CN111431625B - Synthetic and modification method for vocals of cetacea animals - Google Patents

Synthetic and modification method for vocals of cetacea animals Download PDF

Info

Publication number
CN111431625B
CN111431625B CN202010117205.3A CN202010117205A CN111431625B CN 111431625 B CN111431625 B CN 111431625B CN 202010117205 A CN202010117205 A CN 202010117205A CN 111431625 B CN111431625 B CN 111431625B
Authority
CN
China
Prior art keywords
signal
bionic
frequency
time
biomimetic
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.)
Active
Application number
CN202010117205.3A
Other languages
Chinese (zh)
Other versions
CN111431625A (en
Inventor
蒋佳佳
孙中波
段发阶
王宪全
李春月
徐俊宇
杨国梁
李首玉
吕傲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
Original Assignee
Tianjin University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tianjin University filed Critical Tianjin University
Priority to CN202010117205.3A priority Critical patent/CN111431625B/en
Publication of CN111431625A publication Critical patent/CN111431625A/en
Application granted granted Critical
Publication of CN111431625B publication Critical patent/CN111431625B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B11/00—Transmission systems employing sonic, ultrasonic or infrasonic waves
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B13/00—Transmission systems characterised by the medium used for transmission, not provided for in groups H04B3/00 - H04B11/00
    • H04B13/02—Transmission systems in which the medium consists of the earth or a large mass of water thereon, e.g. earth telegraphy

Abstract

The invention discloses a synthetic and modifying method of whale animal crySetting the final sound of a cetacea animal as sT(t) the corresponding biomimetic signal is sb(t) to construct sb(t) mixing sT(t) is divided into M sections, M is a positive integer greater than or equal to 1, and M bionic signal sections are used for simulating s divided into M sectionsT(t); finally, splicing the M bionic signal sections end to obtain sb(t) by modifying s after splicingb(t) the modification of the final sound is realized by the parameter of (t) so as to meet the validity requirements of bionic camouflage hidden underwater sound detection and communication; the method specifically comprises the following steps: (1) construction of a biomimetic signal sb(t) time-frequency expressions of each bionic signal segment; (2) construction of a segmented biomimetic signal sN(t); (3) extracting an envelope of the final sound, and taking the envelope as an envelope of the bionic signal; (4) synthesizing a bionic signal sb(t); (5) if the tonal sound has harmonic, repeating the steps (1) - (4) until the finishing is finished.

