CN1378198A - Position matching method and device for voice synthesizing - Google Patents
Position matching method and device for voice synthesizing Download PDFInfo
- Publication number
- CN1378198A CN1378198A CN 01112404 CN01112404A CN1378198A CN 1378198 A CN1378198 A CN 1378198A CN 01112404 CN01112404 CN 01112404 CN 01112404 A CN01112404 A CN 01112404A CN 1378198 A CN1378198 A CN 1378198A
- Authority
- CN
- China
- Prior art keywords
- signal
- quantization step
- configuration
- shielding
- phonetic synthesis
- 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
Images
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The present invention discloses a metabit matching method and a device for voice synthesizing. Several signal-to-shielding ratio are quantized to produce quantizing orders in defined number. Each of the quantizing orders corresponds to one matching bit, one quantizing signal-to-sielding ratio and one sample number separately. Then, the matching bits are initialized, and the shielding-to-noise ratios of the quantizing orders are calculated according to the matching bits are equantizing signal-to-shielding ratio, one quantizing order with minimum shielding-to-noise radio is found out and partial bits are distributed to those of the quantizing order with minimum shielding-to-noise ratio. Finally, the bit matching table is finished based on the changed total bit number.
Description
The present invention is a kind of position matching method and device, refers to the position matching method and the device of a phonetic synthesis especially.
Develop subband (subband) scrambler that with the human auditory system and can be used to compress the sound input signal that significantly changes usually, such as numerous and diverse voice, multitask sounding and music.The ultimate principle of subband coding is to cut apart input sound spectrum (audio spectrum) for several frequency bands, encodes respectively again.
Bank of filters (filter bank) is commonly used to the separately voice signal of input.Cutting apart sound spectrum for after several frequency bands, each frequency band is by a low-pass filter, again with Nai Shi take a sample than (Nyquist), quantize, coding, multitask (multiplexed) and transmit.Auditory model (psychoacoustic model) calculates one group of data with control quantizer and coding.These data are different according to the reality work of actual coding device.If possible, can use the estimation that shields critical (masking threshold) to do the control of this quantizer.
Aspect receiver, subband is the original frequency position of separating multitask, decipher and be transmitted back to them.Then the subband signal always lumps together, and gives the approximate value of a primary speech signal.
Fig. 1 is traditional subband scrambler.Voice signal S (n) obtains B1 to BN subband signal by N bandpass filter 11 filtering, it is characterized in that each subband signal is represented the energy intensity of voice signal at this frequency band.Voice signal S (n) parallel connection in batches inputs to N bandpass filter 11, and variable n represents n signal frame (frame).
Subband signal B1 ~ BN must each carry out the action of frequency reducing (decimation) by a frequency demultiplier 12.
Each subband can have the different figure place of number to quantize, and these information are built in the allocation list 13 in being.The position allocation list produces according to auditory model 14.Auditory model 14 utilizes the sense of hearing different from and set up the impression degree of being of frequency change, with identical energy, in high frequency, intermediate frequency and the low frequency of the receivable audio frequency of people's ear, its degree of being of experiencing but has difference, possibly can't hear at high frequency and low frequency, and can listen very clearly at intermediate frequency.People promptly utilize this spontaneous phenomenon decision auditory model p (f), it is characterized in that p: energy, f: frequency.
Therefore according to auditory model, people then can determine the position number that each subband is fit to.Influence heavier frequency band, dispose more position, influence lighter frequency band and then dispose less position, do not have (if can be insufficient) even for the figure place of configuration.
Quantizing encoder 15 promptly can quantize each subband signal according to the voice signal S (n) of input for the figure place of configuration in the energy size of each subband and this subband of position allocation list record.For example subband signal B1 allocation list record on the throne can be two for the position number of configuration, then subband signal B1 is being 00 through the numerical value that may occur after frequency demultiplier 12 and the quantizing encoder 15,01,10,11, can be used for representing voice signal S (n) respectively when n signal frame, minimum value, sub-minimum, second largest value and the maximal value of subband signal B1.
