CN100421151C - Adaptive multistage stepping sequence switch method - Google Patents
Adaptive multistage stepping sequence switch method Download PDFInfo
- Publication number
- CN100421151C CN100421151C CNB031524303A CN03152430A CN100421151C CN 100421151 C CN100421151 C CN 100421151C CN B031524303 A CNB031524303 A CN B031524303A CN 03152430 A CN03152430 A CN 03152430A CN 100421151 C CN100421151 C CN 100421151C
- Authority
- CN
- China
- Prior art keywords
- index value
- correlation
- value
- maximum
- pairing
- 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.)
- Expired - Lifetime
Links
Images
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Signal Processing For Digital Recording And Reproducing (AREA)
Abstract
The present invention provides an adaptive multistage stepping timing conversion method which is used for synthesizing S1 <n> and S2 <n> into S3<n>. The method comprises the following steps: (a) a first correlation value of a first index value corresponding to the S1 <n> and the S2 <n> is calculated; (b), the first correlation value is compared with a critical value; (c) if the first correlation value is smaller than the critical value, a correlation value corresponding to an index value in first number after the S1 <n> and the S2 <n> corresponding to the first index value is calculated; if the first correlation value is larger than the critical value, a correlation value corresponding to an index value in second number after the S1 <n> and the S2 <n> corresponding to the first index value is calculated; (d) S3 <n> is generated according to the maximal index value, the S1 <n> and the S2 <n> corresponding to the calculated maximal correlation value.
Description
Technical field
The invention provides a kind of signal synthesis method, particularly a kind of multistage stepping method of adaptability that is applied to sequential conversion (time scaling).
Background technology
Along with the progress of science and technology, some are also more and more as the function that the video-audio playing device of Karaoke and so on can provide, and for example similarly are that audio purifies (audio clean-up), dreamlike sound field (dream), reaches sequential conversion functions such as (time scaling).So-called sequential conversion (being called time stretching, time compression/expansion or timecorrection again) is under the situation that does not influence tone (pitch), change the length of sound signal, promptly change the playback rate (tempo) of this sound signal.
At present, AV device on the market mostly by three kinds of following methods to finish sequential conversion, a kind of is that phase vocoder (Phase Vocoder), a kind of MPEX of being (MinimumPerceived Loss Time Expansion/Compression) another kind then are TimeDomain Harmonic Scaling (TDHS).The mode that Phase vocoder utilizes STFT (ShortTime Fourier Transform) earlier converts sound signal the frequency-region signal (complex Fourier representation) of Fourier form to, utilizes the mode of interpolation and iSTFT (contrary STFT) this frequency-region signal to be converted to the sound signal of changing (time scaled) corresponding to the sequential of this sound signal again.MPEX is developed by Prosoniq recently, and MPEX is a kind of method of simulating human auditory properties, is similar to artificial neural network (artificial neural network).MPEX is according to the sound signal of being included in the specific period, and and then " study " this specific period in the various characteristics of sound signal, to attempt prolongation or to shorten this sound signal.TDHS then is the method that a kind of more general sequential is changed, each correlation (magnitudes of a autocorrelation function) in the correlation table (autocorrelogram) of its elder generation's calculating first sound signal, then postpone this first sound signal to produce second sound signal according to the pairing maximum index value of the maximum related value in this correlation table, and then with this first sound signal with synchronous stack (synchronized overlap-add, SOLA) mode is replicated on this second sound signal, to produce the 3rd sound signal longer than first sound signal.
See also Fig. 1, Fig. 1 is the correlation table 10 of existing TDHS, and correlation table 10 comprises a plurality of correlation R (τ).In general, except maximum related value 12 and near correlation thereof are big, remaining correlation is all very little in the correlation table 10, and the variation of two adjacent correlations is also not too big in the correlation table 10, just, if first correlation 14 is little far beyond maximum related value 12, then also can be much smaller than maximum related value 12 adjacent to second correlation 16 of first correlation 14, accordingly, second correlation, the 16 pairing second index value τ
2Also can be apart from maximum related value 12 pairing index value τ
MaxFar; Otherwise, if third phase pass value 18 is with the difference of 12 of maximum related value when little, then just may be than near maximum related value 12 adjacent to the 4th correlation 20 of third phase pass value 18, accordingly, the 4th correlation 20 pairing the 4th index value τ
4May (for one group in two groups of third phase pass values 18 among Fig. 1 and the 4th correlation 20) will approach maximum index value τ
Max
Correlation table 10 is set up by digital signal processor (DSP), and DSP is special in handling as convolutional calculation (convolution), fast fourier transform (fast Fouriertransform, the usefulness of complex mathematical computing such as FFT).Even so, in order to find out maximum related value 12 and pairing maximum index value τ thereof
Max, and it is not only tediously long but also there is no need fully to use DSP to calculate the process of all correlations in the correlation table 10.
Summary of the invention
Therefore fundamental purpose of the present invention is to provide the sequential conversion of the multistage stepping of a kind of adaptability, in the hope of finding out corresponding to S apace
1[n] and S
2The maximum index value τ of [n]
Max, with synthetic S
1[n] and S
2[n].
According to an aspect of the present invention, the present invention discloses the sequential conversion of the multistage stepping of a kind of adaptability, is used for S
1[n] and S
2[n] synthesizes S
3[n], this method comprises the following step: (a) calculate S
1[n] and S
2[n] is corresponding to first correlation of first index value; (b) relatively this first correlation and critical value; (c) if this first correlation less than this critical value, then calculates S
1[n] and S
2[n] is corresponding to the pairing correlation of first number index value after this first index value; If this first correlation greater than this critical value, then calculates S
1[n] and S
2[n] is corresponding to the pairing correlation of second number index value after this first index value; And (d) according to the pairing maximum index value of maximum related value, the S that calculate
1[n] and S
2[n] produces S
3[n].
In a preferred embodiment of the invention, this first number is greater than 1, and this second number equals 1.
According to another aspect of the present invention, the present invention discloses the sequential conversion of the multistage stepping of a kind of adaptability, is used for S
1[n] and S
2[n] synthesizes S
3[n], this method comprises the following step: (a) with S
2[n] postpones predetermined number to form S
5[n]; (b) calculate S
1[n] and S
5[n] is corresponding to first correlation of first index value; (c) relatively this first correlation and a critical value; (d) if this first correlation less than this critical value, then calculates S
1[n] and S
5[n] is corresponding to the pairing correlation of index value of first number after this first index value; If this first correlation greater than this critical value, then calculates S
1[n] and S
5[n] is corresponding to the pairing correlation of index value of second number after this first index value; And (e) according to the pairing maximum index value of maximum related value, the S that calculate
1[n] and S
5[n] produces S
3[n].
Because method of the present invention is relevant to S in foundation
1[n] and S
2During the correlation table of [n], do not need to calculate one by one correlations all in this correlation table, therefore can save the time that the DSP that is used for setting up this correlation table calculates the required cost of this correlation, so, the calculating usefulness of the computing machine at this DSP place also increased.
Description of drawings
Fig. 1 is the correlation table of existing TDHS.
Fig. 2 is the correlation table of the inventive method.
Fig. 3 is the process flow diagram of the inventive method.
Fig. 4 is that the inventive method is with S
1[n] and S
2[n] synthesizes S
3The synoptic diagram of [n].
Fig. 5 increases the synoptic diagram of sound signal for the inventive method.
Fig. 6 shortens the synoptic diagram of sound signal for the inventive method.
Description of reference numerals
10,30 correlation tables, 12 maximum related value
14 first correlations, 16 second correlations
18 third phase passes are worth 20 the 4th correlations
Th
1The first critical value Th
2Second critical value
Th
3The 3rd critical value
Embodiment
In the process of foundation corresponding to the correlation table of first sound signal and second sound signal, the method 100 in the preferred embodiments of the present invention is according to the pairing correlation of index value in this correlation table and the first critical value th
1And the second critical value th
2Between magnitude relationship, the first critical value th wherein
1Be less than the second critical value th
2, calculate the pairing correlation of index value that is positioned in this correlation table behind this index value.Know clearly it, if the first correlation R (τ in this correlation table
1) less than the first critical value th
1, represent the first correlation R (τ
1) the pairing first index value τ
1Maximum related value R (τ in this correlation table
Max) pairing maximum index value τ
MaxStill have a segment distance, then calculating is positioned at the first index value τ
1After the first predetermined number Δ
1The second index value τ
2The pairing second correlation R (τ
2); If third phase pass value R (τ in this correlation table
3) greater than the first critical value th
1But less than the second critical value th
2, represent third phase pass value R (τ
3) pairing the 3rd index value τ
3Than the first index value τ
1More near maximum index value τ
Max, then calculating is positioned at the 3rd index value τ
3After the second predetermined number Δ
2The 4th index value τ
4Pairing the 4th correlation R (τ
4), the second predetermined number Δ wherein
2Less than the first predetermined number Δ
1If the 5th correlation R (τ in this correlation table
5) greater than the second critical value th
2, represent the 5th correlation R (τ
5) correspondence the 5th index value τ
5Quite near maximum index value τ
Max, then calculate and be next to the 5th index value τ
5After the 6th index value τ
6Pairing the 6th correlation R (τ
6).
See also Fig. 2 and Fig. 3, Fig. 2 is the method 100 pairing correlation tables 30 in the preferred embodiments of the present invention, and Fig. 3 is the process flow diagram of method 100 of the present invention.Method 100 comprises the following step:
Step 102: beginning;
(S
1[n] and S
2[n] will be synthesized and be S
3[n], for convenience of description for the purpose of, suppose S
1[n] and S
2[n] all comprises N signal, certainly S
1[n] and S
2The number of the signal that [n] comprised also can be inequality)
Step 103: with S
2[n] postpones the predetermined number Δ to form S
5[n];
(optical read head (pickuphead) in video-audio playing device is reading S
3The phenomenon of reading of data deficiency (run-in) takes place when [n], so method of the present invention 100 is earlier with S
2After [n] postpones predetermined number, just calculate synthetic S
1[n] and S
2The maximum index value τ that [n] is required
Max, in the present embodiment, the predetermined number Δ equals [N/3])
Step 104: calculate S
1[n] and S
5[n] is corresponding to opening beginning index value τ
1(τ=1) open the beginning correlation R (1), will differentiate correlation R
cSet for and open beginning correlation R (1), and will be corresponding to differentiating correlation R
cDifferentiation index value τ
cSet for and open beginning index value τ
1
Step 106: if (τ
c=N-1), then carry out step 200, otherwise carry out step 108; (if τ
c=N-1 represents R
cBe last correlation in the correlation table 30, correlation table 30 has been set up and has been finished)
Step 108: relatively differentiate correlation R
cWith the first critical value th
1With the second critical value th
2Between size, if differentiate correlation R
cLess than the first critical value th
1(as the R among Fig. 2 (1)) then carry out step 110, if differentiate correlation R
cBetween the first critical value th
1With the second critical value th
2Between (as the τ among Fig. 2
iPairing correlation R (τ
i)), then carry out step 140, (as the τ among Fig. 2
iPairing correlation), if differentiate correlation R
cGreater than the second critical value th
2, then carry out step 170;
(if differentiate correlation R
cGreater than the second critical value th
2, correlation R is differentiated in representative
cPairing differentiation index value τ
cBe positioned at maximum index value τ
MaxNear, then calculate to be next to and differentiate index value τ
cAfter the correlation of index value (as the τ among Fig. 2
jPairing correlation R (τ
j)), otherwise, can ignore and differentiate index value τ
cThe calculating of the pairing correlation of a plurality of index values in back, and direct computational discrimination index value τ
cThe back first predetermined number Δ
1Or the second predetermined number Δ
2The pairing correlation of index value, to save the time that dsp chip is used for calculating the required cost of correlation.Be noted that, in order to find out maximum related value R really
MaxThe τ at place
MaxFor the purpose of, the first critical value th
1And the second critical value th
2Initial set value can not be excessive, for instance, if the second critical value th at the beginning
2Be configured to the 3rd critical value th
3, then according to the judgement of step 108, method 100 is calculating R (τ
j) after, can not calculate R (τ
j+ 1), can calculate R (τ on the contrary
j+ Δ
2), calculate at last R (τ '
Max) (rather than correct R (τ
Max)), and R (τ '
Max) corresponding to index value τ '
Max(rather than correct τ
Max) also just be used to synthetic S mistakenly
3[n])
Step 110: with correlation R (k| τ
c<k<τ
c+ Δ
1, if k<N) all be set at zero, and will differentiate index value τ
cSet (τ for
c=τ
c+ Δ
1), calculate S
1[n] and S
5[n] is corresponding to differentiating index value (τ
c) differentiation correlation R (τ
c), carry out step 106;
Step 140: with correlation R (k| τ
c<k<τ
c+ Δ
2, if k<N) all be set at zero, and will differentiate index value τ
eSet (τ for
c=τ
c+ Δ
2), calculate S
1[n] and S
5[n] is corresponding to differentiating index value τ
cDifferentiation correlation R (τ
c), carry out step 106;
Step 170: will differentiate index value τ
cSet (τ for
c=τ
c+ 1), calculates S
1[n] and S
5[n] is corresponding to differentiating index value τ
cDifferentiation correlation R (τ
c), carry out step 106;
Step 200: find out the maximum related value R in the correlation table 30
MaxPairing maximum index value τ
Max
Step 202: with S
5[n] postpones maximum index value τ
Max, to produce S
4[n];
Step 204: with S
1[n] weighting is synthesized in S
4[n] is to produce S
3[n].
(wherein
S
3[n]=S
1[n] is as 0<=n<([N/3]+τ
Max) time;
S
3[n]=(N-n)/(N-([N/3]+τ
Max)) * S
1[n]+(n-([N/3]+τ
Max))/(N-([N/3]+τ
Max)) * S
4[n-([N/3]+τ
Max)], as ([N/3]+τ
MaxDuring)<=n<N;
S
3[n]=S
4[n-([N/3]+τ
Max)], when N<=n<=(N+[N/3]+τ
Max) time;
Step 300: according to maximum related value R
MaxUpgrade the first critical value th
1And the second critical value th
2
(because S
1[n] and S
2[n] separates from S[n], and S[n] be that sampling is from original signal S
Org(audio or video) therefore continues at S
1[n] and S
2[n] back S[n] in sampled signal, S for example
6[n] and S
7[n] is with S
1[n] and S
2Characteristic between [n] can not gone too far away mutually, so that the maximum related value R that is calculated in the step 200
MaxJust can be used as synthetic S
6[n] and S
7The first critical value th that [n] is required
1And the second critical value th
2The renewal foundation, so just can remove from because of avoiding calculating the τ ' of mistake
MaxAnd set the first too small critical value th especially
1And the second critical value th
2Necessity, the first too small critical value th
1And the second critical value th
2Will make this dsp chip calculate many unnecessary correlations)
Step 302: finish.
See also Fig. 4, Fig. 4 is the S in the preferred embodiments of the present invention
1[n] and S
2[n] synthesizes S
3The synoptic diagram of [n].First 400 among Fig. 4 is the S in the step 102 of display packing 100
1[n] and S
2[n], second portion 402 are τ that the step 103 of display packing 100 is calculated to the step 202
MaxAnd S
4S in the step 204 of [n] and third part 404 display packings 100
1[n] and S
4[n] synthesizes in S
3[n].
In an embodiment of the present invention, the correlation R (k| τ<k<τ+Δ in the step 110 of method 100,140
1.2, if k<N) all is set to zero, however these correlations also can be set to the complete any value that equates or do not equate entirely beyond zero, as long as these correlations all less than, preferably much smaller than maximum related value R
MaxGet final product.
Above-mentioned S
1[n] is if be congruent to S
2[n], i.e. S
1[n] and S
2[n] separates from S[n] same position, as shown in Figure 5, then method 100 is to increase S
1[n].On the contrary, S
1[n] and S
2[n] if unequal, i.e. S
1[n] and S
2[n] separates from S[n] diverse location, as shown in Figure 6, then method 100 is with S
1[n], S
8[n] (being rejected), and S
2[n] shortens to S
3[n].
With respect to existing TDHS, method of the present invention is according to the relaying correlation in the correlation table and the magnitude relationship of critical value, calculate corresponding to the pairing correlation of index value behind the relaying index value of this relaying correlation, owing to do not need to calculate one by one correlations all in this correlation table, therefore can save the time that the DSP that is used for setting up this correlation table calculates the required cost of this correlation, jointly, also increase the operation efficiency of the computing machine at this DSP place.In a preferred embodiment of the invention, the first predetermined number Δ
1And the second predetermined number Δ
2Be respectively 24 and 6, and the first critical value th
1And the second critical value th
2Be respectively R
Max/ 2 and R
Max/ 4 (promptly distinguish amputation (truncate) R
MaxTwo positions in one at end and end), it is original 10% that the calculated amount of DSP is kept to, and the unlikely S that influences
3The quality of [n].
The above only is the preferred embodiments of the present invention, and identical change and the improvement done according to the present patent application claim are all in the scope of claim of the present invention.
Claims (20)
1. the sequential conversion of the multistage stepping of adaptability is used for sequence signal S with audio frequency and video
1[n] and S
2[n] synthesizes S
3[n], this method comprises the following step:
(a) calculate S
1[n] and S
2[n] is corresponding to first correlation of first index value;
(b) relatively this first correlation and critical value;
(c) if this first correlation less than this critical value, then calculates S
1[n] and S
2[n] is corresponding to the pairing correlation of first number index value after this first index value; If this first correlation greater than this critical value, then calculates S
1[n] and S
2[n] corresponding to the pairing correlation of second number index value after this first index value, wherein, second number is less than first number; And
(d) according to the pairing maximum index value of maximum related value, the S that calculate
1[n] and S
2[n] produces S
3[n].
2. the method for claim 1, wherein S
1The signal number that [n] comprised is N
1, and S
2The signal number that [n] comprised is N
2, in the step (d), S
1[n] is that weighting is synthesized in S
4[n] is to produce S
3[n], S
4[n] is with S
2[n] postpones this maximum index value.
3. method as claimed in claim 2, wherein
S
3[n]=S
1[n] is when 0<=n<this maximum index value;
S
3[n]=(N
1-n)/(N
1-this maximum index value) * S
1[n]+(this maximum index value of n-)/(N
1-this maximum index value) * S
4[this maximum index value of n-] is as this maximum index value<=n<N
1The time;
S
3[n]=S
4[this maximum index value of n-] works as N
1<=n<=(N
2+ this maximum index value).
4. the method for claim 1, wherein step (c) also comprises: (e) correlation with the index value skipped is set at zero.
5. the method for claim 1, it also comprises:
(f) upgrade this critical value according to this maximum related value.
6. the method for claim 1, wherein S
1[n] and S
2[n] takes a sample respectively from audio-video signal S
1(t) and S
2(t).
7. method as claimed in claim 6, wherein audio-video signal S
1(t) and S
2(t) separate from original signal.
8. method as claimed in claim 7, wherein this original signal is a sound signal.
9. method as claimed in claim 7, wherein this original signal is a vision signal.
10. method as claimed in claim 7, wherein S
1(t) equal S
2(t).
11. method as claimed in claim 7, wherein S
1(t) be not equal to S
2(t).
12. the method for claim 1, wherein this second number equals 1.
13. the method for claim 1, wherein this first number is greater than 1.
14. the sequential conversion of the multistage stepping of adaptability is used for sequence signal S with audio frequency and video
1[n] and S
2[n] synthesizes S
3[n], this method comprises the following step:
(a) with S
2[n] postpones predetermined number to form S
5[n];
(b) calculate S
1[n] and S
5[n] is corresponding to first correlation of first index value;
(c) relatively this first correlation and a critical value;
(d) if this first correlation less than this critical value, then calculates S
1[n] and S
5[n] is corresponding to the pairing correlation of index value of first number after this first index value; If this first correlation greater than this critical value, then calculates S
1[n] and S
5[n] corresponding to the pairing correlation of index value of second number after this first index value, wherein, second number is less than first number; And
(e) according to the pairing maximum index value of maximum related value, the S that calculate
1[n] and S
5[n] produces S
3[n].
15. method as claimed in claim 14, wherein S
1The signal number that [n] comprised is N
1, and S
2The signal number that [n] comprised is N
2, in the step (e), S
1[n] weighting is synthesized in S
4[n] is to produce S
3[n], S
4[n] is with S
5[n] postpones (this predetermined number+this maximum index value).
16. method as claimed in claim 15, wherein
S
3[n]=S
1[n] is when 0<=n<(this predetermined number+this maximum index value);
S
3[n]=(N
1-n)/(N
1-(this predetermined number+this maximum index value)) * S
1[n]+(n-(this predetermined number+this maximum index value))/(N
1-(this predetermined number+this maximum index value)) * S
4[n-(this predetermined number+this maximum index value)] is as (this predetermined number+this maximum index value)<=n<N
1The time;
S
3[n]=S
4[n-(this predetermined number+this maximum index value)] works as N
1<=n<=(N
2+ this predetermined number+this maximum index value).
17. method as claimed in claim 14, wherein step (d) also comprises: (f) correlation with the index value skipped is set at zero.
18. method as claimed in claim 14, it also comprises:
(g) upgrade this critical value according to this maximum related value.
19. method as claimed in claim 14, wherein this second number equals 1.
20. method as claimed in claim 14, wherein this first number is greater than 1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB031524303A CN100421151C (en) | 2003-07-30 | 2003-07-30 | Adaptive multistage stepping sequence switch method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB031524303A CN100421151C (en) | 2003-07-30 | 2003-07-30 | Adaptive multistage stepping sequence switch method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1581291A CN1581291A (en) | 2005-02-16 |
CN100421151C true CN100421151C (en) | 2008-09-24 |
Family
ID=34579834
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB031524303A Expired - Lifetime CN100421151C (en) | 2003-07-30 | 2003-07-30 | Adaptive multistage stepping sequence switch method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100421151C (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11161273A (en) * | 1997-11-21 | 1999-06-18 | Kawai Musical Instr Mfg Co Ltd | Musical sound generating device and method |
JPH11288290A (en) * | 1998-03-31 | 1999-10-19 | Yamaha Corp | Sound source system using computer software and storage medium |
US6111183A (en) * | 1999-09-07 | 2000-08-29 | Lindemann; Eric | Audio signal synthesis system based on probabilistic estimation of time-varying spectra |
CN1383129A (en) * | 2001-03-27 | 2002-12-04 | 雅马哈株式会社 | Waveform generating method and appts. thereof |
-
2003
- 2003-07-30 CN CNB031524303A patent/CN100421151C/en not_active Expired - Lifetime
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11161273A (en) * | 1997-11-21 | 1999-06-18 | Kawai Musical Instr Mfg Co Ltd | Musical sound generating device and method |
JPH11288290A (en) * | 1998-03-31 | 1999-10-19 | Yamaha Corp | Sound source system using computer software and storage medium |
US6111183A (en) * | 1999-09-07 | 2000-08-29 | Lindemann; Eric | Audio signal synthesis system based on probabilistic estimation of time-varying spectra |
CN1383129A (en) * | 2001-03-27 | 2002-12-04 | 雅马哈株式会社 | Waveform generating method and appts. thereof |
Also Published As
Publication number | Publication date |
---|---|
CN1581291A (en) | 2005-02-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6073100A (en) | Method and apparatus for synthesizing signals using transform-domain match-output extension | |
CN101859583B (en) | Noise reducing apparatus and noise reducing method | |
TW518557B (en) | Continuously variable time scale modification of digital audio signals | |
CN104464726B (en) | A kind of determination method and device of similar audio | |
JPH06266390A (en) | Waveform editing type speech synthesizing device | |
US20100192753A1 (en) | Karaoke apparatus | |
JPS623439B2 (en) | ||
CN105931634B (en) | Audio screening technique and device | |
EP3440564B1 (en) | Audio fingerprinting based on audio energy characteristics | |
JP2010015152A (en) | Method for time scaling of sequence of input signal values | |
CN107170432A (en) | A kind of music generating method and device | |
CN113241082A (en) | Sound changing method, device, equipment and medium | |
CN1164084A (en) | Sound pitch converting apparatus | |
JP3576800B2 (en) | Voice analysis method and program recording medium | |
CN100421151C (en) | Adaptive multistage stepping sequence switch method | |
JPH0193795A (en) | Enunciation speed conversion for voice | |
Wong et al. | Fast SOLA-based time scale modification using modified envelope matching | |
Pfitzinger | DFW-based spectral smoothing for concatenative speech synthesis. | |
JPS59223499A (en) | Phoneme recognition equipment | |
JP4442239B2 (en) | Voice speed conversion device and voice speed conversion method | |
CN1244901C (en) | Non-linear overlapping method for sequence switch | |
US20190019525A1 (en) | Audio processing method and audio processing device | |
US7337109B2 (en) | Multiple step adaptive method for time scaling | |
Röbel | Neural network modeling of speech and music signals | |
JPH06222794A (en) | Voice speed conversion method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CX01 | Expiry of patent term |
Granted publication date: 20080924 |
|
CX01 | Expiry of patent term |