard deviation and said mean peak value.
Claim 8
Independent Claim
115 words
A method for discriminating a digital speech sound comprising dividing digital speech signals into blocks each consisting of a predetermined number of samples, and making a decision for each of said blocks as to whether or not the speech sound is voiced, said method further comprising the steps of transforming signals of each of said blocks into data on the frequency scale, finding low frequency range energies based on said data on the frequency scale, finding high frequency range energies based on said data on the frequency scale, and deciding whether or not the signals of each of said blocks are voiced based on said low frequency range energies and said high frequency range energies.
Claim 9
Dependent Claim
31 words
The method as claimed in claim 8 wherein said low frequency range energies and said high frequency range energies are demarcated from each other at a frequency of approximately 2 kHz.
Claim 10
Dependent Claim
48 words
The method as claimed in claim 8 or 9 comprising the step of finding the ratio between said low frequency range energies and said high frequency range energies, said ratio being used as basis in deciding whether or not the signals of each of said blocks are voiced.
Claim 11
Independent Claim
57 words
The method as claimed in any of claims 8 to 10 comprising the steps of finding the mean signal level of signals of each of said blocks from low frequency range energies and high frequency range energies, said ratio being used as basis in deciding whether or not the signals of each of said blocks are voiced.
Claim 12
Dependent Claim
66 words
The method as claimed in claim 10 comprising the steps of finding the ratio between low frequency range energies and high frequency range energies and the mean signal level from said low frequency range energies and said high frequency range energies, said ratio and the mean signal level being used as basis in deciding whether or not the signals of each of said blocks are voiced.
Claim 13
Independent Claim
81 words
The method as claimed in any of claims 8 to 12 comprising the steps of dividing signals of said block into plural sub-blocks, finding statistical characteristics of signals of each of said sub-blocks, finding a bias of said statistical characteristics of said signals on the time domain, and deciding whether said signals of said blocks are voiced based on said low frequency range energies and said high frequency range energies and said bias of said statistical characteristics on the time domain.
Claim 14
Dependent Claim
53 words
The method as claimed in claim 13 wherein a decision as to whether or not the signals of each of said blocks are voiced is given based on the ratio between said low frequency range energies and said high frequency range energies and said bias of said statistical characteristics on the time domain.
Claim 15
Dependent Claim
65 words
The method as claimed in claim 13 or 14 wherein a mean signal level is found based on said low frequency range energies and said high frequency range energies and a decision as to whether or not the signals of each of said blocks are voiced is given based on said mean signal level and said bias of said statistical characteristics on the time domain.
Claim 16
Independent Claim
81 words
The method as claimed in any of claims 13 to 15 wherein a mean signal level and the ratio between low frequency range energies and high frequency range energies are found based on said low frequency range energies and said high frequency range energies and a decision as to whether or not the signals of each of said blocks are voiced is given based on said ratio, mean signal level and the bias of said statistical characteristics on the time domain.
Claim 17
Independent Claim
132 words
A method for discriminating a digital speech sound comprising dividing digital speech signals into blocks each consisting of a predetermined number of samples, and making a decision as to whether or not the speech sound is voiced for each of said blocks, said method further comprising the steps of finding an effective value of signals in each of a plurality of sub-blocks divided from each of said blocks, finding a standard value and a mean value of said signals of each block based on the effective value as found for each of said sub-blocks, finding a normalized standard value based on said standard deviation and said mean value, and giving a decision as to whether or not the signals of each of said block are voiced based on said normalized standard deviation.
Claim 18
Dependent Claim
68 words
The method as claimed in claim 17 comprising the steps of frequency-analyzing the signals of each of said blocks to find spectral intensities at each frequency, finding energy distribution based on the spectral intensity at each point on the frequency domain, and giving a decision as to whether or not the signals of each of said blocks are voiced from said normalized standard deviation and said energy distribution.
Claim 19
Dependent Claim
44 words
The method as claimed in claim 18 wherein said spectral intensities at each point of the frequency domain are divided into groups of low-range frequency and high-range frequency and energy distribution is found based on the ratio between the energies of the respective groups.