All subband signal B1 ~ BN's then can be by the encoded voice data of finishing of multiplexer MUX output behind quantization encoding, and is stored, up to each signal time-frame all encode finish till.
See also Fig. 2.When people's desire is taken out above-mentioned encoded voice data, then need be deciphered to it, and the process of decoding the process with coding is just opposite basically.Encoded voice data X (n) can take out the data of each subband through separating multiplexer 21 earlier, imports code translator 22 more respectively and solves data.22 whiles of code translator could solve data with reference to last rheme allocation list.
The data b1 that solves ~ bN then can do the action of a raising frequency through frequency-booster 23, promptly gives sampling (upsample), imports each bandpass filter 24 and mixer 25 again, voice signal S (n) reduction is returned and broadcast.
The traditional bit collocation method is limiting aspect the sound quality of rebuilding to some extent.The principle of traditional bit configuration is that overall noise shielding ratio (noise-to-mask) is got at individual signal frame (being generally 10-30ms) and minimized, thereby the figure place that must not can use above each signal frame of institute's bit quantity of using.For example, when position ratio (bit rate) is K ms for Bbits/sec and signal frame length, can be used to do the figure place that subband and limit are put information (sideinformation) coding on the signal frame " adb " be calculated as follows:
adb＝B÷1000×K
Shielding factor (masking effects) exists many signal frames can't be by the sense of hearing of people's ear institute usually.In traditional coded program, the signal frame that these can't the sense of hearing but is configured the position of some equally.Therefore can make the effect of position configuration have a greatly reduced quality.
In consumer products, be demand invariably with the low cost.Above-mentioned signal frame that can't the sense of hearing but is configured the way of position, obviously increases cost on foot.
Purpose of the present invention promptly is to improve the shortcoming of traditional method, makes the configuration of each all be configured in useful signal frame, is then filtered for signal frame that can't the sense of hearing, refuses configuration bit, makes the position configuration more efficient, and then reduces cost.
For reaching above-mentioned purpose, the present invention proposes a kind of position matching method of phonetic synthesis, and it is with a total bit amount, is dispensed to an allocation list.At first, provide a plurality of signal shielding ratios.Then, quantize this a plurality of signal shielding ratios, produce the quantization step of certain number, these quantization step systems correspond to a configuration figure place, respectively and quantize a signal shielding number of samples when.These configuration figure places of initialization.Quantize the signal shielding ratio according to these configuration figure places and these, calculate one of each quantization step shielding noise ratio.In these shielding noise ratios, find out the quantization step of a minimal shielding noise ratio, and distribute the configuration figure place of the part of this total bit amount to the quantization step of this minimal shielding noise ratio.At last, in response to the change of this total bit amount, finish the action that distributes this allocation list.
Because the present invention is when configuration bit, be that the total bit amount is once disposed on the allocation list that puts in place, when configuration, specially all signal shielding ratios are divided into a certain number quantization step, at the weight of each quantization step, disposed suitable configuration figure place again, like this then can fill part the high signal frame of weight is given in the position configuration, and for signal frame that can't the sense of hearing, then can dispose, thereby can increase the efficient of position configuration, and then reduce cost.
It is characterized in that,, these signal noise ratios are to be determined by the voice signal of an input and an auditory model.The voice signal of input is after the integral body input is finished, and determines these signal noise ratios again.This total bit amount is the whole figure places of configuration to the voice signal of this input.
The position allocation list comprises a time coordinate and a frequency band coordinate, the figure place that on behalf of a time and a frequency band, the content of one lattice distributed.
Quantize these a plurality of signal shielding ratios and comprise the following step: these a plurality of signal shielding ratios of classifying, find out these quantization steps; These number of samples of decision in each quantization step are in order to the number of the signal shielding ratio that writes down each quantization step; And in each quantization step, find out these and quantize the signal shielding ratio, in order to represent signal shielding ratios all in these quantization steps.
These configuration figure places are in order to represent the number of each quantization step configuration bit.Setting these configuration figure places during these configuration figure places of initialization all is zero.
This shielding noise ratio of these quantization steps is that this configuration figure place according to these quantization steps is multiplied by this signal shielding ratio that a constant deducts these quantization steps again and obtains.This constant is 6.02, in order to represent pairing signal ratio of gains of the every increase of this configuration figure place.
The part of this total bit amount is the number of samples of the quantization step of this minimal shielding noise ratio.Give the configuration of configuration figure place of the quantization step of this minimal shielding noise ratio according to the number of samples of the quantization step of this minimal shielding noise ratio, the configuration figure place of the quantization step of this minimal shielding noise ratio is to add one.In addition, be dispensed to the configuration figure place of the quantization step of this minimal shielding noise ratio according to the part of this total bit amount, this total bit amount is decrescence to change.That is decrescence to zero, this allocation list is to have assigned in response to this total bit amount.
In addition, the present invention proposes a kind of position inking device of phonetic synthesis, is in order to the voice signal according to an input, produces an allocation list, carries out the action of a signal Processing for a speech processor.It comprises an auditory model, a digital storage unit, quantizer.Auditory model produces a plurality of signal shielding ratios in order to filter this whole voice signals.The digital storage unit is electrically connected to this auditory model, uses to store these signal shielding ratios.Quantizer is electrically connected to this digital storage unit, to use these signal shielding ratios of classification be certain number quantization step, and, distribute each quantization step one configuration figure place, to produce this allocation list according to the weight of each quantization step.
It is characterized in that,, this signal Processing action that this speech processor carries out, the action of encoding according to this allocation list, and the weight of each quantification amount depends on the size of some signal shielding ratios that each quantization step is contained.
The present invention will have more deep understanding by following accompanying drawing and detailed description:
Fig. 1: be traditional subband scrambler.
Fig. 2: be traditional subband code translator.
Fig. 3: be subband scrambler of the present invention.
Fig. 4: be position of the present invention configuration flow.
Fig. 5: be application examples of the present invention.
See also Fig. 3.The position inking device of phonetic synthesis of the present invention mainly comprises an auditory model 35, a digital storage unit 37, quantizer 38, in order to voice signal S (n) according to an input, produce an allocation list 39, carry out the action of a signal Processing for a speech processor 301.Auditory model 35 produces a plurality of signal shielding ratios 36 in order to filter whole voice signal S (n).Digital storage unit 37 is electrically connected to this auditory model 35, uses to store these signal shielding ratios 36.Quantizer 38 is electrically connected to this digital storage unit 37, uses these signal shielding ratios 36 of classification and be the quantization step of certain number, and according to the weight of each quantization step, distribute each quantization step one to dispose figure place, to produce this allocation list 39.
It is characterized in that,, this signal Processing action that this speech processor 301 carries out be the action of encoding according to this allocation list, and the weight of each quantization step is depended on the size of some signal shielding ratios that each quantization step is contained.
In addition, (n i) represents i subband sampling value in n the signal frame through the subband signal S of frequency reducing.All subband sampling values (subband samples) all are stored in the storage element 31.Simultaneously, each subband signal shielding is calculated by auditory model 35 than SMR (n) in n the signal frame, and is stored in the storage element 37.
In other words, voice signal S (n) can be introduced into position of the present invention inking device 302, the signal frame of once all finishing dealing with all, and, dispose suitable configuration figure place, to obtain an allocation list 39 according to the weight of each signal frame.
Thus, import and be stored in the voice signal (energy of each subband) of storage element 31, when subband code sample device 32 is encoded, can be according to the figure place of each time of record and subband planted agent configuration on the allocation list of position, give quantization encoding, and then deposit the data X (n) of quantization encoding and a position allocation list (claiming the limit to put information) in ROM (read-only memory) as last, for the usefulness of decoding, the process of deciphering is then identical with tradition.
The present invention is main to be to reach the efficient that improves configuration bit and reduce cost by different position matching methods.And the flow process of position of the present invention configuration is as shown in Figure 4, and it comprises the following step:
41: the definition following parameters:
QL: the number of signal shielding ratio quantization step.Voice signal can be stored earlier before the action that quantizer 38 quantizes at storage element 37 through all signal shielding ratios that auditory model filters, that is there has been the individual signal shielding ratio of N*T (plural number) in the storage element, it is characterized in that, N is the number (equaling the number of bandpass filter 11 among Fig. 3) of subband, T is a time point altogether, i.e. signal frame number.Many shielding ratios of this NT can be classified as a QL certain number quantization step, NT＞QL earlier in quantizer.
NQL (i): the number of samples of i quantization step.I.e. contained subband number in i signal shielding ratio quantization step.Each subband can comprise a signal shielding ratio, total NT signal shielding ratio, so after being categorized as QL quantization step, and the signal shielding ratio that has each quantization step to contain not wait number, promptly each quantization step has different number of samples.
SMR (i): quantize the signal shielding ratio, represent the signal shielding ratio of i quantization step.As mentioned above, contain the signal shielding ratio that has number to differ at each quantization step, but when quantizing, must in each quantization step, determine a numerical value that is used for representing all signal shielding ratios in each quantization step, so people are with this numerical value signal shielding ratio that quantizes by name.Quantize the intermediate value representative of the signal shielding ratio in then desirable each quantization step of signal shielding ratio.
BQL (i): the configuration figure place of i quantization step.Because people have been categorized as QL quantization step, and each quantization step can correspond to one and quantize a signal shielding number of samples when, get final product so when configuration bit, only need be configured, when a quantization step disposes one, promptly represent all samplings of this quantization step (subband) all to mix one each quantization step.
TB: the total bit amount promptly is used for to input speech signal all available figure places of encoding.In the process of configuration bit, then can reduce gradually till distribution is over (TB=0).
42: the configuration figure place of each quantization step of initialization.Because the configuration figure place of each quantization step is started from scratch usually, so being configuration figure place BQL (i) with each quantization step at the beginning, people are made as 0, grammer then is BQL (i)=0, for i=1...QL.
43: the shielding noise ratio MNR (i) (Mask to Noise Ratio) that calculates the i quantization step.Shielding noise ratio MNR (i) is signal shielding ratio SMR (i) generation according to the configuration bit BQL (i) of i quantization step and i quantization step.Its calculating formula is as follows:
MNR(i)＝BQL(i)×6.02-SMR(i)
It is characterized in that constant 6.02 is in order to represent pairing signal ratio of gains of the every increase of this configuration figure place, this is the general rule of analog digital conversion.
44: find out minimum shielding noise ratio MNR (k).It is characterized in that k represents k quantization step, that is at the shielding noise ratio minimum of k quantization step (at signal shielding ratio SMR (k) the maximum then).The minimum weight of represent it to influence the sense of hearing of shielding noise ratio MNR is the highest, so people must dispose earlier the high person of weight, every configuration once, its weight then can reduce by one (being equivalent to 6.02dB), turns one's head to compare with the shielding noise ratio of other quantization step again.
45: upgrade the total bit amount.Owing to disposed some positions to the subband of each sampling of the highest quantization steps of above-mentioned weight, so available relatively total bit amount can reduce relatively.
46: if the total bit amount can then be got back to step 43 and repeat above-mentioned action for (TB＞0) of configuration in addition.Otherwise then finish the action of configuration bit.
47: finish the action of configuration bit.After configuration bit finishes, in each lattice in the allocation list on the throne, there is the figure place that each subband distributed of representing each time (signal frame).In Fig. 3, subband sampling scrambler 32 can be according to the position allocation list 39 that this flow process produced, in addition to the speech signal coding of input.
Fig. 5 is common speech synthetic device, comprises ROM (read-only memory) (ROM), random access memory (RAM), digital signal processor (DSP), digital analog converter (D/A) and loudspeaker (Speaker) etc.Last rheme allocation list and encoded voice signal are stored in the read only memory ROM, and the function of digital signal processor DSP removes to handle the voice signal of rebuilding (reconstructed) pulse wave coding modulation (PCM) for these encoded data being deciphered and put in order the synthetic action of group filtering.These voice PCM data is temporarily stored in the random memory RAM and plays usefulness for loudspeaker.Certainly before playing, can with digital analog converter D/A digital signal be converted to simulating signal earlier.The conversion ratio of digital analog converter D/A is then controlled by digital signal processor DSP.
