CN106157978A - Speech signal processing device and audio signal processing method - Google Patents
Speech signal processing device and audio signal processing method Download PDFInfo
- Publication number
- CN106157978A CN106157978A CN201510177492.6A CN201510177492A CN106157978A CN 106157978 A CN106157978 A CN 106157978A CN 201510177492 A CN201510177492 A CN 201510177492A CN 106157978 A CN106157978 A CN 106157978A
- Authority
- CN
- China
- Prior art keywords
- value
- sampled
- window
- interpolated
- parameter function
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Transmission Systems Not Characterized By The Medium Used For Transmission (AREA)
Abstract
The present invention provides a kind of speech signal processing device and audio signal processing method.The present invention calculates the value of the interpolated parameter function corresponding with sampled signal window according to three sampled values of continuous print in sampled signal window, and calculates in down-scaled signals window the interpolated value between adjacent two sampled points according to the value of interpolated parameter function.There is the situation of distorted signals in voice signal after the present invention can be effectively prevented from frequency reducing.
Description
Technical field
The present invention relates to a kind of signal processing apparatus, particularly relate to a kind of speech signal processing device and voice
Signal processing method.
Background technology
For Hearing Impaired, it often cannot clearly receive the voice signal of higher-frequency, example
Such as consonant signal, but the signal for low frequency can clearly be heard.But, signal is being dropped
After Pin, owing to time span is elongated, the signal value between two sampled signals of continuous print need to utilize interpolation
Mode is tried to achieve, if such as by acoustical signal when high-frequency signal reduces to the low frequency signal of only half frequency,
Time span will become twice originally, the new signal between sampled signal and sampled signal, it is necessary to by interior
The mode inserted is tried to achieve.Owing to the Property comparison of acoustical signal is close to string ripple, if with the side of general arithmetic mean
Formula asks for the signal value of interpolation, often makes the signal after frequency reducing the situation of distorted signals occur.
Summary of the invention
The present invention provides a kind of speech signal processing device and audio signal processing method, can be effectively prevented from
There is the situation of distorted signals in voice signal after frequency reducing.
The speech signal processing device of the present invention includes processing unit, and it receives the sampled signal including sequence
The sampled speech signal of window, and calculate and each sampling according to three sampled values of continuous print in each sampled signal window
The value of the interpolated parameter function that signal window is corresponding, down sample voice signal, to produce the fall including sequence
Frequently the down-scaled signals of signal window, calculates each fall according to the value of interpolated parameter function corresponding to each down-scaled signals window
Frequently interpolated value between adjacent two sampled points in signal window.
In one embodiment of this invention, above-mentioned speech signal processing device also includes sampling unit, its coupling
Connecing processing unit, primary speech signal of sampling, to produce sampled speech signal, in processing unit also judges
Whether the value inserting parametric function is less than higher limit and more than or equal to lower limit, if the value of interpolated parameter function is not
Less than higher limit or not more than or equal to lower limit, the value of modified interpolation parametric function.
In one embodiment of this invention, if wherein the value of interpolated parameter function is more than or equal to higher limit, will
The value of interpolated parameter function is modified to higher limit, if the value of interpolated parameter function is less than lower limit, by interpolation
The value of parametric function is modified to lower limit.
In one embodiment of this invention, above-mentioned higher limit and lower limit are associated with the frequency of primary speech signal
Rate and the sample frequency of sampling unit.
In one embodiment of this invention, above-mentioned processing unit is also according to continuous print three in each sampled signal window
Trigonometric function relation between individual sampled value calculates the interpolated parameter function that each sampled signal window is corresponding.
In one embodiment of this invention, above-mentioned interpolated parameter function is trigonometric function.