Description

Synthetic and modification method for vocals of cetacea animals
Technical Field
The invention belongs to the field of underwater acoustic signal processing, and particularly relates to a synthetic and modification method of whale animal cry.
Background
Since signals are actively transmitted, Active Sonar Detection (ASD) systems and Underwater Acoustic Communication (UAC) systems are easily discovered by Underwater surveillance systems, and researchers in various countries have proposed many methods to improve concealment of ASD signals and UAC signals over the past decades. Traditional ASD signal and UAC signal concealment techniques can be divided into three categories: low Probability of Interception (LPI) technologies (first class), LPD (Low Probability of Detection, LPD) technologies (second class), and biomimetic camouflage concealed underwater sound Detection and communication technologies (third class).
The LPI technology continuously changes the parameters of the traditional artificial signals through the technologies of frequency hopping, time hopping and the like to improve the difficulty of identifying the signals so as to realize covert detection and communication. However, the artificial signals adopted by the technology have obvious characteristics and are easily identified by an enemy underwater reconnaissance system.
The LPD technique enables concealment of ASD and UAC by hiding the signal in the ocean background noise by reducing the signal-to-noise of the traditional artificial signal. But can still be detected by energy detection, energy spectral density analysis, etc., and the low signal-to-noise ratio signal can greatly reduce the distance and reliability of detection and communication.
Different from the LPI technology and the LPD technology, the idea of the bionic camouflage concealed underwater sound detection and communication technology is as follows: the sonar signals or the communication signals are disguised as the cry of the cetacea animals which are widely existed in the sea, so as to trap the enemy, and the communication and the sonar signals are mistakenly filtered out as the noise of the sea animals generated by the cetacea animals, thereby realizing the concealment of the communication and the sonar signals.
The design of the bionic signal is the key of bionic camouflage concealed detection and communication. More specifically, the biomimetic signal should meet the validity (e.g., signal transmission distance, communication rate, and detection accuracy) and concealment requirements of the ASD and UAC. The effectiveness of the ASD and UAC depends, among other things, on the detection (or communication) performance of the biomimetic signal. In addition, the concealment depends on the disguising ability of the bionic signal, and further, the disguising ability of the bionic signal depends on the similarity of the bionic signal and the sound of a real cetacea animal. Therefore, both the disguising ability and effectiveness of the biomimetic signal must be considered.
The design of the bionic signal comprises synthesis and modification of the signal. In order to ensure the disguising ability, the acoustic characteristics (such as waveform shape, frequency distribution and time-frequency distribution) of the sound of the real cetacea animals are required to be as close as possible when synthesizing the bionic signals. However, a biomimetic signal with high camouflage capability does not necessarily meet the performance requirements of ASD and UAC. Therefore, in order to ensure effectiveness, the biomimetic signal also needs to be modified, which means that parameters of the biomimetic signal need to be fine-tuned according to the performance requirements of the ASD and UAC.
For many years, some progress has been made in the design of biomimetic signals based on the sounds of cetacea animals. Some researchers used original whale sounds to construct the biomimetic signal. However, the original cetacea sound database is limited and it is difficult to find an original cetacea sound that meets the ASD and UAC validity requirements. Because of these limitations of using original whale sounds, some researchers have attempted to create suitable biomimetic signal models to mimic original whale animal sounds.
The Tonal sound is an important component of the sounds of animals of the order Cetaceae, which are produced by odontocetis and bearwhales, which are sister branches containing all existing whales. Since the acoustic properties of the tonal sound vary widely among animals of the order cetacea, this sound is generally defined as a frequency modulated signal.
In order to meet the validity requirements of the biomimetic signal, the biomimetic signal model should parameterize the pitch so that the parameters of the biomimetic signal can be conveniently set according to the performance requirements of the ASD and UAC. In addition, in order to meet the camouflage requirement of the bionic signal, the bionic signal model should realize high-similarity simulation of various types of final sound.
The current tonal sound model can be largely divided into two categories. The first type of model synthesizes a biomimetic signal by weighted superposition of sinusoidal signals. However, such models can only extract the contour of the final sound, and cannot parameterize the contour, so that the validity requirement cannot be met; the second type of model is constructed based on a conventional Frequency Modulation (FM) signal model, such as Linear Frequency Modulation (LFM) signals and Hyperbolic Frequency Modulation (HFM) signals. However, because the time-frequency structure of the final sound is complex and various, the degree of freedom of the traditional FM signal model is low, and high-similarity simulation of various final sounds cannot be realized, the requirement of disguise cannot be met.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a method for synthesizing and modifying the vocals of cetacea animals, which can meet the requirements of effectiveness and disguise of hidden underwater sound detection and communication.
The purpose of the invention is realized by the following technical scheme:
a method for synthesizing and modifying vocals of animals of the order Cetaceae comprises setting the final sound of an animal of the order Cetaceae as sT(t) the corresponding biomimetic signal is sb(t) to construct sb(t) mixing sT(t) is divided into M sections, M is a positive integer greater than or equal to 1, and M bionic signal sections are used for simulating s divided into M sectionsT(t); finally, splicing the M bionic signal sections end to obtain sb(t) by modifying s after splicingb(t) the modification of the final sound is realized by the parameter of (t) so as to meet the validity requirements of bionic camouflage hidden underwater sound detection and communication; the method specifically comprises the following steps:
(1) construction of a biomimetic signal sb(t) time-frequency expressions of each bionic signal segment;
(2) construction of a segmented biomimetic signal sN(t)；
(3) Extracting an envelope of the final sound, and taking the envelope as an envelope of the bionic signal;
(4) synthesizing a bionic signal sb(t)；
(5) If the tonal sound has harmonic, repeating the steps (1) - (4) until the finishing is finished.
Preferably, the specific implementation steps of step (1) are as follows:
(101) setting two Bionic signal models, namely a Power Frequency Modulation (PFMB) signal model and a Sinusoidal Frequency Modulation (SFMB) signal model; each of the PFMB signal model and the SFMB signal model includes 4 sub-signal models, a bionic signal sb(t) each biomimetic signal segment in (t) is constructed based on the 8 sub-signal models;
(102) obtaining expressions of four sub-PFMB signal models according to the time-frequency expressions and the function graphs of the 4 sub-PFMB signal models;
(103) obtaining expressions of four sub SFMB signal models according to the time-frequency expressions and the function graphs of the 4 sub SFMB signal models;
(104) by modifying the parameters in the PFMB signal model and the SFMB signal model, the method comprises the following steps: signal duration T, signal bandwidth B, signal carrier frequency fCA curvature regulating factor alpha, a curvature regulating factor beta, a slope regulating factor k and a slope regulating factor h, so that the bionic signal sb(t) in the presence of a terminal soundT(t) increasing the biomimetic signal s while being similarb(t) performance of underwater exploration or communication;
(105) to ensure sb(t) and sT(t) high similarity between them, sbThe smoothness of the time-frequency spectrum profile of (t) should be equal to sT(t) is coincident, i.e. at sbAnd (t) at the connection position of two adjacent bionic signal sections, the frequency and the time-frequency spectrum contour slope are continuously changed.
Preferably, the step (2) is implemented by the following steps:
(201)sN(t) mth bionic signal segment sb,mThe expression of (t) is:
Figure BDA0002391859850000031
wherein phib,m(t) is a phase expression, fb,m(t) is constructed based on a time-frequency representation of a sub-PFMB signal or sub-SFMB signal, sb,m(t) is sb(t) M (M ═ 1,2, …, M) number of biomimetic signal segments, sb,m(T) has a duration of Tb,m；
(202) Translating each bionic signal segment on a time axis to form an m-th bionic signal segment sb,m(T) time of translation is TD,mThe expression after translation is:
sb,m(t-TD,m)＝cos[φb,m(t-TD,m)+Δφb,m],TD,m≤t≤Tb,m+TD,m
in the formula, delta phib,mFor compensating the factor for the phaseThe function of the bionic signal segment is to eliminate the phase jump, delta phi, at the splicing position of two adjacent bionic signal segmentsb,mThe expression of (a) is:
Δφb,m＝φb,m-1(Tb,m-1)+…φb,1(Tb,1)
wherein, the first bionic signal segment sb,1(t) phase compensation factor Δ φb,1＝0；
(203) Substituting M translated bionic signal segments into an expression sN(t)＝sb,1(t-TD,1)+…+sb,m(t-TD,m)+…sb,M(t-TD,M) To obtain a segmented bionic signal sN(t)。
Preferably, the specific implementation steps of step (3) are as follows:
(301) for the final soundT(t) Short-time Fourier Transform (STFT) to obtain sT(t) time frequency spectrum, Xk[l]Is s isT(t) STFT result of kth data block, l represents the l spectrum of STFT, τkThe time corresponding to the first sampling point in the kth data block;
(302) at time τkSegment bionic signal sN(t) has a frequency of fkIn the frequency range [ fk-ΔB-,fk+ΔB+]In, find Xk[l]Peak value P ofk＝max|Xk(l)|；
(303) Bionic signal sbThe envelope of (t) is at time τkThe values of (d) are expressed as:
Ab(τk)＝KAPk
wherein, KAFor amplitude correction factors, by modifying KACan make sbThe envelope of (t) is close to sT(t) envelope.
Preferably, step (4) segments the bionic signal sN(t) and envelope Ab(t) substitution into the formula sb(t)＝Ab(t)·sN(t), then a biomimetic signal s can be obtainedb(t)；
If the signal sound has harmonic wave in the step (5), the corresponding bionic signal expression is as follows:
Figure BDA0002391859850000041
wherein s isb_1(t) represents the fundamental wave, sb_r(t) represents the subharmonic of R (R is more than or equal to 2 and less than or equal to R); sb_1(t) and sb_r(t) was constructed in accordance with steps (1) to (4).
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
(1) the invention can realize underwater communication and underwater detection with extremely strong concealment. By simulating the time-frequency spectrum contour and time-domain waveform enveloping characteristics of the cetacea animal tonal sound, the underwater communication or active sonar signals are disguised as the cetacea animal tonal sound, so that an enemy mistakenly considers that the received signals come from the cetacea animal and filters the signals, the effects of disguising the underwater communication and the active sonar device are achieved, the concealment of the underwater communication and the active sonar device is improved, and the probability of the underwater communication and the active sonar device discovered by the enemy is reduced.
(2) The invention has strong universality. By adjusting the harmonic quantity of the bionic signals, the quantity and time domain waveform envelope of the bionic signal segments of each harmonic and the time-frequency spectrum contour of each bionic signal segment, high-similarity simulation of various types of final sound and some whale animal cry with complex time-frequency spectrum contours is achieved.
(3) The invention can conveniently adjust the parameters of the bionic signal such as bandwidth, fundamental frequency, duration and the like, thereby improving the underwater communication and underwater detection performance of the bionic signal by properly changing the parameters of the bionic signal.
(4) The method establishes a mathematical model for the time-frequency spectrum contour of the cetacea animal final sound, and the time-frequency spectrum contour of the cetacea animal final sound is an important basis for classifying and identifying the cetacea animal population, so that the method provides help for the biological research of the cetacea animal.
Drawings
FIG. 1 shows waveforms and time-frequency diagrams of the signal sound of the cetacea, hexapod.
FIG. 2 shows f corresponding to different slope adjustment factors k in the present inventionPO(t) function diagram
FIG. 3a shows f corresponding to different curvature adjustment factors alpha in the present inventionPO(t) function diagram.
FIG. 3b shows f corresponding to different curvature adjustment factors alpha in the present inventionPX(t) function diagram.
FIG. 3c shows f for different curvature adjustment factors α in the present inventionPY(t) function diagram.
FIG. 3d shows f for different curvature adjustment factors α in the present inventionPZ(t) function diagram.
FIG. 4 shows f corresponding to different slope adjustment factors h in the present inventionSO(t) function diagram.
FIG. 5a shows f corresponding to different curvature adjustment factors β in the present inventionSO(t) function diagram.
FIG. 5b shows f for different curvature adjustment factors β in the present inventionSX(t) function diagram.
FIG. 5c shows f for different curvature adjustment factors β in the present inventionSY(t) function diagram.
FIG. 5d shows f for different curvature adjustment factors β in the present inventionSZ(t) function diagram.
FIG. 6 shows a time-frequency plot of a sinusoidal type final sound in the present invention.
Fig. 7 shows the time-frequency spectrum profile of three bionic signal segments in the invention.
Fig. 8a shows the waveform of a segmented biomimetic signal in the present invention.
FIG. 8b shows a time-frequency diagram of a segmented biomimetic signal in the present invention.
FIG. 9 shows the waveform and envelope of a sinusoidal type final sound in the present invention.
Figure 10a shows the waveform of a sinusoidal biomimetic signal in the present invention.
Fig. 10b shows a time-frequency diagram of a sinusoidal bionic signal in the invention.
Fig. 11a shows a waveform of constant frequency type final sound in the present invention.
Fig. 11b shows a waveform of a constant frequency type bionic signal in the present invention.
FIG. 11c is a time-frequency diagram of the constant frequency type final sound in the present invention.
FIG. 11d is a time-frequency diagram of a constant frequency type bionic signal according to the present invention.
Fig. 12a shows a waveform of a tone-up type final sound in the present invention.
Fig. 12b shows the waveform of the up-modulated type bionic signal in the present invention.
FIG. 12c shows a time-frequency diagram of a frequency up modulated type final sound in the present invention.
Fig. 12d shows a time-frequency diagram of the up-modulated type bionic signal in the invention.
Fig. 13a shows a waveform of a lower frequency modulated type final sound in the present invention.
Fig. 13b shows the waveform of the lower frequency modulation type bionic signal in the invention.
FIG. 13c shows a time-frequency diagram of the lower frequency modulated type final sound in the present invention.
Fig. 13d shows a time-frequency diagram of a lower frequency modulation type bionic signal in the invention.
Fig. 14a shows the waveform of the concave type final sound in the present invention.
Fig. 14b shows the waveform of the concave bionic signal in the invention.
FIG. 14c shows a time-frequency plot of the concave type of signal sound in the present invention.
FIG. 14d shows a time-frequency diagram of concave bionic signals in the invention.
Fig. 15a shows a waveform of the convex type final sound in the present invention.
Fig. 15b shows a waveform of a convex-shaped biomimetic signal in the present invention.
FIG. 15c shows a time-frequency diagram of the convex type final sound in the present invention.
FIG. 15d shows a time-frequency diagram of a convex biomimetic signal in the present invention.
Fig. 16a shows a waveform of signature while in the present invention.
Fig. 16b shows a waveform of a signature while bionic signal in the present invention.
Fig. 16c shows a time-frequency diagram of signature while in the present invention.
Fig. 16d shows a time-frequency diagram of a signature while bionic signal in the invention.
Detailed Description
The invention is described in further detail below with reference to the figures and specific examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a method for synthesizing and modifying a vocalization of a whale animal, which can meet the requirements of effectiveness and concealment of bionic camouflage concealed underwater sound detection and communication.
In fig. 1, the present invention provides waveforms and time-frequency diagrams of six types of cetacea, which can be divided into six types, i.e., a constant frequency type 1, an up-modulation type 2, a down-modulation type 3, a concave type 4, a convex type 5, and a sinusoidal type 6, according to the time-frequency spectrum profile of the cetacea. Wherein the frequency of the constant frequency type 1 does not substantially change with time; the frequency of the up-modulation type 2 increases as time increases; the frequency of the lower sweep type 3 decreases with increasing time; the frequency of the concave form 4 decreases and then increases; the frequency of the lobes 5 increases and then decreases; the frequency of the sinusoid 6 increases then decreases then increases and so on, or decreases then increases then decreases and so on.
Statistical analysis is carried out on the cetacea animal, and it is found that the time-frequency spectrum profile of the cetacea has complex frequency modulation characteristics, the frequency modulation characteristics of the frequency spectrum profiles are different when different types of the cetacea are used, and the frequency modulation characteristics of different parts in one type of the cetacea may also be different, but the characteristics can be divided into two types: the first type is that the absolute value of the slope of the time-frequency spectrum contour changes monotonously, such as monotonously increases; the second type is that the absolute value of the slope of the spectral profile increases and then decreases.
In order to realize high-precision matching of frequency spectrum outlines of cetacea animals during the tonal sound detection and meet the requirements of bionic camouflage hidden underwater sound detection and communication camouflage, the invention designs a method for synthesizing the tonal sound by splicing multiple sections of bionic signals. A total sound sT(t) the corresponding biomimetic signal is sb(t) (ii) a Is constructed of sb(t) first, s isT(t) dividing into M (M is a positive integer not less than 1) segments, and then simulating s divided into M segments using M bionic signal segmentsTAnd (t) finally splicing the M bionic signal sections end to end. S comprising M biomimetic signal segmentsbThe expression of (t) is:
sb(t)＝Ab(t)·sN(t) (1)
wherein A isb(t) is the envelope, sN(t) is a segmented biomimetic signal, the amplitude of which is normalized, sNThe expression of (t) is:
sN(t)＝sb,1(t-TD,1)+…+sb,m(t-TD,m)+…sb,M(t-TD,M) (2)
wherein s isb,m(t) is sb(t) M (M ═ 1,2, …, M) number of biomimetic signal segments, sb,m(T) has a duration of Tb,m. Time delay TD,mRepresenting a set of signals sb,1(t),…,sb,m-1(t) the sum of the durations of all signals, expressed as:
TD,m＝Tb,m-1+…+Tb,1 (3)
wherein, the first bionic signal segment sb,1The delay of (T) is typically set to T D,10, i.e. sb,1(t-TD,1)＝sb,1(t)。
Due to the bionic signal s obtained as described aboveb(t) parameterizing the final sound by modifying sbAnd (t) modifying the final sound by using parameters such as frequency range, period and the like so as to meet the validity requirements of bionic camouflage hidden underwater sound detection and communication.
The specific implementation steps for synthesizing and modifying the cetacea animal final sound are as follows:
the first step is to construct a biomimetic signal sbAnd (t) time-frequency expression of each bionic signal segment.
Firstly, aiming at the time-Frequency spectrum outline characteristics of the final sound, the invention designs two bionic signal models, namely Power Frequency Modulation bionic Bionic, PFMB) signal model, and Sinusoidal Frequency Modulation Bionic (SFMB) signal model. Each of the PFMB signal model and the SFMB signal model includes 4 sub-signal models, sbEach biomimetic signal segment in (t) is constructed based on the 8 sub-signal models.
Further, a time-frequency expression f of 4 sub-PFMB signal modelsPO(t)、fPX(t)、fPY(t) and fPZ(t) are respectively:
Figure BDA0002391859850000071
Figure BDA0002391859850000081
Figure BDA0002391859850000082
Figure BDA0002391859850000083
where 0 ≦ T ≦ T, T represents the signal duration, B is the signal bandwidth, fCFor the signal carrier frequency, the curvature is adjusted by a factor alpha (alpha)>0) The function of the slope adjusting factor k (k is more than or equal to 0 and less than or equal to B/T) is to adjust the slope of the time-frequency expression at the position where T is 0 and T is T;
f corresponding to different slope regulating factors kPOAs shown in fig. 2, the function graph of (t) shows that f is 0kHz/s, 8 is 5kHz/s, and 9 is 10kHz/sPO(T) the slope at T-0 and T-T, continuously varying as k varies;
f corresponding to different curvature adjusting factors alphaPO(t) is shown in FIG. 3a, where f is the function of the different curvature adjustment factors αPX(t) is shown in FIG. 3b, different curvature adjustmentsF for factor alphaPY(t) is shown in FIG. 3c, where f is the function of the different curvature adjustment factors αPZThe function graph of (t) is shown in FIG. 3 d. From the time-frequency curve 10 when α is 0.5, the time-frequency curve 11 when α is 1, and the time-frequency curve 12 when α is 2 in fig. 3a, f can be seenPO(t) the curvature varies continuously with a; as can be seen from fig. 3a, 3b, 3c and 3d, fPX(t) and fPO(t) about line f (t) ═ B/2+ fCSymmetry, fPY(t) and fPO(T) symmetrical about a straight line T-T/2, fPZ(t) and fPO(T) about point (T/2, B/2+ f)C) Symmetry;
further, the expression of the four sub-PFMB signal models is:
Figure BDA0002391859850000084
Figure BDA0002391859850000085
Figure BDA0002391859850000086
Figure BDA0002391859850000087
further, time-frequency expression f of four sub-SFMB signal modelsSO(t)、fSX(t)、fSY(t) and fSZ(t) are respectively:
Figure BDA0002391859850000088
Figure BDA0002391859850000089
Figure BDA0002391859850000091
Figure BDA0002391859850000092
the curvature adjusting factor beta (beta >0) is used for adjusting the curvature of the time-frequency expression, and the slope adjusting factor h (h is more than or equal to 0 and less than or equal to B/T) is used for adjusting the slope of the time-frequency expression at T-0 and T-T;
f corresponding to different slope regulating factors hSOThe function graph of (t) is shown in fig. 4, and from the time-frequency curve 13 when h is 0kHz/s, the time-frequency curve 14 when h is 5kHz/s, and the time-frequency curve 15 when h is 10kHz/s, f can be seenSO(T) a slope at T-0 and T-T that varies continuously with h;
f corresponding to different curvature regulating factors betaSO(t) is shown in FIG. 5a, where f is the function of the different curvature adjustment factors βSX(t) is shown in FIG. 5b, where f is the function of the different curvature adjustment factors βSY(t) is shown in FIG. 5c, where f is the function of the different curvature adjustment factors βSZ(t) is shown in FIG. 5 d;
from the time-frequency curve 16 when β is 0.5, the time-frequency curve 17 when β is 1, the time-frequency curve 18 when β is 2 and the time-frequency curve 19 when β is 4 in fig. 5a, f can be seenSO(t) the curvature varies continuously with β;
as can be seen from fig. 5a, 5b, 5c and 5d, fSX(t) and fSO(t) about line f (t) ═ B/2+ fCSymmetry, fSY(t) and fSO(T) symmetrical about a straight line T-T/2, fSZ(t) and fSO(T) about point (T/2, B/2+ f)C) Symmetry;
further, the expressions of the four sub-SFMB signal models are:
Figure BDA0002391859850000093
Figure BDA0002391859850000094
Figure BDA0002391859850000095
Figure BDA0002391859850000096
by modifying the parameters in the PFMB signal model and the SFMB signal model, the method comprises the following steps: signal duration T, signal bandwidth B, signal carrier frequency fCA curvature regulating factor alpha, a curvature regulating factor beta, a slope regulating factor k and a slope regulating factor h, so that the bionic signal sb(t) in the presence of a terminal soundT(t) increasing the biomimetic signal s while being similarb(t) performance of underwater exploration or communication;
further, to ensure sb(t) and sT(t) high similarity between them, sbThe smoothness of the time-frequency spectrum profile of (t) should be equal to sT(t) is coincident, i.e. at sbAnd (t) at the connection position of two adjacent bionic signal sections, the frequency and the slope of the time-frequency spectrum contour should be continuously changed. In particular, for s as described above containing M biomimetic signal segmentsb(t), if M is more than or equal to 2, the 1 st bionic signal segment sb,1(t)(0≤t≤Tb,1) Has a time-frequency spectrum profile of fb,1(t), 2 nd bionic signal segment sb,2(t)(0≤t≤Tb,2) Has a time-frequency spectrum profile of fb,2(t), then analogizing the above steps, M (M is more than or equal to 1 and less than or equal to M) th bionic signal segment sb,m(t)(0≤t≤Tb,m) Has a time-frequency spectrum profile of fb,m(t) of (d). F is thenb,1(t) and fb,2(t) the condition must be satisfied: f. ofb,1(Tb,1)＝fb,2(0) And is
Figure BDA0002391859850000101
Then use it toBy analogy, fb,m(t) and fb,m+1(t) the condition must be satisfied: f. ofb,m(Tb,1)＝fb,m+1(0) And is
Figure BDA0002391859850000102
In the simulation, fig. 6 shows a time-frequency diagram of a sinusoidal type final sound, and based on a time-frequency spectrum contour 20 of the sinusoidal type final sound in fig. 6, time-frequency spectrum contours of three bionic signal segments in fig. 7 are constructed, wherein a first contour segment 21, a middle contour segment 22 and a last contour segment 23 are respectively based on the time-frequency expression fSO(t)、fSY(t) and fSO(t) is constructed.
The second step is to construct a segmented bionic signal sN(t)。
sN(t) mth bionic signal segment sb,mThe expression of (t) is:
Figure BDA0002391859850000103
wherein phib,m(t) is a phase expression, fb,m(t) is constructed based on a time-frequency expression of one sub-PFMB signal or sub-SFMB signal;
further, each bionic signal segment is translated on a time axis, and the mth bionic signal segment sr,m(T) time of translation is TD,mThe expression after translation is:
sb,m(t-TD,m)＝cos[φb,m(t-TD,m)+Δφb,m],TD,m≤t≤Tb,m+TD,m (21)
in the formula, delta phib,mIs a phase compensation factor, and has the function of eliminating the phase mutation, delta phi, at the splicing part of two adjacent bionic signal segmentsb,mThe expression of (a) is:
Δφb,m＝φb,m-1(Tb,m-1)+…φb,1(Tb,1) (22)
wherein, the first bionic signal segment sb,1(t) phase compensation factor Δ φb,1＝0；
Further, substituting M translated bionic signal segments into an expression sN(t)＝sb,1(t-TD,1)+…+sb,m(t-TD,m)+…sb,M(t-TD,M) Then a segmented bionic signal s can be obtainedN(t) of (d). In the simulation, based on the time-frequency spectrum profiles of the three bionic signal segments in fig. 7, the waveform of the segmented bionic signal shown in fig. 8a and the time-frequency diagram of the segmented bionic signal shown in fig. 8b are obtained.
And the third step is to extract the envelope of the final sound and take the envelope as the envelope of the bionic signal.
First, for the final soundT(t) Short-time Fourier Transform (STFT) to obtain sT(t) time frequency spectrum. Xk[l]Is s isT(t) STFT result of kth data block, l represents the l spectrum of STFT, τkThe time corresponding to the first sampling point in the kth data block;
further, at time τkSegment bionic signal sN(t) has a frequency of fkIn the frequency range [ fk-ΔB-,fk+ΔB+]In, find Xk[l]Peak value P ofk＝max|Xk(l)|；
Further, a bionic signal sbThe envelope of (t) is at time τkThe values of (d) are expressed as:
Ab(τk)＝KAPk (23)
wherein, KAFor amplitude correction factors, by modifying KACan make sbThe envelope of (t) is close to sT(t) envelope.
In the simulation, the STFT uses a Hamming window of N points,
Figure BDA0002391859850000111
fig. 9 shows the waveform and envelope of a sinusoidal type final sound, wherein the envelope 25 of the sinusoidal type final sound is extracted from the waveform 24 of the sinusoidal type final sound.
The fourth step is to synthesize a bionic signal sb(t)。
Segment bionic signal sN(t) and envelope Ab(t) is substituted into the formula (1), then the bionic signal s can be obtainedb(t) of (d). In the simulation, the waveform of the segmented bionic signal in fig. 8a and the envelope 25 of the sinusoidal type final sound in fig. 9 are used to obtain the waveform of the sinusoidal type bionic signal shown in fig. 10a and the time-frequency diagram of the sinusoidal type bionic signal shown in fig. 10 b.
And fifthly, if the final sound has harmonic waves, the corresponding bionic signal expression is as follows:
Figure BDA0002391859850000112
wherein s isb_1(t) represents the fundamental wave, sb_r(t) represents the subharmonic of R (2. ltoreq. R. ltoreq.R). sb_1(t) and sb_r(t) are constructed in accordance with the methods of the first to fourth steps described above.
Further, this example simulates the remaining 5 types of final sound, including five types, i.e., constant frequency type 1, up-modulation type 2, down-modulation type 3, concave type 4, and convex type 5;
FIG. 11a is a waveform of a constant frequency type final sound, FIG. 11b is a waveform of a constant frequency type bionic signal, FIG. 11c is a time-frequency diagram of a constant frequency type final sound, and FIG. 11d is a time-frequency diagram of a constant frequency type bionic signal;
fig. 12a is a waveform of the up-modulated type final sound, fig. 12b is a waveform of the up-modulated type bionic signal, fig. 12c is a time-frequency diagram of the up-modulated type final sound, and fig. 12d is a time-frequency diagram of the up-modulated type bionic signal;
fig. 13a is a waveform of a lower-frequency-modulation type final sound, fig. 13b is a waveform of a lower-frequency-modulation type bionic signal, fig. 13c is a time-frequency diagram of a lower-frequency-modulation type final sound, and fig. 13d is a time-frequency diagram of a lower-frequency-modulation type bionic signal;
FIG. 14a is a waveform of a concave type tonal sound, FIG. 14b is a waveform of a concave type bionic signal, FIG. 14c is a time-frequency diagram of a concave type tonal sound, and FIG. 14d is a time-frequency diagram of a concave type bionic signal;
fig. 15a is a waveform of a convex type final sound, fig. 15b is a waveform of a convex type bionic signal, fig. 15c is a time-frequency diagram of a convex type final sound, and fig. 15d is a time-frequency diagram of a convex type bionic signal;
in summary, in addition to the final sound, cetacea produce frequency modulated sounds of long duration and more complex spectral profiles, such as "signature whistle". Since these whale sounds can be divided into a plurality of simple frequency modulated sounds, the present invention can also be used to synthesize these whale sounds with complex time-frequency spectral profiles. Fig. 16a is a waveform of a signature while fig. 16b is a waveform of a signature while fig. 16c is a time-frequency diagram of a signature while fig. 16d is a time-frequency diagram of a signature while bionic signal, wherein the waveform of the signature while bionic signal is composed of 21 bionic signals.
The present invention is not limited to the above-described embodiments. The foregoing description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention, and the above specific embodiments are merely illustrative and not restrictive. Those skilled in the art can make many changes and modifications to the invention without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (3)

1. A method for synthesizing and modifying vocals of cetacea is characterized in that a final sound of the cetacea is set as sT(t) the corresponding biomimetic signal is sb(t) to construct sb(t) mixing sT(t) is divided into M sections, M is a positive integer greater than or equal to 1, and M bionic signal sections are used for simulating s divided into M sectionsT(t); finally, splicing the M bionic signal sections end to obtain sb(t) by modifying s after splicingb(t) the modification of the final sound is realized by the parameter of (t) so as to meet the validity requirements of bionic camouflage hidden underwater sound detection and communication; the method specifically comprises the following steps:
(1) construction of a biomimetic signal sb(t) time-frequency expressions of each bionic signal segment; the concrete implementation steps of the step (1) are as follows:
(101) setting up two biomimetic signal models, i.e.A power frequency modulation bionic PFMB signal model and a sine frequency modulation bionic SFMB signal model; each of the PFMB signal model and the SFMB signal model includes 4 sub-signal models, a bionic signal sb(t) each biomimetic signal segment in (t) is constructed based on the 8 sub-signal models;
(102) obtaining expressions of four sub-PFMB signal models according to the time-frequency expressions and the function graphs of the 4 sub-PFMB signal models;
(103) obtaining expressions of four sub SFMB signal models according to the time-frequency expressions and the function graphs of the 4 sub SFMB signal models;
(104) by modifying the parameters in the PFMB signal model and the SFMB signal model, the method comprises the following steps: signal duration T, signal bandwidth B, signal carrier frequency fCA curvature regulating factor alpha, a curvature regulating factor beta, a slope regulating factor k and a slope regulating factor h, so that the bionic signal sb(t) in the presence of a terminal soundT(t) increasing the biomimetic signal s while being similarb(t) performance of underwater exploration or communication;
(105) to ensure sb(t) and sT(t) high similarity between them, sbThe smoothness of the time-frequency spectrum profile of (t) should be equal to sT(t) is coincident, i.e. at sb(t) where two adjacent bionic signal segments are connected, the frequency and the slope of the time-frequency spectrum contour should be kept continuously changed;
(2) construction of a segmented biomimetic signal sN(t)；
(3) Extracting an envelope of the final sound, and taking the envelope as an envelope of the bionic signal;
(4) synthesizing a bionic signal sb(t); segment bionic signal sN(t) and envelope Ab(t) substitution into the formula sb(t)＝Ab(t)·sN(t), then a biomimetic signal s can be obtainedb(t)；
(5) If the tonal sound has harmonic waves, repeating the steps (1) to (4) until the modification is finished; if the signal sound has harmonic, the corresponding bionic signal expression is as follows:
Figure FDA0003057919770000021
wherein s isb_1(t) represents the fundamental wave, sb_r(t) represents the subharmonic of R (R is more than or equal to 2 and less than or equal to R); sb_1(t) and sb_r(t) was constructed in accordance with steps (1) to (4).
2. The method for synthesizing and modifying vocals of animals in the order of cetacea as claimed in claim 1, wherein the step (2) is implemented by the following steps:
(201)sN(t) mth bionic signal segment sb,mThe expression of (t) is:
Figure FDA0003057919770000022
wherein phib,m(t) is a phase expression, fb,m(t) is constructed based on a time-frequency representation of a sub-PFMB signal or sub-SFMB signal, sb,m(t) is sb(t) mth bionic signal segment, sb,m(T) has a duration of Tb,m(ii) a Wherein M is 1,2, …, M;
(202) translating each bionic signal segment on a time axis to form an m-th bionic signal segment sb,m(T) time of translation is TD,mThe expression after translation is:
sb,m(t-TD,m)＝cos[φb,m(t-TD,m)+Δφb,m],TD,m≤t≤Tb,m+TD,m
in the formula, delta phib,mIs a phase compensation factor, and has the function of eliminating the phase mutation, delta phi, at the splicing part of two adjacent bionic signal segmentsb,mThe expression of (a) is:
Δφb,m＝φb,m-1(Tb,m-1)+…φb,1(Tb,1)
wherein, the first bionic signal segment sb,1(t) phase compensation factor Δ φb,1＝0；
(203) Substituting M translated bionic signal segments into an expression sN(t)＝sb,1(t-TD,1)+…+sb,m(t-TD,m)+…sb,M(t-TD,M) To obtain a segmented bionic signal sN(t)。
3. The method for synthesizing and modifying vocals of animals in the order of cetacea as claimed in claim 1, wherein the step (3) is implemented by the following steps:
(301) for the final soundT(t) Short-time Fourier Transform (STFT) to obtain sT(t) time frequency spectrum, Xk[l]Is s isT(t) STFT result of kth data block, l represents the l spectrum of STFT, τkThe time corresponding to the first sampling point in the kth data block;
(302) at time τkSegment bionic signal sN(t) has a frequency of fkIn the frequency range [ fk-ΔB-,fk+ΔB+]In, find Xk[l]Peak value P ofk＝max|Xk(l)|；
(303) Bionic signal sbThe envelope of (t) is at time τkThe values of (d) are expressed as:
Ab(τk)＝KAPk
wherein, KAFor amplitude correction factors, by modifying KACan make sbThe envelope of (t) is close to sT(t) envelope.
CN202010117205.3A 2020-02-25 2020-02-25 Synthetic and modification method for vocals of cetacea animals Active CN111431625B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010117205.3A CN111431625B (en) 2020-02-25 2020-02-25 Synthetic and modification method for vocals of cetacea animals

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010117205.3A CN111431625B (en) 2020-02-25 2020-02-25 Synthetic and modification method for vocals of cetacea animals

Publications (2)

Publication Number Publication Date
CN111431625A CN111431625A (en) 2020-07-17
CN111431625B true CN111431625B (en) 2021-07-27

Family

ID=71547150

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010117205.3A Active CN111431625B (en) 2020-02-25 2020-02-25 Synthetic and modification method for vocals of cetacea animals

Country Status (1)

Country Link
CN (1) CN111431625B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112182868B (en) * 2020-09-21 2023-03-24 天津大学 Shallow sea active sonar detection method based on whale/dolphin click sound calling
CN112491765B (en) * 2020-11-19 2022-08-12 天津大学 CPM modulation-based identification method for whale-imitating animal whistle camouflage communication signal
CN112702125B (en) * 2020-12-16 2022-06-03 天津大学 Communication signal recognition method imitating whale cry based on linear frequency modulation signal modulation
CN112953652B (en) * 2021-01-27 2022-04-15 浙江大学 Whale whistle-like communication method based on segmented time-frequency contour time delay modulation
CN113259016B (en) * 2021-04-23 2022-09-13 西北工业大学 Underwater sound concealed communication method based on dolphin whistle parameter model

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105391501A (en) * 2015-10-13 2016-03-09 哈尔滨工程大学 Dolphin click simulating underwater acoustic communication method based on time-frequency spectrum translation

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9490911B2 (en) * 2013-03-15 2016-11-08 Fairfield Industries Incorporated High-bandwidth underwater data communication system
CN104217722B (en) * 2014-08-22 2017-07-11 哈尔滨工程大学 A kind of dolphin whistle signal time-frequency spectrum contour extraction method
CN105227246B (en) * 2015-10-13 2017-10-31 哈尔滨工程大学 It is a kind of to utilize the underwater acoustic communication method for being segmented the imitative dolphin whistle signal of LFM signals
CN106789794B (en) * 2016-11-30 2019-05-28 厦门大学 A kind of bionical communication means modulated using dolphin signal of communication frequency modulation
CN106899357B (en) * 2017-03-13 2020-04-03 哈尔滨工程大学 Disguised and concealed underwater communication device simulating dolphin whistle
CN109347568B (en) * 2018-09-05 2021-04-20 哈尔滨工程大学 Dolphin whistle-imitating continuous phase multi-element frequency modulation underwater acoustic communication method

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105391501A (en) * 2015-10-13 2016-03-09 哈尔滨工程大学 Dolphin click simulating underwater acoustic communication method based on time-frequency spectrum translation

Also Published As

Publication number Publication date
CN111431625A (en) 2020-07-17

Similar Documents

Publication Publication Date Title
CN111431625B (en) Synthetic and modification method for vocals of cetacea animals
CN109800700B (en) Underwater acoustic signal target classification and identification method based on deep learning
CN110111803B (en) Transfer learning voice enhancement method based on self-attention multi-kernel maximum mean difference
CN106503336B (en) A kind of modeling of dolphin ticktack acoustical signal and synthetic method
CN109887496A (en) Orientation confrontation audio generation method and system under a kind of black box scene
Wang et al. ia-PNCC: Noise Processing Method for Underwater Target Recognition Convolutional Neural Network.
CN105869651A (en) Two-channel beam forming speech enhancement method based on noise mixed coherence
CN110109058A (en) A kind of planar array deconvolution identification of sound source method
CN110456332A (en) A kind of underwater sound signal Enhancement Method based on autocoder
Jiang et al. Synthesis and modification of cetacean tonal sounds for underwater bionic covert detection and communication
Jiang et al. Disguised bionic sonar signal waveform design with its possible camouflage application strategy for underwater sensor platforms
CN113049080A (en) GDWC auditory feature extraction method for ship radiation noise
CN112382301A (en) Noise-containing voice gender identification method and system based on lightweight neural network
Shi et al. Fusion feature extraction based on auditory and energy for noise-robust speech recognition
CN107332642B (en) Low probability of intercept waveform design method based on the synthesis of ocean mammal signal
CN107635181B (en) Multi-address sensing source feedback optimization method based on channel learning
Li et al. Performance analysis for focused beamformers in passive underwater acoustic localization
CN116705056A (en) Audio generation method, vocoder, electronic device and storage medium
CN104407341A (en) Camouflaged and hidden underwater sound source device and method using whale sound and simulating sounding characteristics of whale
Guo et al. Underwater target detection and localization with feature map and CNN-based classification
CN115561739A (en) Underwater acoustic signal counterfeiting method based on WaveGAN
Wu et al. Low-frequency Underwater Acoustic Signal Denoising Method in the Shallow Sea with a Low Signal-to-noise Ratio
CN113409817B (en) Audio signal real-time tracking comparison method based on voiceprint technology
Stark et al. Single channel speech separation using source-filter representation
Xiang et al. A weekly supervised speech enhancement strategy using cycle-gan

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant