WO2007006623A1 - Method and apparatus for detecting correlation between a received signal and a candidate spreading sequence - Google Patents
Method and apparatus for detecting correlation between a received signal and a candidate spreading sequence Download PDFInfo
- Publication number
- WO2007006623A1 WO2007006623A1 PCT/EP2006/063308 EP2006063308W WO2007006623A1 WO 2007006623 A1 WO2007006623 A1 WO 2007006623A1 EP 2006063308 W EP2006063308 W EP 2006063308W WO 2007006623 A1 WO2007006623 A1 WO 2007006623A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- time period
- maximum value
- correlation
- max
- spreading sequence
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
- H04B1/69—Spread spectrum techniques
- H04B1/707—Spread spectrum techniques using direct sequence modulation
- H04B1/7073—Synchronisation aspects
- H04B1/7075—Synchronisation aspects with code phase acquisition
- H04B1/70754—Setting of search window, i.e. range of code offsets to be searched
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
- H04B1/69—Spread spectrum techniques
- H04B1/707—Spread spectrum techniques using direct sequence modulation
- H04B1/7073—Synchronisation aspects
- H04B1/7075—Synchronisation aspects with code phase acquisition
- H04B1/70751—Synchronisation aspects with code phase acquisition using partial detection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
- H04B1/69—Spread spectrum techniques
- H04B1/707—Spread spectrum techniques using direct sequence modulation
- H04B1/7073—Synchronisation aspects
- H04B1/7075—Synchronisation aspects with code phase acquisition
- H04B1/70755—Setting of lock conditions, e.g. threshold
Definitions
- the invention relates to a method and to an apparatus for detecting correlation between a received signal and a candidate spreading sequence that may have been used to watermark said signal.
- Audio watermarking technologies like e.g. psycho-acoustical shaped spread spectrum perform correlation to retrieve the watermark (WM) data.
- the correla- tion can be followed by a spectral whitening, as explained e.g. in EP-A-1594122.
- a watermarked audio signal is recorded in an acoustical environment it will contain echoes and reverberation so that the detection of the correct correlation becomes difficult. In many applications this prob- lem is ignored, meaning that simply the maximum of the correlation is used (or the maximum at time-lag zero, if the correlation is already synchronised) for its evaluation.
- One solution in the presence of distinct echoes is to use a Rake receiver, which contains several correlators and calcu- lates one correlation for each echo, the location of which echo must be known and must be basically stationary.
- the absolute values of the cor- related signals or correlations are formed, these values are low-pass filtered and the correlation is evaluated in two different time ranges: one where the direct sound is expected and the other one where the first echo is expected.
- the maximum in both ranges is calculated and based on these maxima different criteria are used to choose the correct correlation result from several candidate correlations .
- the inventive method is suited for detecting correlation between a received signal and a candidate spreading sequence that may have been used to watermark said signal, said method including the steps:
- the inventive apparatus is suited for detecting correlation between a received signal and a candidate spreading sequence that may have been used to watermark said signal, said apparatus including:
- - means being adapted for correlating a current block of received signal values with one or more candidate spreading sequences; means being adapted for calculating the absolute values of the correlation signals and lowpass filtering these absolute values; - means being adapted for evaluating said lowpass filtered absolute correlation signal values in a first time period in which the direct sound is expected and in a second, different time period in which the first echo is expected, thereby determining the maximum value in said first time period and the maximum value in said second time period, and for choosing the correct correlation signal result from several candidate correlation signals based on these maxima, by comparison with threshold values.
- Fig. 1 correlation signals without disturbance between WM embedding and WM extracting
- Fig. 2 correlation signals in the presence of an acoustic path between embedding and extracting and low noise level
- Fig. 3 correlation signals in the presence of the acoustic path between embedding and extracting and higher noise level
- FIG. 4 the correlation signals of Fig. 3 after the inven- tive processing has been applied;
- Fig. 5 flowchart of the inventive method;
- Fig. 6 block diagram of the inventive apparatus.