The audio signal processing method of the present invention comprises the following steps.Sampling primary speech signal, to produce
Sampled speech signal including the sampled signal window of sequence.Adopt according to continuous print three in each sampled signal window
Sample value calculates the value of interpolated parameter function corresponding to each sampled signal window.Down sample voice signal, to produce
The down-scaled signals of the raw down-scaled signals window including sequence.According to the interpolated parameter letter that each down-scaled signals window is corresponding
The value of number calculates the interpolated value in each down-scaled signals window between adjacent two sampled points.
In one embodiment of this invention, above-mentioned audio signal processing method also includes, it is judged that interpolated parameter
Whether the value of function is less than higher limit and more than or equal to lower limit, if the value of interpolated parameter function is not less than upper
Limit value or more than or equal to lower limit, the value of modified interpolation parametric function.
In one embodiment of this invention, if wherein the value of interpolated parameter function is more than or equal to higher limit, will
The value of interpolated parameter function is modified to higher limit, if the value of interpolated parameter function is less than lower limit, by interpolation
The value of parametric function is modified to lower limit.
In one embodiment of this invention, wherein higher limit and lower limit are associated with the frequency of primary speech signal
Rate and the sample frequency of sampling unit.
In one embodiment of this invention, above-mentioned audio signal processing method includes, according to each sampled signal
In window, the trigonometric function relation between three sampled values of continuous print calculates the interpolated parameter that each sampled signal window is corresponding
Function.
In one embodiment of this invention, above-mentioned interpolated parameter function is trigonometric function.
Based on above-mentioned, embodiments of the invention calculate according to three sampled values of continuous print in sampled signal window
The value of the interpolated parameter function corresponding with sampled signal window, and calculate fall according to the value of interpolated parameter function
Frequently interpolated value between adjacent two sampled points in signal window, to obtain accurate interpolated value, is effectively prevented from fall
There is the situation of distorted signals in voice signal after Pin.
For the features described above of the present invention and advantage can be become apparent, special embodiment below, and coordinate
Accompanying drawing is described in detail below.
Accompanying drawing explanation
Fig. 1 is shown as the schematic diagram of the speech signal processing device of one embodiment of the invention；
Fig. 2 illustrates the schematic diagram of the down-scaled signals of one embodiment of the invention；
Fig. 3 illustrates the schematic flow sheet of the audio signal processing method of one embodiment of the invention.
Description of reference numerals:
102: processing unit；
104: sampling unit；
S1: primary speech signal；
S2: sampled speech signal；
S3: down-scaled signals；
Wm, Wm+1: down-scaled signals window；
S (2n), s (2n+2), s (2n+4), s (2n+6), s (2n+8): sampled point；
S (2n+1), s (2n+3), s (2n+5), s (2n+7): interpolated point；
S302～S312: step.
Detailed description of the invention
Fig. 1 is shown as the schematic diagram of the speech signal processing device of one embodiment of the invention, refer to Fig. 1.
Speech signal processing device includes processing unit 102 and sampling unit 104, and processing unit 102 couples
Sampling unit 104, wherein processing unit 102 can such as be implemented with CPU, and list of sampling
Unit 104 then can such as implement with logic circuit, but is not limited.Sampling unit 104 can be sampled former
Beginning voice signal S1, to produce sampled speech signal S2, wherein sampled speech signal S2 includes a sequence
Sampled signal window.Processing unit 102 can be according to three sampled value meters of continuous print in each sampled signal window
Calculate the value of the interpolated parameter function corresponding with each sampled signal window, additionally, also can down sample voice letter
Number S2 is to produce the down-scaled signals of the down-scaled signals window including a sequence, and according to each down-scaled signals window institute
The value of corresponding interpolated parameter function calculates the interpolation in each down-scaled signals window between adjacent two sampled points
Value, wherein interpolated parameter function is trigonometric function, such as SIN function or cosine function, but not as
Limit.
For example, Fig. 2 illustrates the schematic diagram of the down-scaled signals of one embodiment of the invention, refer to Fig. 2.