Claim 20
Dependent Claim
31 words
The method as claimed in claim 19 wherein said low frequency range energies and said high frequency range energies are demarcated from each other at a frequency of approximately 2 kHz.
Claim 21
Independent Claim
60 words
The method as claimed in any of claims 18 to 20 comprising the step of finding the mean level of the signals of each of the blocks from aid energy distribution and wherein a decision as to whether the signals of each of the blocks are voiced based on said normalized standard deviation, energy distribution and the mean signal level.
Claim 22
Independent Claim
126 words
An apparatus for discriminating a digital speech sound by dividing digital speech signals into blocks each consisting of a predetermined number of samples, and making a decision as to whether or not the speech sound is voiced for each of said blocks, said apparatus comprising means (13) for dividing signals of said block into plural sub-blocks, means (16) for finding statistical characteristics of signals of each of said sub-blocks, means (17) for finding a bias on the time domain of statistical characteristics of signals outputted from said means (16) for finding statistical characteristics of signals of each of said sub-blocks, and means (18) for deciding whether said signals of said blocks are voiced based on bias data outputted from said means (17) for finding the dispersion.
Claim 23
Dependent Claim
41 words
The apparatus as claimed in claim 22 wherein statistical characteristics of the signals of each of the sub-blocks are calculated by said means (16) for finding statistical characteristics based on the standard deviation of the signals of each of the sub-blocks.
Claim 24
Dependent Claim
43 words
The apparatus as claimed in claim 22 or 23 wherein statistical characteristics of the signals of each of the sub-blocks are calculated by said means (16) for finding statistical characteristics based on the effective value of the signals of each of the sub-blocks.
Claim 25
Independent Claim
63 words
The apparatus as claimed in any of claims 22 to 24 further comprising arithmetic mean calculating means (23) for finding an arithmetic mean of statistical characteristics of signals and geometric mean calculating means (24) for finding a geometric mean of statistical characteristics of signals, a bias on the time domain of said statistical characteristics of the signals being found from these mean values.
Claim 26
Dependent Claim
27 words
The apparatus as claimed in claim 25 further comprising means (17) for finding a ratio between the arithmetic mean and the geometric mean based on said ratio.
Claim 27
Independent Claim
41 words
The apparatus as claimed in any of claims 22 to 26 wherein the statistical characteristics of the signals are calculated by said means (16) for finding statistical characteristics based on a peak value of the signals of each of the sub-blocks.
Claim 28
Dependent Claim
84 words
The apparatus as claimed in claim 27 wherein said means (16) for finding statistical characteristics comprise standard deviation calculating means (28) for finding the standard deviation of the signals of each of said blocks, mean peak value calculating means (27) for calculating a mean peak value from the peak value of the signals of each of the sub-blocks, and bias calculating means for finding the bias of statistical characteristics of the signals from the ratio between the standard deviation and the mean peak value.
Claim 29
Independent Claim
122 words
An apparatus for discriminating a digital speech sound by dividing digital speech signals into blocks each consisting of a predetermined number of samples, and making a decision for each of said blocks, as to whether or not the speech sound is voiced said apparatus comprising frequency data calculating means (33) for transforming signals of each of said blocks into frequency-domain data, means (34a) for finding low frequency range energies based on said frequency-domain data, means (34b) for finding high frequency range energies based on said frequency-domain data, and decision means (37) for giving a decision as to whether or not the signals of each of the blocks are voiced based on said low frequency range energies and said high frequency range energies.
Claim 30
Dependent Claim
72 words
The apparatus as claimed in claim 29 wherein an upper limit of the frequency in said means (34a) for finding low frequency range energies is approximately 2 kHz and a lower limit of the frequency in said means (34b) for finding high frequency range energies is approximately 2 kHz, said low frequency range energies and said high frequency range energies being demarcated from each other at said frequency of approximately 2 kHz.
Claim 31
Dependent Claim
49 words
The apparatus as claimed in claim 29 or 30 further comprising means (35) for finding the ratio between said low frequency range energies and said high frequency range energies, said ratio being used as basis in deciding whether or not the signals of each of said blocks are voiced.
Claim 32
Dependent Claim
55 words
The apparatus as claimed in claim 29 further comprising means (15, 17) for finding the mean signal level of signals of each of said blocks from low frequency range energies and high frequency range energies, said ratio being used as basis in deciding whether or not the signals of each of said blocks are voiced.
Claim 33
Dependent Claim
67 words
The apparatus as claimed in claim 31 further comprising means (35) for finding the ratio between low frequency range energies and high frequency range energies and the mean signal level from said low frequency range energies and said high frequency range energies, said ratio and the mean signal level being used as basis in deciding whether or not the signals of each of said blocks are voiced.
Claim 34
Independent Claim
84 words
The apparatus as claimed in any of claims 29 to 33 further comprising means (13, 15, 17) for dividing signals of said block into plural sub-blocks, finding statistical characteristics of signals of each of said sub-blocks, finding a bias of said statistical characteristics of said signals on the time domain, and deciding whether said signals of said blocks are voiced based on said low frequency range energies and said high frequency range energies and said bias of said statistical characteristics on the time domain.
Claim 35
Dependent Claim
56 words
The apparatus as claimed in claim 34 further comprising means (39) for giving a decision as to whether or not the signals of each of said blocks are voiced based on the ratio between said low frequency range energies and said high frequency range energies and said bias of said statistical characteristics on the time domain.
Claim 36
Dependent Claim
64 words
The apparatus as claimed in claim 34 further comprising means for finding a mean signal level based on said low frequency range energies and said high frequency range energies and giving a decision as to whether or not the signals of each of said blocks are voiced based on said mean signal level and said bias of said statistical characteristics on the time domain.
Claim 37
Independent Claim
84 words
The apparatus as claimed in any of claims 29 to 36 further comprising means for finding a mean signal level and the ratio between low frequency range energies and high frequency range energies based on said low frequency range energies and said high frequency range energies and means for giving a decision as to whether or not the signals of each of said blocks are voiced based on said ratio, mean signal level and the bias of said statistical characteristics on the time domain.
Claim 38
Independent Claim
140 words
An apparatus for discriminating a digital speech sound by dividing digital speech signals into blocks each consisting of a predetermined number of samples, and making a decision for each of said blocks as to whether or not the speech sound is voiced, said apparatus comprising means (16) for finding an effective value of signals in each of a plurality of sub-blocks divided from each of said blocks, means (62) for finding a standard deviation and a mean value of said signals of each block based on the effective value as found for each of said sub-blocks, means (63) for finding a normalized standard value based on said standard deviation and said mean value, and means (64) for giving a decision as to whether or not the signals of each of said block are voiced based on said normalized standard deviation.
Claim 39
Dependent Claim
79 words
The apparatus as claimed in claim 38 further comprising means (53, 54) for frequency-analyzing signals of each of said blocks to find spectral intensities at each point on the frequency domain, means (55) for finding energy distribution based on the spectral intensity at each point of the frequency domain, and means (64) for giving a decision as to whether or not the signals of each of said blocks are voiced from said normalized standard deviation and said energy distribution.
Claim 40
Dependent Claim
44 words
The apparatus as claimed in claim 39 wherein said spectral intensities at each point of the frequency domain are divided into groups of low-range frequency and high-range frequency and energy distribution is found based on the ratio between the energies of the respective groups.
Claim 41
Dependent Claim
63 words
The apparatus as claimed in claim 39 or 40 further comprising means (56) for finding the mean level of the signals of each of the blocks from said energy distribution, with said normalized standard deviation, energy distribution and the mean signal level being used as a basis in giving a decision as to whether the signals of each of the blocks are voiced.
References Cited (Backward Citations)
US4625327A
Cited
US4637046A
Cited
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.