- Fig. 1 shows three correlation signals plotted together, without disturbance by any additional acoustic signal path (or after passing a channel without noise) between WM embedding and WM extracting, whereby the peak representing the bit value of the watermark data is clearly present or absent .
- Binary Phase Shift Keying i.e. the encoding of the WM data in the sign of the spreading sequence, is not very robust against an acoustic path because reflections, echoes and reverberation will shift the phase of the signal.
- Fig. 2 shows three correlation signals plotted together.
- Code Shift Keying uses different spreading sequences for encoding the WM data. In this case not the sign or polarity of the peak is to be detected, but which one of the candidate correlation signals shows a characteristic maxi- mum.
- Fig. 3 shows three correlation signals plotted together whereby different spreading sequences were used for water- marking an audio signal which has been transmitted and recorded in an office room as an example for a normal acoustic environment. In this case it is impossible to decide, following correlation of the received signal with corresponding candidate spreading sequences, which one of the spreading sequences had been embedded in the current WM data frame of the original audio signal.
- the invention takes advantage of the following facts : - A correlation at reception side with the clean, original signal gives a clear, only a few samples wide, positive or negative peak as depicted in Fig. 1. If there is an additional acoustic path between embedder and receiver, the peak is wider and has negative as well as positive parts as depicted in Fig. 2.
- the peak in the correlation from direct sound is by definition approximately at time-lag zero, and the peak from first early reflections shows up not later than approximately 3ms .
- the inventive processing includes the following steps :
- a first time range in these low-pass filtered correlation values e.g. a range from -10 samples to +10 samples (if the audio signal is already synchronised)
- selecting a second time range in these low-pass filtered correlation values e.g. a range from -10 samples to -150 samples (if the audio signal is already synchronised) ;
- Cio"* is the number of the smoothed correlation the spread- ing sequence of which leads to the largest value in the first time range
- Max ⁇ o is the corresponding maximum value in this time range
- Max ⁇ o is the maximum of the correlation with the second largest maximum in the first time range
- 0x50"*- is the number of the smoothed correlation the spreading sequence of which leads to the largest value in the second time range
- Max;L5o i the maximum of the correlation with the second largest maximum in the second time range.
- the first time lag or period is e.g. -10 to +10 samples and the second time lag is e.g. -10 to -150 samples) .
- Max ⁇ g is greater than Max ⁇ g multiplied with a constant second factor smaller than the first constant fac- tor, e.g. 1.2
- the corresponding sequence number C ⁇ g ⁇ is equal to sequence number C ⁇ o ⁇ and if Max 150" ⁇ ⁇ s greater than Max;L5o multiplied with a constant third factor smaller than the second constant factor, e.g. 1.1
- Max ⁇ o ⁇ is not greater than Max ⁇ g ⁇ multiplied with the second factor
- Max ⁇ 5Q is greater than Max ⁇ 5Q multiplied with a constant fourth factor of e.g. 1.3. If true, it is assumed that the corresponding candidate spreading sequence (C ⁇ 5o-*-) is the correct spreading sequence and Max ⁇ o ⁇ is considered as indicating the searched sequence number C ⁇ 5o-*-. If not true, no spreading sequence can be detected.
- a Hanning window which is also called 'raised cosine window' , is a commonly used filter window in digital signal processing.
- the invention improves correlation peak de- tection and evaluation also in cases where the actual signal and its echoes are ⁇ smeared', i.e. where the correlation does not show single peaks but groups of peaks of different height and/or polarity, as shown in Fig. 2. Therefore the invention provides improved flexibility, e.g. for non- stationary echoes .
- the invention improves tremendously the detection rate of a correlation-based processing in the presence of an acoustic path, as is apparent from the comparison between Fig. 3 and Fig. 4.
- Concerning Fig. 4 in an exemplary test measurement the number of correctly identified correlations increased by 16%, the number of false detections decreased by 20%, and the number of non-identifiable results decreased by over 90%, as compared to the known processing.
- the inventive apparatus receives as input signal IPS a current block of input samples which preferably are whitened in a whitening stage WHT. Thereafter correlations with several candidate spreading sequences taken from a memory SPRS are carried out in correlation means CORR. The cor- relation signal absolute values are formed and lowpass filtered in stage ABSLPF. Thereafter the calculations according to Fig. 5 are carried out in a calculator CALC, which outputs information items indicating the number of the candidate spreading sequence found or indicating that no matching spreading sequence has been determined for the current block of input samples .
Abstract
Audio watermarking technologies use correlation to retrieve the watermark data. The inventive method for detecting correlation between audio signals includes the steps of calculating the absolute values of the correlated signals, lowpass filtering these absolute values, evaluating the correlation in two different time ranges, namely one where the direct sound is expected and the other where the first echo is expected, calculating the maxima in both ranges, and choosing the correct correlation from several candidate correlations based on these maxima.
Description
Method and apparatus for detecting correlation between a received signal and a candidate spreading sequence
The invention relates to a method and to an apparatus for detecting correlation between a received signal and a candidate spreading sequence that may have been used to watermark said signal.
Background
Audio watermarking technologies like e.g. psycho-acoustical shaped spread spectrum perform correlation to retrieve the watermark (WM) data. To improve the decoding, the correla- tion can be followed by a spectral whitening, as explained e.g. in EP-A-1594122. If a watermarked audio signal is recorded in an acoustical environment it will contain echoes and reverberation so that the detection of the correct correlation becomes difficult. In many applications this prob- lem is ignored, meaning that simply the maximum of the correlation is used (or the maximum at time-lag zero, if the correlation is already synchronised) for its evaluation. One solution in the presence of distinct echoes is to use a Rake receiver, which contains several correlators and calcu- lates one correlation for each echo, the location of which echo must be known and must be basically stationary.
Invention
However, this needs more processing power than the inventive solution described below. Furthermore it is to be expected that Rake receivers are working well in the presence of known discrete echoes, but do not work very well with smeared or time-varying echoes .
A problem to be solved by the invention is to improve the correlation detection in particular for watermarking technologies, so that it is possible to make the correct correlation result decision even under the above described worse conditions. This problem is solved by the method disclosed in claim 1. An apparatus that utilises this method is disclosed in claim 2.
According to the invention, the absolute values of the cor- related signals or correlations are formed, these values are low-pass filtered and the correlation is evaluated in two different time ranges: one where the direct sound is expected and the other one where the first echo is expected. The maximum in both ranges is calculated and based on these maxima different criteria are used to choose the correct correlation result from several candidate correlations .
In principle, the inventive method is suited for detecting correlation between a received signal and a candidate spreading sequence that may have been used to watermark said signal, said method including the steps:
- correlating a current block of received signal values with one or more candidate spreading sequences; calculating the absolute values of the correlation sig- nals and lowpass filtering these absolute values; evaluating said lowpass filtered absolute correlation signal values in a first time period in which the direct sound is expected and in a second, different time period in which the first echo is expected, thereby determining the maximum value in said first time period and the maximum value in said second time period;
- choosing the correct correlation signal result from several candidate correlation signals based on these maxima, by comparing with threshold values .
In principle the inventive apparatus is suited for detecting
correlation between a received signal and a candidate spreading sequence that may have been used to watermark said signal, said apparatus including:
- means being adapted for correlating a current block of received signal values with one or more candidate spreading sequences; means being adapted for calculating the absolute values of the correlation signals and lowpass filtering these absolute values; - means being adapted for evaluating said lowpass filtered absolute correlation signal values in a first time period in which the direct sound is expected and in a second, different time period in which the first echo is expected, thereby determining the maximum value in said first time period and the maximum value in said second time period, and for choosing the correct correlation signal result from several candidate correlation signals based on these maxima, by comparison with threshold values.
Advantageous additional embodiments of the invention are disclosed in the respective dependent claims.
Drawings
Exemplary embodiments of the invention are described with reference to the accompanying drawings, which show in: Fig. 1 correlation signals without disturbance between WM embedding and WM extracting; Fig. 2 correlation signals in the presence of an acoustic path between embedding and extracting and low noise level; Fig. 3 correlation signals in the presence of the acoustic path between embedding and extracting and higher noise level;
Fig. 4 the correlation signals of Fig. 3 after the inven-
tive processing has been applied; Fig. 5 flowchart of the inventive method; Fig. 6 block diagram of the inventive apparatus.
Exemplary embodiments
Fig. 1 shows three correlation signals plotted together, without disturbance by any additional acoustic signal path (or after passing a channel without noise) between WM embedding and WM extracting, whereby the peak representing the bit value of the watermark data is clearly present or absent .
Binary Phase Shift Keying (BPSK), i.e. the encoding of the WM data in the sign of the spreading sequence, is not very robust against an acoustic path because reflections, echoes and reverberation will shift the phase of the signal.
Fig. 2 shows three correlation signals plotted together.
Disturbance by one or more additional signal paths (or after passing a corresponding noisy channel) between WM embedding and WM extracting leads to a peak representing the watermark data bit value no more being clearly present or absent.
Code Shift Keying (CSK) uses different spreading sequences for encoding the WM data. In this case not the sign or polarity of the peak is to be detected, but which one of the candidate correlation signals shows a characteristic maxi- mum.
Even this decision is impossible to take if too much distortion is introduced between WM data insertion and WM data extraction, for example by echoes, reverberation or noise.
Fig. 3 shows three correlation signals plotted together whereby different spreading sequences were used for water-
marking an audio signal which has been transmitted and recorded in an office room as an example for a normal acoustic environment. In this case it is impossible to decide, following correlation of the received signal with corresponding candidate spreading sequences, which one of the spreading sequences had been embedded in the current WM data frame of the original audio signal.
The invention takes advantage of the following facts : - A correlation at reception side with the clean, original signal gives a clear, only a few samples wide, positive or negative peak as depicted in Fig. 1. If there is an additional acoustic path between embedder and receiver, the peak is wider and has negative as well as positive parts as depicted in Fig. 2.
- False peaks, which do not represent the embedded sequence correlation result, are often only one sample wide.
- Correct peaks, which are due to the embedded sequence, are derived either from the correlation result for the direct sound path, or from correlation with first early reflections . All other early and all late reflections are in normal acoustic conditions so much damped that their correlation peak hides in the noise floor.
- If the incoming signal is already synchronised, the peak in the correlation from direct sound is by definition approximately at time-lag zero, and the peak from first early reflections shows up not later than approximately 3ms .
Based on these facts, the inventive processing includes the following steps :
- calculating the correlation signal for the whitened sound signal for all candidate spreading bit sequences;
- taking the absolute values of each correlation signal; - low-pass filtering these absolute values, for example by averaging six successive values weighted by a Hanning win-
dow;
- selecting a first time range in these low-pass filtered correlation values, e.g. a range from -10 samples to +10 samples (if the audio signal is already synchronised) , and selecting a second time range in these low-pass filtered correlation values, e.g. a range from -10 samples to -150 samples (if the audio signal is already synchronised) ;
- carrying out the processing depicted in Fig. 5, in which: Cio"*" is the number of the smoothed correlation the spread- ing sequence of which leads to the largest value in the first time range,
Max^o is the corresponding maximum value in this time range,
Max^o is the maximum of the correlation with the second largest maximum in the first time range,
0x50"*- is the number of the smoothed correlation the spreading sequence of which leads to the largest value in the second time range,
Max;L5o is the corresponding maximum value in this time range,
Max;L5o is the maximum of the correlation with the second largest maximum in the second time range.
As shown in Fig. 5, for a current block of audio signal sam- pies the values
are ca-l~ culated. However, since for the first decision step the values of Maxχ5o-*- and Max^g^ are not used, they can also be calculated after this step. If Max^o^ is greater than Max^g^ multiplied with a constant first factor of e.g. 1.3, then Maxio"*" is considered as indicating the searched sequence number C^O"*"-
Otherwise, it is compared also in the second, wider area at a later time lag (the first time lag or period is e.g. -10 to +10 samples and the second time lag is e.g. -10 to -150 samples) . If Max^g is greater than Max^g multiplied with a constant second factor smaller than the first constant fac-
tor, e.g. 1.2, and if the corresponding sequence number C^g^ is equal to sequence number C^o^ and if Max150"^ ^s greater than Max;L5o multiplied with a constant third factor smaller than the second constant factor, e.g. 1.1, then it is as- sumed that this candidate spreading sequence Cχø"'"==Ci50"'" is the correct spreading sequence and that MaXi0 1 indicates the searched sequence number C^o"*"-
If the latter two conditions are not true, or if Max^o^ is not greater than Max^g^ multiplied with the second factor, it is checked whether Max^5Q is greater than Max^5Q multiplied with a constant fourth factor of e.g. 1.3. If true, it is assumed that the corresponding candidate spreading sequence (Cχ5o-*-) is the correct spreading sequence and Max^o^ is considered as indicating the searched sequence number Cχ5o-*-. If not true, no spreading sequence can be detected.
A Hanning window, which is also called 'raised cosine window' , is a commonly used filter window in digital signal processing. A Hanning window w[n] of length M is calculated as : w [n] = 0 . 5 - 0 . 5 cos (2π n/M) for 0 < n < M , w[n] = 0 for all other n , wherein Λn' is the sample number.
Advantageously, the invention improves correlation peak de- tection and evaluation also in cases where the actual signal and its echoes are Λ smeared', i.e. where the correlation does not show single peaks but groups of peaks of different height and/or polarity, as shown in Fig. 2. Therefore the invention provides improved flexibility, e.g. for non- stationary echoes .
The invention improves tremendously the detection rate of a correlation-based processing in the presence of an acoustic path, as is apparent from the comparison between Fig. 3 and Fig. 4. Concerning Fig. 4, in an exemplary test measurement the number of correctly identified correlations increased by
16%, the number of false detections decreased by 20%, and the number of non-identifiable results decreased by over 90%, as compared to the known processing.
In Fig. 6, the inventive apparatus receives as input signal IPS a current block of input samples which preferably are whitened in a whitening stage WHT. Thereafter correlations with several candidate spreading sequences taken from a memory SPRS are carried out in correlation means CORR. The cor- relation signal absolute values are formed and lowpass filtered in stage ABSLPF. Thereafter the calculations according to Fig. 5 are carried out in a calculator CALC, which outputs information items indicating the number of the candidate spreading sequence found or indicating that no matching spreading sequence has been determined for the current block of input samples .
Claims
1. Method for detecting correlation between a received signal (IPS) and a candidate spreading sequence (SPRS) that may have been used to watermark said signal, said method including the steps :
- correlating (CORR) a current block of received signal values with one or more candidate spreading sequences
(SPRS) ; - calculating (ABSLPF) the absolute values of the correlation signals and lowpass filtering these absolute values; evaluating (CALC) said lowpass filtered absolute correlation signal values in a first time period in which the direct sound is expected and in a second, different time period in which the first echo is expected, thereby determining the maximum value (Max^g ) i-n said first time period and the maximum value (Max^5Q ) in said second time period;
- choosing the correct correlation signal result from sev- eral candidate correlation signals based on these maxima, by comparing with threshold values .
2. Apparatus for detecting correlation between a received signal (IPS) and a candidate spreading sequence (SPRS) that may have been used to watermark said signal, said apparatus including:
- means (CORR) being adapted for correlating a current block of received signal values with one or more candidate spreading sequences (SPRS) ; - means (ABSLPF) being adapted for calculating the absolute values of the correlation signals and lowpass filtering these absolute values; means (CALC) being adapted for evaluating said lowpass filtered absolute correlation signal values in a first time period in which the direct sound is expected and in a second, different time period in which the first echo is expected, thereby determining the maximum value (Max^o ) in said first time period and the maximum value (Maxχ5o-*-) in said second time period, and for choosing the correct correlation signal result from several candi- date correlation signals based on these maxima, by comparison with threshold values.
3. Method according to claim 1, or apparatus according to claim 2, wherein said lowpass filtering is carried out by averaging six successive values weighted by a Hanning window.
4. Method according to claim 1 or 3, or apparatus according to claim 2 or 3, wherein said first time period is in the range -10 to +10 samples and said second time period is in the range -10 to -150 samples.
5. Method according to one of claims 1, 3 and 4, or apparatus according to one of claims 2 to 4, wherein also a second largest maximum value (Max^g ) i-n said first time period and a second largest maximum value (Max^5Q ) in said second time period are determined from said lowpass filtered absolute correlation signal values, and wherein: - if said maximum value (Max^g ) i-n said first time period is greater than said second largest maximum value
(Max^o ) i-n said first time period multiplied with a constant first factor, then said maximum value is considered as indicating that the corresponding candidate spreading sequence (C^O"*") is the correct spreading sequence; - if not true, it is determined whether said maximum value (Max^g ) i-n said first time period is greater than said second largest maximum value (Max^o ) i-n said first time period multiplied with a constant second factor smaller than said first constant fac- tor, and if said maximum value (Max^5Q ) in said second time period is greater than said second largest maximum value (Maxχ5Q^) in said second time period multiplied with a constant third factor smaller than said second constant factor, and if the corresponding candidate spreading sequence
(Cio"*") in said first time period is identical to the corresponding candidate spreading sequence (C^g-*-) in said second time period, and in case these three conditions are met it is assumed that this candidate spreading sequence (C^g^ (Ciso^) is the correct spreading sequence; if the latter two conditions of said three conditions are not true, or if the first one of said three conditions is not true, it is determined whether or not said maximum value (Maxχ5o-*-) in said second time period is greater than said second largest maximum value (Max^5Q ) in said second time period multiplied with a constant fourth factor, and if true it is assumed that the corresponding candidate spreading sequence (C^g-*-) is the correct spreading sequence.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP05014955.8 | 2005-07-11 | ||
| EP05014955 | 2005-07-11 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2007006623A1 true WO2007006623A1 (en) | 2007-01-18 |
Family
ID=36917384
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/EP2006/063308 WO2007006623A1 (en) | 2005-07-11 | 2006-06-19 | Method and apparatus for detecting correlation between a received signal and a candidate spreading sequence |
Country Status (1)
| Country | Link |
|---|---|
| WO (1) | WO2007006623A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2081187A1 (en) * | 2008-01-21 | 2009-07-22 | Deutsche Thomson OHG | Method and apparatus for determining whether or not a reference pattern is present in a received and possibly water-marked signal |
| CN109784126A (en) * | 2017-11-10 | 2019-05-21 | 富士通株式会社 | Data cutting method and device, article detection method and device |
| US11183198B2 (en) | 2012-10-15 | 2021-11-23 | Digimarc Corporation | Multi-mode audio recognition and auxiliary data encoding and decoding |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0756387A2 (en) * | 1995-07-25 | 1997-01-29 | Nokia Mobile Phones Ltd. | Serial acquisition system with adaptive threshold and optimal decision for spread spectrum systems |
| US5737361A (en) * | 1994-11-22 | 1998-04-07 | Samsung Electronics Co., Ltd. | Receiver in a direct-sequence spread-spectrum communication system using a window filter |
| US5818868A (en) * | 1995-04-14 | 1998-10-06 | Agence Spatiale Europeenne | Direct-sequence spread-spectrum receiver including code acquisition and detection using an autoadaptive threshold |
| US6892175B1 (en) * | 2000-11-02 | 2005-05-10 | International Business Machines Corporation | Spread spectrum signaling for speech watermarking |
-
2006
- 2006-06-19 WO PCT/EP2006/063308 patent/WO2007006623A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5737361A (en) * | 1994-11-22 | 1998-04-07 | Samsung Electronics Co., Ltd. | Receiver in a direct-sequence spread-spectrum communication system using a window filter |
| US5818868A (en) * | 1995-04-14 | 1998-10-06 | Agence Spatiale Europeenne | Direct-sequence spread-spectrum receiver including code acquisition and detection using an autoadaptive threshold |
| EP0756387A2 (en) * | 1995-07-25 | 1997-01-29 | Nokia Mobile Phones Ltd. | Serial acquisition system with adaptive threshold and optimal decision for spread spectrum systems |
| US6892175B1 (en) * | 2000-11-02 | 2005-05-10 | International Business Machines Corporation | Spread spectrum signaling for speech watermarking |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2081187A1 (en) * | 2008-01-21 | 2009-07-22 | Deutsche Thomson OHG | Method and apparatus for determining whether or not a reference pattern is present in a received and possibly water-marked signal |
| US8281137B2 (en) | 2008-01-21 | 2012-10-02 | Thomson Licensing | Method and apparatus for determining whether or not a reference pattern is present in a received and possibly watermarked signal |
| US11183198B2 (en) | 2012-10-15 | 2021-11-23 | Digimarc Corporation | Multi-mode audio recognition and auxiliary data encoding and decoding |
| CN109784126A (en) * | 2017-11-10 | 2019-05-21 | 富士通株式会社 | Data cutting method and device, article detection method and device |
| CN109784126B (en) * | 2017-11-10 | 2022-11-18 | 富士通株式会社 | Data cutting method and device and article detection method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6516079B1 (en) | Digital watermark screening and detecting strategies | |
| JP4560269B2 (en) | Silence detection | |
| EP2175444A1 (en) | Method and apparatus for regaining watermark data that were embedded in an original signal by modifying sections of said original signal in relation to at least two different reference data sequences | |
| JP2003506991A (en) | Auxiliary data detection in information signals | |
| NZ566782A (en) | Neural network classifier for separating audio sources from a monophonic audio signal | |
| RU2482553C2 (en) | Method and apparatus for determining presence of reference pattern in received signal, possibly watermarked | |
| AU2005315882B2 (en) | Apparatus and method for determining a maximum correlation | |
| WO2003087770A3 (en) | Matched filtration with experimental noise determination for denoising, peak picking and quantitation in lc-ms | |
| EP2656341A1 (en) | A method and an apparatus for performing a voice activity detection | |
| JP5605574B2 (en) | Multi-channel acoustic signal processing method, system and program thereof | |
| WO2007006623A1 (en) | Method and apparatus for detecting correlation between a received signal and a candidate spreading sequence | |
| US6865529B2 (en) | Method of estimating the pitch of a speech signal using an average distance between peaks, use of the method, and a device adapted therefor | |
| US20050015241A1 (en) | Method for detecting the quantization of spectra | |
| WO2005078655A1 (en) | Watermark detection | |
| US8374229B2 (en) | Method for the detection and generation of a useful signal and associated devices and communications system | |
| EP2709102A1 (en) | Method and apparatus for determining an optimum frequency range within a full frequency range of a watermarked input signal | |
| Abu-El-Quran et al. | Pitch-based feature extraction for audio classification | |
| US10811032B2 (en) | Data aided method for robust direction of arrival (DOA) estimation in the presence of spatially-coherent noise interferers | |
| Lim et al. | Acoustic blur kernel with sliding window for blind estimation of reverberation time | |
| EP2835799A1 (en) | Method and apparatus for detecting a watermark symbol in a section of a received version of a watermarked audio signal | |
| Ning et al. | A novel echo detection scheme based on autocorrelation of power cepstrum | |
| Abu-El-Quran et al. | Adaptive pitch-based speech detection for hands-free applications | |
| Angelopoulos et al. | Evaluating the impact of spectral estimators on frequency domain feature classification applications for pipe leakage detection | |
| Cao et al. | A modified method for acoustic pulse extraction under noisy environments | |
| CN115811357A (en) | Optical fiber end event positioning method for OTDR (optical time Domain reflectometer) sampling signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| WWW | Wipo information: withdrawn in national office |
Country of ref document: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 06777358 Country of ref document: EP Kind code of ref document: A1 |