In fig. 2, the part of black circle is the sampled point of sampling unit 104, and the part of soft dot is then
The interpolated point calculated by processing unit 102.It is assumed herein that m-th is adopted in sampled speech signal S2
In sample signal window, the sampled value at time point n isWherein m is positive integer, and n is 0 or just
Integer.It addition, in the present embodiment, frequency reducing letter obtained after sampled speech signal S2 is carried out frequency reducing
The half of the frequency that frequency is sampled speech signal S2 of number S3, if assuming m in down-scaled signals S3
Time point n in individual down-scaled signals window Wm (the m-th sampled signal window of its corresponding sampled speech signal S2)
Sampled value be sm(n), then before and after frequency reducing, the corresponding relation of same sampled point can be shown below:
Processing unit 102 can calculate and each according to three sampled values of continuous print in each sampled signal window
The interpolated parameter function that sampled signal window is corresponding, such as, the interpolation ginseng corresponding to m-th sampled signal window
Number function CmG () can be according to sampling unit 104 three sampled values of continuous sampling in sampled signal windowAndBetween trigonometric function relation try to achieve, in sampled signal
Interpolated parameter function corresponding in the time range of window can be shown below:
Wherein g is 0 or positive integer, CmG () is the interpolated parameter function functional value at time point g, interpolation
Parametric function CmG () is trigonometric function.
Produce owing to speech signal processing device there may be noise during signal processing, and cause
The composition that the value of the interpolated parameter function calculated comprises noise, so asks for affecting processing unit 102
The degree of accuracy of interpolated value.Whether processing unit 102 can fall within default by the value judging interpolated parameter function
In the range of whether inspect the value of interpolated parameter function by noise jamming, such as can determine whether interpolated parameter letter
Whether the value of number is less than higher limit and more than or equal to lower limit, if the value of interpolated parameter function is not less than the upper limit
It is worth or more than or equal to lower limit, then the value of representation parameter function is by noise jamming, processing unit 102
Can the value of modified interpolation parametric function, to remove the noise contribution included in the value of interpolated parameter function.
Such as, if the value of interpolated parameter function is more than or equal to higher limit, processing unit 102 can be by interpolated parameter letter
The value of number is modified to higher limit, if the value of interpolated parameter function is less than lower limit, processing unit 102 can be by
The value of interpolated parameter function is modified to lower limit, and if the value of interpolated parameter function less than higher limit and is more than
Equal to lower limit, then it is not required to the value of interpolation parametric function is modified.For example, in the reality of Fig. 2
Execute in example, interpolated parameter function CmG the correcting mode of the value of () can be with following formula subrepresentation:
The most above-mentioned higher limit and lower limit are respectively 1 and 0.5 in the embodiment of fig. 2, if voice
Signal processing apparatus by effect of noise, and makes interpolated parameter function during signal processing
CmG the value of () is more than or equal to 1, then processing unit 102 is by interpolated parameter function CmG the value of () is modified to 1,
And if interpolated parameter function CmG the value of () is less than 0.5, then processing unit 102 is by interpolated parameter function Cm(g)
Value be modified to 0.5.It should be noted that higher limit and the lower limit of formula (3) are only exemplary embodiment,
It is not limited thereto.Wherein the situation of higher limit and the interference of lower limit visual actual noise adjusts, such as
Higher limit and lower limit can be adjusted according to the sample frequency of the frequency of primary speech signal Yu sampling unit.
After the value obtaining interpolated parameter function, processing unit 102 just can be counted according to interpolated parameter function
Calculate the interpolated value between adjacent two sampled points in down-scaled signals window.As a example by the embodiment of Fig. 2, believe in frequency reducing
In number window Wm interpolated point s (2n+1) between the sampled point s (2n), s (2n+2) of sampling unit 104 with
And the interpolated point s (2n+3) between sampled point s (2n+2), s (2n+4) can distinguish shown in following formula:
In formula (4), formula (5), n is 0 or positive even numbers.The rest may be inferred, sampled point in other down-scaled signals windows
Between interpolated value can also identical mode try to achieve, such as sampled point in the down-scaled signals window Wm+1 of Fig. 2
Interpolated point s (2n+5) between s (2n+4), s (2n+6) and between sampled point s (2n+6), s (2n+8)
Interpolated point s (2n+7) can also the embodiment of Fig. 2 try to achieve, those skilled in the art should be according to above-mentioned
The method of embodiment pushes away to obtain its embodiment, thus does not repeats them here.