By above-mentioned diagram and explanation, as can be known, principal feature of the present invention is on the position matching method, by the idea of vector quantization (vector quantization), produce the position allocation list of whole voice signal, the action of encoding according to this again, therefore can get rid of and to dispose unnecessary shortcoming on the subband of the sense of hearing, improved the efficient of position configuration, and then can save memory space, reduced cost.
Yet in spirit of the present invention and scope, can make many variations, these variations all should be included in appending claims.
Claims (21)
1. the position matching method of a phonetic synthesis with a total bit amount, is dispensed to an allocation list, comprises the following step:
A plurality of signal shielding ratios are provided;
Quantize this a plurality of signal shielding ratios, produce the quantization step of certain number, these quantization steps correspond to a configuration figure place, respectively and quantize a signal shielding number of samples when;
These configuration figure places of initialization;
Quantize the signal shielding ratio according to these configuration figure places and these, calculate a shielding noise ratio of each quantization step;
In these shielding noise ratios, find out the quantization step of a minimal shielding noise ratio, and distribute the configuration figure place of the part of this total bit amount to the quantization step of this minimal shielding noise ratio; And
According to the change of this total bit amount, finish the action that distributes this allocation list.
2. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, these signal noise ratios are by the voice signal and the auditory model decision of an input.
3. the position matching method of phonetic synthesis as claimed in claim 2 is characterized in that, the voice signal of this input determines these signal noise ratios again after the integral body input is finished.
4. the position matching method of phonetic synthesis as claimed in claim 3 is characterized in that, this total bit amount is the whole figure places of configuration to the voice signal of this input.
5. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, this allocation list comprises a time coordinate and a frequency band coordinate.
6. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, the figure place that on behalf of a time and a frequency band, the content of lattice of this allocation list distributed.
7. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, these a plurality of signal shielding ratios of this quantification comprise the following step:
This a plurality of signal shielding ratios of classifying are found out these quantization steps;
These number of samples of decision in each quantization step are in order to the number of the signal shielding ratio that writes down each quantization step; And
In each quantization step, find out these and quantize the signal shielding ratio, in order to represent signal shielding ratios all in these quantization steps.
8. the position matching method of phonetic synthesis as claimed in claim 7 is characterized in that, these configuration figure places are in order to represent the number of each quantization step configuration bit.
9. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, these configuration figure places of initialization are that these configuration figure places of setting all are zero.
10. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, this shielding noise ratio of these quantization steps is that this configuration figure place according to these quantization steps is multiplied by this signal shielding ratio that a constant deducts these quantization steps again and obtains.
11. the position matching method of phonetic synthesis as claimed in claim 10 is characterized in that, this constant is 6.02, in order to represent pairing signal ratio of gains of the every increase of this configuration figure place.
12. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, the part of this total bit amount is the number of samples of the quantization step of this minimal shielding noise ratio.
13. the position matching method of phonetic synthesis as claimed in claim 12, it is characterized in that, according to the configuration of the configuration figure place of the quantization step of several these minimal shielding noise ratios of sampling of the quantization step of this minimal shielding noise ratio, the configuration figure place of the quantization step of this minimal shielding noise ratio is to add one.
14. the position matching method of phonetic synthesis as claimed in claim 1 is characterized in that, is dispensed to the configuration figure place of the quantization step of this minimal shielding noise ratio according to the part of this total bit amount, this total bit amount decrescence changes.
15. the position matching method of phonetic synthesis as claimed in claim 14 is characterized in that, decrescence to zero, this allocation list is to have assigned according to this total bit amount.
16. the position inking device of a phonetic synthesis, the voice signal in order to according to an input produces an allocation list, carries out the action of a signal Processing for a speech processor, comprises:
One auditory model in order to filter this whole voice signals, produces a plurality of signal shielding ratios;
One digital storage unit is electrically connected to this auditory model, uses to store these signal shielding ratios; And
One quantizer is electrically connected to this digital storage unit, to use these signal shielding ratios of classification be certain number quantization step, and, distribute each quantization step one configuration figure place, to produce this allocation list according to the weight of each quantization step.
17. the position inking device of phonetic synthesis as claimed in claim 16 is characterized in that, this signal Processing action that this speech processor carries out, the action of encoding according to this allocation list.
18. the location means of phonetic synthesis as claimed in claim 16 is characterized in that, this allocation list comprises a time coordinate and a frequency band coordinate.
19. the position inking device of phonetic synthesis as claimed in claim 16 is characterized in that the content of lattice of this allocation list is a figure place of representing a time and a frequency band to be distributed.
20. the position inking device of phonetic synthesis as claimed in claim 16 is characterized in that these quantization step tool one number of samples are in order to the number of the signal shielding ratio that writes down each quantization step.
21. the position inking device of phonetic synthesis as claimed in claim 16 is characterized in that the size of some signal shielding ratios that each quantization step is contained is depended in the weight system of each quantization step.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB011124040A CN1173330C (en) | 2001-03-29 | 2001-03-29 | Position matching method and device for voice synthesizing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB011124040A CN1173330C (en) | 2001-03-29 | 2001-03-29 | Position matching method and device for voice synthesizing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1378198A true CN1378198A (en) | 2002-11-06 |
| CN1173330C CN1173330C (en) | 2004-10-27 |
Family
ID=4659279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB011124040A Expired - Fee Related CN1173330C (en) | 2001-03-29 | 2001-03-29 | Position matching method and device for voice synthesizing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1173330C (en) |
-
2001
- 2001-03-29 CN CNB011124040A patent/CN1173330C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN1173330C (en) | 2004-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1110145C (en) | Scalable audio coding/decoding method and apparatus | |
| CN1154085C (en) | Scalable audio coding/decoding method and apparatus | |
| KR100335609B1 (en) | Scalable audio encoding/decoding method and apparatus | |
| CN1046608C (en) | Apparatus and method for data compression using signal-weighted quantizing bit allocation | |
| EP1749296B1 (en) | Multichannel audio extension | |
| CN1527306B (en) | Method and apparatus for coding and/or decoding digital data using bandwidth expansion technology | |
| CN1702974B (en) | Method and apparatus for encoding/decoding a digital signal | |
| CN1822508B (en) | Method and apparatus for encoding and decoding digital signals | |
| CN101055720B (en) | Method and apparatus for encoding and decoding an audio signal | |
| CN1248824A (en) | Audio signal coding device and method, decoding device and method | |
| CN1153365C (en) | Transfer system adopting different coding principle | |
| RU97122037A (en) | METHOD AND DEVICE FOR SCALABLE ENCODING / DECODING OF THE AUDIO SIGNAL | |
| CN101105940A (en) | Audio frequency encoding and decoding quantification method, reverse conversion method and audio frequency encoding and decoding device | |
| CN1138254C (en) | Audio signal comprssing coding/decoding method based on wavelet conversion | |
| CN1154084C (en) | Audio coding/decoding technology based on pseudo wavelet filtering | |
| CN1173330C (en) | Position matching method and device for voice synthesizing | |
| CN1273955C (en) | Method and device for coding and/or decoding audip frequency data using bandwidth expanding technology | |
| CN108417219A (en) | A kind of audio object decoding method being adapted to Streaming Media | |
| CN1318904A (en) | Practical sound coder based on wavelet conversion | |
| CN1290078C (en) | Method and device for coding and/or devoding audio frequency data using bandwidth expanding technology | |
| CN1065400C (en) | Compatible AC-3 and MPEG-2 audio-frequency code-decode device and its computing method | |
| CN2929904Y (en) | Signal coding and de-coding device | |
| CA2194199C (en) | Encoding system and decoding system for audio signals | |
| CN1460247A (en) | ADPCM speech coding system with specific step-size adaptation | |
| JP3297238B2 (en) | Adaptive coding system and bit allocation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C06 | Publication | ||
| PB01 | Publication | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20041027 |