As it has been described above, the present embodiment is the interpolated value utilizing trigonometric function to come between estimating sampling point, according to interior
Slotting parametric function calculates the interpolated value in down-scaled signals window between adjacent two sampled points, due to trigonometric function
Characteristic is more similar to the characteristic of acoustical signal, therefore merely utilizes arithmetical average compared to prior art
Asking for interpolated value, the calculation of the present embodiment can obtain more accurate interpolated value, and can effectively keep away
Exempt from the voice signal after frequency reducing and the situation of distorted signals occurs.
Fig. 3 illustrates the schematic flow sheet of the audio signal processing method of one embodiment of the invention, refer to figure
3.From above-described embodiment, the audio signal processing method of speech signal processing device can include following step
Suddenly.First, primary speech signal of sampling, to produce the sampled speech of the sampled signal window including a sequence
Signal (step S302).Then, each sampling is calculated according to three sampled values of continuous print in each sampled signal window
The value (step S304) of the interpolated parameter function that signal window is corresponding, wherein interpolated parameter function can be according to respectively adopting
In sample signal window, the trigonometric function relation between three sampled values of continuous print calculates and obtains, and interpolated parameter function can
For trigonometric function.Afterwards, can then judge whether the value of interpolated parameter function is less than higher limit and is more than
In lower limit (step S306), if the value of interpolated parameter function is not less than higher limit or not more than or equal to lower limit
Value, then the value (step S308) of modified interpolation parametric function, to remove unnecessary noise.The wherein upper limit
The situation of value and the interference of lower limit visual actual noise adjusts, such as can be according to the frequency of primary speech signal
The sample frequency of rate and sampling unit adjusts higher limit and lower limit, and the repairing of the value of interpolated parameter function
Positive way can for example, when the value of interpolated parameter function is more than or equal to higher limit, by interpolated parameter function
Value be modified to higher limit, when interpolated parameter function value less than lower limit time, by interpolated parameter function
Value is modified to lower limit.After the value having revised interpolated parameter function, can then down sample voice signal,
To produce the down-scaled signals (step S310) of the down-scaled signals window including a sequence, then believe according to each frequency reducing
The value of number interpolated parameter function that window is corresponding calculates the interpolated value in each down-scaled signals window between adjacent two sampled points
(step S312).If on the contrary, the value of interpolated parameter function is less than higher limit and more than or equal to lower limit,
Then it is directly entered step S310, down sample voice signal.
In sum, embodiments of the invention utilize trigonometric function to carry out the interpolated value between estimating sampling point, also
Just it is based on interpolated parameter function to the interpolated value calculating in down-scaled signals window between adjacent two sampled points, due to
The characteristic of trigonometric function is more similar to the characteristic of acoustical signal, therefore compared to prior art, can obtain more
For accurate interpolated value, and the voice signal after frequency reducing can be effectively prevented from the situation of distorted signals occurs.
Last it is noted that various embodiments above is only in order to illustrate technical scheme, rather than to it
Limit；Although the present invention being described in detail with reference to foregoing embodiments, the common skill of this area
Art personnel it is understood that the technical scheme described in foregoing embodiments still can be modified by it,
Or the most some or all of technical characteristic is carried out equivalent；And these amendments or replacement, and
The essence not making appropriate technical solution departs from the scope of various embodiments of the present invention technical scheme.
Claims (10)
1. a speech signal processing device, it is characterised in that including:
Processing unit, receives the sampled speech signal of the sampled signal window including sequence, and according to each described
In sampled signal window, three sampled values of continuous print calculate the interpolated parameter letter corresponding with each described sampled signal window
The value of number, sampled speech signal described in frequency reducing, to produce the down-scaled signals of the down-scaled signals window including sequence,
Phase in each described down-scaled signals window is calculated according to the value of interpolated parameter function corresponding to each described down-scaled signals window
Interpolated value between adjacent two sampled points.
Speech signal processing device the most according to claim 1, it is characterised in that also include:
Sampling unit, couples described processing unit, and primary speech signal of sampling, to produce described sampling language
Tone signal, described processing unit also judges whether the value of described interpolated parameter function is less than higher limit and is more than
Equal to lower limit, if the value of described interpolated parameter function is not less than described higher limit or not more than or equal to described
Lower limit, revises the value of described interpolated parameter function.
Speech signal processing device the most according to claim 2, it is characterised in that if described interpolation
The value of parametric function is more than or equal to described higher limit, the value of described interpolated parameter function is modified on described
Limit value, if the value of described interpolated parameter function is less than described lower limit, by the value of described interpolated parameter function
It is modified to described lower limit.
Speech signal processing device the most according to claim 3, it is characterised in that described higher limit
The frequency of described primary speech signal and the sample frequency of described sampling unit it is associated with described lower limit.
Speech signal processing device the most according to claim 1, it is characterised in that described process list
Unit also calculates each institute according to trigonometric function relation between three sampled values of continuous print in each described sampled signal window
State the interpolated parameter function that sampled signal window is corresponding.
Speech signal processing device the most according to claim 5, it is characterised in that described interpolation is joined
Number function is trigonometric function.
7. an audio signal processing method, it is characterised in that including:
Sampling primary speech signal, to produce the sampled speech signal of the sampled signal window including sequence；
Calculate and each described sampled signal window pair according to three sampled values of continuous print in each described sampled signal window
The value of the interpolated parameter function answered；
Sampled speech signal described in frequency reducing, to produce the down-scaled signals of the down-scaled signals window including sequence；With
And
Each described down-scaled signals window is calculated according to the value of interpolated parameter function corresponding to each described down-scaled signals window
In interpolated value between adjacent two sampled points.
Audio signal processing method the most according to claim 7, it is characterised in that also include:
Judge whether the value of described interpolated parameter function is less than higher limit and more than or equal to lower limit, if described
The value of interpolated parameter function, not less than described higher limit or not more than or equal to described lower limit, is revised described interior
Insert the value of parametric function.
Audio signal processing method the most according to claim 8, it is characterised in that if described interpolation
The value of parametric function is more than or equal to described higher limit, the value of described interpolated parameter function is modified on described
Limit value, if the value of described interpolated parameter function is less than described lower limit, by the value of described interpolated parameter function
It is modified to described lower limit.
Audio signal processing method the most according to claim 9, it is characterised in that the described upper limit
Value and described lower limit are associated with the frequency of described primary speech signal and the sampling frequency of described sampling unit
Rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510177492.6A CN106157978B (en) | 2015-04-15 | 2015-04-15 | Speech signal processing apparatus and speech signal processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510177492.6A CN106157978B (en) | 2015-04-15 | 2015-04-15 | Speech signal processing apparatus and speech signal processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106157978A true CN106157978A (en) | 2016-11-23 |
| CN106157978B CN106157978B (en) | 2020-04-07 |
Family
ID=57336970
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510177492.6A Active CN106157978B (en) | 2015-04-15 | 2015-04-15 | Speech signal processing apparatus and speech signal processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106157978B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5987082A (en) * | 1996-07-30 | 1999-11-16 | Sony Corporation | Playback apparatus and playback method |
| US20100278356A1 (en) * | 2004-04-01 | 2010-11-04 | Phonak Ag | Audio amplification apparatus |
| CN101981612A (en) * | 2008-09-26 | 2011-02-23 | 松下电器产业株式会社 | Speech analyzing apparatus and speech analyzing method |
| CN102496374A (en) * | 2011-12-16 | 2012-06-13 | 河海大学常州校区 | Hearing compensation method |
| US20140288926A1 (en) * | 2009-09-11 | 2014-09-25 | Texas Instruments Incorporated | Method and system for interference suppression using blind source separation |
| WO2015006112A1 (en) * | 2013-07-08 | 2015-01-15 | Dolby Laboratories Licensing Corporation | Processing of time-varying metadata for lossless resampling |
| CN104378075A (en) * | 2008-12-24 | 2015-02-25 | 杜比实验室特许公司 | Audio signal loudness determination and modification in the frequency domain |
-
2015
- 2015-04-15 CN CN201510177492.6A patent/CN106157978B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5987082A (en) * | 1996-07-30 | 1999-11-16 | Sony Corporation | Playback apparatus and playback method |
| US20100278356A1 (en) * | 2004-04-01 | 2010-11-04 | Phonak Ag | Audio amplification apparatus |
| CN101981612A (en) * | 2008-09-26 | 2011-02-23 | 松下电器产业株式会社 | Speech analyzing apparatus and speech analyzing method |
| CN104378075A (en) * | 2008-12-24 | 2015-02-25 | 杜比实验室特许公司 | Audio signal loudness determination and modification in the frequency domain |
| US20140288926A1 (en) * | 2009-09-11 | 2014-09-25 | Texas Instruments Incorporated | Method and system for interference suppression using blind source separation |
| CN102496374A (en) * | 2011-12-16 | 2012-06-13 | 河海大学常州校区 | Hearing compensation method |
| WO2015006112A1 (en) * | 2013-07-08 | 2015-01-15 | Dolby Laboratories Licensing Corporation | Processing of time-varying metadata for lossless resampling |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106157978B (en) | 2020-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7970609B2 (en) | Method of estimating sound arrival direction, sound arrival direction estimating apparatus, and computer program product | |
| US8271051B2 (en) | Method and system for double-end talk detection, and method and system for echo elimination | |
| US10026418B2 (en) | Abnormal frame detection method and apparatus | |
| US7912230B2 (en) | Howling detection device and method | |
| EP3091534A1 (en) | Method and apparatus for processing speech signal according to frequency domain energy | |
| EP2465113B1 (en) | Method, computer program product and system for determining a perceived quality of an audio system | |
| CN102479504A (en) | Speech determination apparatus and speech determination method | |
| US20120020484A1 (en) | Audio Signal Quality Prediction | |
| EP2608201A2 (en) | Signal processing apparatus and signal processing method | |
| US20160217787A1 (en) | Speech recognition apparatus and speech recognition method | |
| US9495973B2 (en) | Speech recognition apparatus and speech recognition method | |
| JP2008116954A (en) | Generation of sample error coefficients | |
| EP3007171A1 (en) | Signal processing device and signal processing method | |
| CN106157978A (en) | Speech signal processing device and audio signal processing method | |
| US20120089357A1 (en) | Method and apparatus for identifying anomalies of a signal | |
| KR101615563B1 (en) | Diagnosis method of fatigue crack and diagnosis device | |
| TWI566241B (en) | Voice signal processing apparatus and voice signal processing method | |
| JP6703840B2 (en) | Sound quality evaluation device | |
| US20190066714A1 (en) | Method, information processing apparatus for processing speech, and non-transitory computer-readable storage medium | |
| US10297268B2 (en) | Voice signal processing apparatus and voice signal processing method | |
| CN106254001A (en) | A kind of SRS signal channel quality measuring method and device | |
| US9761242B2 (en) | Voice signal processing apparatus and voice signal processing method | |
| CN105868553A (en) | Track data processing method | |
| EP1516466B1 (en) | Method for estimating noise auto-correlation | |
| CN105845152A (en) | Method for detecting audio signal echoes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |