TWI342515B - Methods, apparatus and machine accessible mediums for embedding watermarks - Google Patents

Description

1342515 发明, 发明发明: Related Applications Cross-Reference The present invention claims to apply for a US Provisional Patent Application No. 6Q/478,626 and its application on May 14, 2004. Priority is claimed in U.S. Provisional Patent Application Serial No. 6/571 '258, the contents of which are incorporated herein by reference. TECHNICAL FIELD OF THE INVENTION The present invention relates generally to media measurements, and more particularly to methods and apparatus for building watermarks in a compressed digital data stream. [Prior Art] In modern television or radio stations, the compressed digital data stream is typically used to carry video and/or audio material for transmission. For example, in the United States for digital television broadcasting. The Advanced Television Systems c〇mmiUee (ATSC) standard adopts the M〇ving picture Experts Group (MPEG) standard (for example, the Moving Picture Experts Group - 1 'Mobile Map) Like the special group one 2 'moving image expert group one 3, mobile image special group one 4 seeking) 'playing with video content and for engaging with audio content (ie, advanced television system committee) Standard: digital audio dust reduction (AC-3), digital version of the audio compression standard (see 'A-3, also known as d〇1 by Digi tal®) « The 1342515 AC-3 compression standard is based on a perceptual digital audio coding technique that reduces the amount of data required to regenerate the original audio signal while minimizing perceptible distortion.In particular, the AC_3 compacts the cognitive human ear to be unable to perceive changes in the spectral frequency energy at a particular spectral frequency less than the mask energy at the spectral frequencies. The mask energy is characterized by an audio segment based on the characteristics of the tone and similar noise of the audio segment. Different known psycho-acoustic models can be used to determine the mask energy at a particular spectral frequency. Furthermore: the AC rhythm standard provides a multi-channel digital for digital television, high-definition television, digital audio and video, digital video and satellite transmission for broadcasting special sound effects (eg, nuclear surround sound). Audio format (for example, 5.1 channel format). The current TV or radio station uses watermark technology, built-in watermarks for video and/or audio data compressed according to the contraction criteria.

The group is willing to shrink the standard such as the A (a) compression standard and the mobile map. Specialized group advanced audio coding (A

Repeated standards. Typically, the watermark transmission 2 10 codlng' AAC p is uniquely identifying the broadcaster and/or the digits of the order. Typically, the watermark is used at a location (e.g., taken out of the poor 4 - X less than 1 fixed, and due to the location of the body consumption) or the home group can be used to evaluate individual homes and / / live The viewing behavior of the group is to generate rated information. - from ====Print technology is designed to be analogous to broadcast (four) to become an uncompressed digit; conversion analog program residuals " alpha watermark data into the uncompressed 1342515: digital data stream, and Converting the watermarked data machine to an analog format prior to transmission. In the transition towards a fully digital broadcast environment in which the compressed video and audio streams are transmitted from the broadcast network to the regional branch, the watermark data system may need to be built in directly or placed into a compressed digital data stream. in. The existing watermarking technology may need to decompress the compressed digital data stream into a time domain sample, place the watermark into the time domain sample, and recompress the watermarked time domain sample into a watermark. Compressed digital data stream. Such decompression/compression may result in degradation of the quality of the media content in the compressed digital data stream. Furthermore, existing decompression//compression techniques require additional equipment and the delay in the audio component of a broadcast is in some cases unreachable. Furthermore, the method used by the regional broadcast branch to receive the compressed digital data stream from its parent network and to place the regional content through the complex bonding device prevents re-compression of the digital data streams Previously converted from a compressed digital data stream into a time domain (uncompressed) signal. SUMMARY OF THE INVENTION According to one aspect of the present invention, a method of providing a built-in watermark includes: identifying one or more frames that combine a compressed digital data stream to decompress the one or more frames Each of the following is to identify a plurality of sets of conversion coefficients, and to modify the plurality of sets of conversion coefficients to build the watermark. According to another aspect of the present invention, the present invention provides a device for built-in water 0 1342515 printing. : - an identifier for identifying one or more frames combined with a compressed digital data stream; a decompressor 'for decompressing each of the one or more frames to identify the complex number a set of conversion coefficients; and a modifier for modifying the plurality of sets of conversion coefficients to build the watermark. According to another aspect of the present invention, the present invention provides a machine-accessible medium having a plurality of instructions that, when executed, cause a machine to recognize one or more combined with a compressed digit. a frame of data streams, decompressing each of the one or more frames to identify a plurality of sets of conversion coefficients; and modifying the plurality of sets of conversion coefficients to construct the watermark. [Embodiment] Generally, a method and apparatus for constructing a watermark in a compressed digital data stream are disclosed. The methods and apparatus disclosed herein can be used to build a watermark into a dust-reduced digital data stream (i.e., not before decompression of the compressed digital data stream). Thus, the methods and apparatus disclosed herein eliminate the need for compressed digital data streams to be subject to a plurality of decompression/compression cycles. Such complex decompression/compression cycles are typically not exemplified by television broadcast networks. The branch of the road accepts (i.e., 'a plurality of decompression/compression cycles may significantly degrade the quality of the media content in the compressed digital data stream). Prior to broadcasting, for example, the methods and apparatus disclosed herein can be used to decompress a modified discrete cosine transform combined with a compressed digital data stream formatted according to the AC-3 compression standard. ( 1342515

Modified Discrete Cosine Transform (MDCT) coefficient ° The mantissas of the decompressed discrete cosine transform coefficient groups can be modified to be built-in watermarks, which unintentionally increase the compressed digital data stream. Upon receiving the compressed digital data stream, a receiving device (eg, an on-board television metering device at a media consumption location) can retrieve the built-in watermark information from an uncompressed analog output. For example, the output from a speaker of a television set. The retrieved watermark information can be used to identify media sources and/or programs (e.g., broadcast stations) that are incorporated in a media consumption location (e.g., viewed, heard, etc.). The source and program identification information can then be used in a known manner to generate rating information and/or any other information that can be used to assess viewing behavior associated with individual home and/or home groups. A wide range of service providers 1 1 ◦, a TV 1 2 0, a remote control device 2 5 and a receiving device "〇, the broadcast system 1 〇 is measured using a listener measurement system. The components of the broadcast system i 0 0 are connected in any well known manner. For example, the television-2 0 is set in a viewing area located within a home in which one or more people live. 1 5 Π > Φ ^ 5 ◦ 'These-some or all of the '..., an audience measurement research study. The receiving device "; is connected to the set-top box of the television 120 (SetT〇pB〇x: s personal: 象 ' '--a digital camera' a personal camera, a digital video player, and the like. The viewing area 1342515 50 can be viewed by a person located in the viewing area. Among the U or a plurality of homes, the measuring device 1 is constructed by the receiving device 13 〇 Transfer to the TV = 根 = root = signal to identify the viewing information. The test: the member loses .... and provide the viewing information and other tuning and materials to a data for sale ^, 'Fanbeiji sigh 1 8 Q. Phase 1 7 Q can be implemented using any desired combination of hardwired and wireless communication links, including, for example, the Internet 'Ethernet connection, a digital subscriber loop: °The tongue line '- a mobile phone system, a coaxial cable, etc. The data collection device i 8 can be constructed to process and/or store data received by the measuring device 1 400 to generate a rating Information. The service provider can use any of the 11 Implemented by the provider, such as, for example, a cable television service provider 丄2 2, a severance television service provider 1 14 , and/or a satellite television service provider 1 16 . The television i 2 〇 Receiving a plurality of television signals transmitted by the service provider 丄1 through a plurality of channels, and is adapted to process and display television signals provided in any format, such as: national television standards Committee (NTSC) television signal format, high definition television (HDTV) signal format, Advanced Television Systems Committee (

Advanced Television Systems Committee (ATSC) TV signal format 'Phase Alternating Phase (PAL) TV signal format, a Digital Video Broadcasting (DVB) TV signal format in the Association of Radio and Industry (Association of Radio 10 1342515

Industries and Business, ARIB) TV signal formats and more. The user-operated remote control device 1 2 5 allows a user (eg, 'home member 1 60 0 ) to cause the television 1 220 to tune to a signal transmitted on a channel 2 and receive the signal, and causes the television 1 2 Process and present or transmit the program plan or media content contained in the signals to the desired channel. The processing performed by the television i 2 can include, for example, 'removing a video and/or audio component transmitted through the received signal, causing the video component to be displayed on a combined television 1 2 0 Above the screen/display, and causing the audio component to be emitted by the team that combined the TV 1 2 0. The programming content included in the television signal may include, for example, a television show, an advertisement, a video game, a web page, a still image, and/or is currently being served by the service provider. A preview of other programmatic content provided by or provided by the service provider in the future. The components shown in Figure i are shown as separate structures within the broadcast system 100. The functions implemented by some of these structures may be in a single_unit or may use rain. Or more separate components are implemented. For example, although the television i 2 〇 and the receiving device 130 are displayed as separate structures, the television 丄 20 and the 接收 receiving A are set to 1 〇 可以 can be integrated into _ single units (for example 'An integrated digital TV set.' In another example, the electricity 2 0 '6 Xuan receiving device 1 3 〇 and / or the measuring device " 〇 can be integrated into a single unit. In order to evaluate the individual living Hongjie g η β na 豕 into shell 1 60 0//or the watch line of the home group 1342515 is, a watermark built-in system (for example, the watermark built-in system of Fig. 2) can encode uniquely to identify the broadcast by the service provider ii 0 a watermark of the broadcaster and/or the program in the signal. The watermark built-in system can be implemented at the service provider 1 10 such that the plurality of media signals (eg, television) transmitted by the service provider 1 1 Each of the signals includes one or more watermarks. According to the selection of the home member j 6 ', the receiving device i 3 can transmit the media signal transmitted to the shoulder camera C. And receiving the signal, and causing the television to process and present or transmit the channel plan included in the signals or the channel desired by the media, and the measuring device 1 4 Transferred by the receiving device 1 3 to the battery The watermark information is recognized by the video/audio output signal of 1 2 0. Therefore, the measuring device 14 provides the watermark information and other tuning and/or demographic data to the data collecting device 1 through the 4 network 1 7 〇. 8 0. In Fig. 2, an exemplary watermark built-in system 2 includes a built-in device 21A and a watermark source 22〇. The built-in device 2^1 0 is constructed by the watermark source 2 2 The watermark information 2 3 〇 is inserted into a compressed digital data stream 2 4 。. The compressed digital stream 2 4 0 can be compressed according to an audio compression standard, edge frequency compression Standards such as the AC_3 standard and/or the Moving Picture Experts Group, and the Moving Picture Experts Group-Advanced Audio Coding (MPEG-AAC) compression standard, either of which can be used A block of audio signals is processed by a pre-empted number of digitized samples from each block. The source of the 12 1342515 compressed digital data stream 2 (not shown) may be one, For example, a rate of 4 8 kHz is sampled to form an audio block, as described below. Typically, audio compression techniques such as those according to the AC-3 compression standard use overlapping audio blocks and the modified dispersion. Cosine transform (M〇di f丨ed

The Discrete Cosine Transform (MDCT) algorithm converts an audio signal into a compressed digital data stream (e.g., the compressed digital data stream 2 4 0 of Figure 2). Two different block sizes (i.e., short blocks and long blocks) can be used based on the dynamic characteristics of the sampled audio signal. For example, the AC-3 short block can be used to minimize the pre-echo for the transient segment of the 5H audio scream, and the AC-3 long block can be used to achieve the non-transient of the audio signal. High compression gain of the state segment. According to the AC-3 compression standard, an AC-3 long block is derived from a block of 51 time-domain audio samples by a modified discrete cosine transform algorithm. . In the case of the AC-3 long block, the 51 1 time domain samples are connected by a block of a previous (old) 256 time domain samples and a current (new) 2 The block of 5 6 time domain samples is then followed by 'to generate a 5 1 2 time domain sample audio block. The ac-3 long block contains 256 conversion coefficients. According to the same standard, an Ac~3 short block is derived from a pair of overlapping overlapping time-domain sample blocks of audio. Each sample block contains 256 samples. The AC-3 short block contains modified discrete cosine transform coefficients for these time domain short blocks. The time domain knows that the block produces 1 2 8 coefficients each time resulting in a total of 2 5 6 coefficients combined. Therefore, both the AC-3 long block and the AC-3 short block 13 1342515 contain the same number of modified μ m # & The root 摅1 moving image expert group is an advanced one. According to the one short block system, there are 1 2 8 彳H η 1 cases, η 〇 / w samples 'and a long block system contains 0 2 4 sample. As is well known, audio compression technology reduces the number of bits needed to dry an audio signal. In the example of Figure 3, an uncompressed digital data stream containing a plurality of 256 samples of the time domain audio block 31〇^ is generally shown as AO, Al, Α2 A3 Α4 « , '丁', And A5. The modified discrete cosine transform > ambiguous method processes the ambiguous audio 抽 n - 曰 neck & block 3 1 〇 to generate a modified discrete cosine transform coefficient group, which is generally displayed as (4), 隠, ΜΑ3, ΜΑ4 &, (its t, ·5 silk display). For example, the modified discrete cosine transform algorithm can process the audio blocks Α0 and Α1 to produce the modified discrete cosine transform coefficient group μα〇. 'The audio block Α0 and A1 are concatenated to produce a 5 丄 2 sample audio block that has been modified by the discrete cosine transform using the modified sigma = cosine transform algorithm to produce 2 5 6 The modified discrete cosine transform coefficient group MA0 is modified by a discrete cosine transform coefficient. Similarly, the audio blocks A1 and A2 can be processed to produce the modified discrete cosine transform coefficient set MA1. Therefore, the audio block A1 is an overlapping audio block 'because it is used to generate the modified discrete cosine transform coefficient sets Μ A 0 and M A1 . In a similar manner, the modified discrete cosine transform algorithm is used to convert the audio blocks A2 and A3 to produce the modified discrete cosine transform coefficient group MA2, converting the audio blocks A3 and A4' to generate The modified discrete cosine transform coefficient group MA3 is modified, and the audio block A4 λ ^ ΜΑ 4 〇Α 〇Α 〇Α 5 is converted to generate the modified discrete cosine transform coefficient group _ 匕 ° 曰 曰 区 系 2 system is an overlapping audio Block = is used to generate the modified discrete cosine transform coefficient group leg and the day i ghost A3 is an overlapping audio block, which is used to generate the modified discrete cosine transform coefficients Μ MA2 and MA3, The audio block Α 4 is an overlapping audio block that is used for generation. Hai modified the discrete cosine transform coefficient groups ΜΑ3 and ΜΑ4, etc. [This modified discrete cosine transform coefficient group 3 2〇 together forms the I-reduced digital data stream 2 4 0. As will be described in detail below, the built-in device 2 of FIG. 2 can be built or placed into the compressed digital data by the watermark information or watermark 2 3 0 from the watermark source 2 2 . Stream 2 4 〇. The watermark 230 can be used, for example, to uniquely identify the broadcaster and/or program so that media consumption information (e.g., viewing information) and/or rating mechanism can be generated. Therefore, the built-in slot 2 1 产生 generates a compressed digital stream 250 0 for transmission. In the example of Fig. 4, the built-in device 2i includes an identification unit 4 1 〇 ' a decompression unit 420, a modification unit 430 and a recompression unit 464. Although the operation of the built-in device is described below in accordance with the AC-3 compression standard, the built-in device is configured to operate in an additional or other compression standard, such as by way of example. , the mobile image expert group - advanced audio coding compression standard. The operation of the built-in device 210 is described in more detail with reference to Fig. 5. : Open % 'The identification unit 41 〇 is constructed to identify the digitized data stream 2 4 〇 one of the numbers n U of the $ 个 box 5 1 〇 'The compressed / stream 2 4 〇 - part It is shown as A and B in Fig. 5 by way of illustration. As described above, the compressed digital data stream u 糸 can be a digital data stream compressed according to the AC-3 standard (below

: Called AC-3 data stream). Although the AC_3 data stream 2 4 can contain a plurality of channels, for the purpose of _, 言 > B aA ΛΓ 0 - 4 4 for the purpose of 间 / ,, the following example describes the - lean stream 2 4 0 is Contains only - one channel. The Ac_3 data stream 2 4 〇 巾 ' g some boxes 5 1 Q each of the series contains a plurality of modified discrete cosine transform coefficient groups 5 2 Q. According to the cardiac compression standard, for example:: Each of the blocks 51 includes six modified discrete cosine transform coefficient groups (ie, six "audio blocks,"). For example, box A The modified discrete cosine transform coefficient groups _, mA1, MA2, MA3, MA4, MA5 and MA6 are included, and the frame B contains the modified discrete cosine transform coefficient groups ΜΒ0, ΜΒ1, ΜΒ2, ΜΒ3, Μβ4 MB5 and MB6. The identification unit 4 1 is also constructed to identify the "header information" associated with the parent of the box 5, such as, for example, the number of channels in combination with the ac-3 data stream. Although the exemplary AC-3 data stream 2 4 contains only one channel as described above, an exemplary compressed digital data stream having a plurality of channels is described below in connection with Figures 7 and 8. Referring again to FIG. 5, the decompression unit 4 2 构成 is constructed to decompress the modified discrete cosine transform coefficient group 5 2 〇 to determine the compression information 16 1342515. The modified discrete cosine transform coefficient is decompressed into a pseudo number ^ And an index such that mk = Mk * 2-n. The number of bits used to represent the alias Mk of each of the modified discrete cosine transform coefficient groups 5 2 修改 modified discrete cosine transform coefficients may be based on a known quantized look-up table 6 0 0 published in the 3 compression standard (For example, the quantization look-up table 6 of Figure 6) is determined. In the example of Figure 6, the quantization look-up table 6 提供 provides false numbers or bit patterns and false values for modified discrete cosine transform coefficients represented by a 4-bit number. As described in more detail below, the alias Mk can be changed (e.g., increased) to represent a modified value of the modified discrete cosine transform coefficient to have a built-in watermark in the AC-3 data stream 2 4 0 in. Referring to FIG. 5, the modification unit 430 is constructed to perform inverse conversion of each of the modified discrete cosine transform coefficient groups 5 2 0 to generate a time domain audio block 5 3 0 'by being exemplified as TAO, TA3'', TA4', TA4'', TA5', TA5, ', TBO', TBO,, TB1, TB1,, and TB5, (TAO, to TA3, and TB2, to TB4, ' Not shown). The modifying unit 430 performs an inverse conversion operation 'to generate a previous (old) time domain audio block (which is represented as a first-order block) combined with the 256 sample time domain audio block and The current (new) time domain audio block (which is represented as a second-order block) group 'the 2 256 sample time-domain audio blocks are concatenated' to form a s-AC AC-3 data stream 2 4 The modified discrete cosine transform coefficient group 5 2 0 is modified by 0. For example, the modifying unit 430 performs an inverse transformation on the modified discrete cosine transform coefficient group MA5 to generate time domain blocks TA4', and TA5, for which the modified discrete cosine transform system 18 1342515 is arrayed. The MBO implements an inverse transformation, and generates a domain block D5, and TB0, for which the modified discrete cosine transform is performed, and the coefficient group is subjected to an inverse transformation to generate a % domain block Τβ〇. ,, and _ , , π , & TB1 usurped. In this way, the modification is early: 糸 generates a reconstructed time domain audio block 5 4 〇, which provides the member of the original time domain audio block of the AC: _3f stream 2 4 Q in order to generate a far reconstruction The time domain audio block HQ, the modification unit 430 can be based on, for example, the known Bradley time domain semaphore cancellation (Ding De D(10) ain Alias

CanCellatl〇n, TDAC) technology is added to the time domain audio block. The Princen-Bradley time domain false signal cancellation technology is described in the sound, voice and signal processing ASSP reported by the president of the Institute of Electrical and Electronics Engineers (EEE). - 35, No. 5, No. 1153 - 1161 (1986), entitled "Analysis of Analytical/Synthesized Filter Bank Design Based on Time Domain False Signal Elimination". For example, the modification unit 4 3 The time domain audio block TA5 can be reconstructed by adding the first-order time domain audio block TA5' and the second-order time domain audio block TA5' by using the Princen-Bradley time domain false signal cancellation technology. That is, TA5R). Similarly, the modification unit 430 can use the Princen-Bradley time domain false signal cancellation technology to combine the first-order time domain audio block TB0 and the second-order time domain audio block TB0'. 'Adding and reconstructing the time domain audio block TB0 (ie, 'TB0R). In this way, the original time domain audio block used to form the AC-3 data stream 2 4 0 is reconstructed without solution Compress the AC-3 data stream 2 4 0 (to be detailed below) In detail, this means that the watermark 2 3 0 can be directly built in or inserted into the AC-3 data stream 2 4 0. 19 1342515 The modification unit 4 3 Q is also constructed to form a time domain audio to be reconstructed. Block 5 4 0 is generated and dropped into the audio block 5 5 ◦, which is substantially, water: time domain TA0W) TA4W, TA5W, TB0W, TB1W and TB5W are shown as TA: W, Jane, TA3W, Zhan Cong and Lan Xing are not shown in the second, the watermark 2 3 〇, the modification unit 4 3 藉 is connected by solid: = reconstructed time domain audio block to generate - 5 i 2 samples Vertical = „Generate—a modifiable time domain audio block. For example, ^: early: 4 3 0 can be concatenated with the reconstructed time domain: > hired to produce - 512 sample audio blocks. Then, the unit can insert the watermark 23〇 into the second=change 葙F for the time domain reconstructed in the field::° and the 512 sample audio block formed by the job to generate the watermark The time domain audio block is bribed and fine. As described in the U.S. Patent No.... I!, 870 and 6,621,881 are code programs that can be used to insert the watermark 2300 into the reconstructed time domain (4) block 54〇1 U.S. Patent No. 6, 272, 176, 6, 504, 870, and 6/621 '881, the disclosure of which is incorporated herein by reference.

σ is inserted into a 512 sample audio block. For example, J ^ / 3⁄4 - 4 solid ^ I solid sample audio block system with a bit built-in or inserted watermark 9 q A JLA . z 3 0 - shell material. In particular, the special frequency having the indices f 1 and f 2 can be modified or increased to insert a data bit that is combined with the water mark 2 3 〇 of the 20 1342515. One . In order to insert a bit "1", for example, the power at the first spectral frequency of the combined index f^.x can be increased or increased to become - 0^ a frequency adjacent (eg, by The maximum value of one pupil power within "- 2, ίΗ, n, n + i and the frequency adjacent to fl'0 .z). This is the second spectral frequency of the combination index f2. The power of the system can be reduced or increased to become a spectral power guide within a frequency proximity (eg, a frequency proximity defined by f2-9 f〇, ' , i2, f2 + l, and ί2 + 2) | +min. Conversely, in order to insert a bit two u Γ \ >» 1 兀◦ , the power system at the first spectral frequency combined with φ ή丨f ! ', f 1 becomes a The regional spectral power is minimum and the combined index [2, the power at the second spectral frequency is increased to a regional spectral power maximum. The return ^ 5 map' is based on the watermarked time domain audio block 5 5 0 4 changed earlier, 4 3 0 system produced a modified discrete cosine transform system with watermark Array 5 6 0 , which is shown by MA as MA0W.MA4W, MA5W, MB0W and MB5W. V block MA1W, MA2W, MA3W, MB1W, MB2W, MB3W ^ m ^ ) After the above example, the modification unit 4 3 〇 based on the watermarked time domain audio blocks TA5W and TB0W to generate a modified discrete cosine transform coefficient group MA5W with a watermark of the righteous h1 degree king 4. Clearly, the modification unit 4 3 The watermarked time domain audio blocks TA5W and TB〇w are connected to form a 51 1 sample audio block, and the 5 } 2 sample audio blocks are converted into the watermark modified modified cosine transform. A coefficient group, which can replace the original modified discrete cosine transform coefficient group ma5 21 ^ 42515, which is described in detail below.

The difference between the modified discrete cosine transform coefficient set 5 2 〇 and the modified off-gas cosine transform coefficient set 5 6 系 represents the heart data/2 :: in the _ item due to built-in or insertion of the watermark 23 :: change. In the description of Fig. 6, for example, the modifying unit 4 3 modifies the discrete cosine transform coefficient group MA5 by replacing the modified discrete cosine transform coefficient group MM with its corresponding watermark. A quantized lookup table (e.g., the reading 6 of FIG. 6) can be used to determine a new dummy of the modified discrete cosine transform coefficient combined with the modified discrete cosine transform coefficient set of the watermark. A numerical value to replace the old false value of the modified discrete cosine transform coefficient combined with the watermark modified modified cosine transform coefficient set 5 2 0 . Therefore, the new false value represents a change in the AC-3 data stream 2 4 0 or the increase in the AC-3 data stream 2 4 0 due to the built-in or insertion of the watermark 2 3 0 .

Figure 6 of the Qing Mai Review, the quantified table 6 of the demonstration contains the pseudo-numbers used for a demonstration (the 25 digits of the quantified false numbers and false values, which are at -0.93 to +0 The scope of 9333. Although the exemplary quantization look-up table 60 provides the information of the modified discrete cosine transform coefficients in combination with the 4-bit, the AC_3 compression standard provides a combination of each modification. A quantized look-up table of other suitable number of bits of the discrete cosine transform coefficients. To show that the modified unit 430 can have a special modified dispersion having an alias Mk included in the modified discrete cosine transform coefficient group MA5 One way of the cosine transform coefficient mk, assuming that the original denominator is -〇. 2666 (ie, -4/1 5 ). Use this quantization to look at 22 1342515 Table 6 Ο Ο 'corresponds to the modified discrete cosine transform In the coefficient group MA5, the special modified discrete cosine transform coefficient is determined to be 〇1〇1°. The watermark modified modified cosine transform coefficient group MA5W contains one watermark with one The value WMk is modified by the discrete cosine transform coefficient wmk. Furthermore, the new false value of the modified discrete discrete cosine transform parameter wmk of the watermark modified modified discrete cosine transform coefficient group MA5W is assumed to be _ 〇· 4300, which is between the fake numbers 0 011 and 01 0 0. In other words, in this example, the watermark 2 3 0 causes the original false value 2666 and the watermarked false value _ ( 0 A difference between 4300 - 667 1667. In order to build or insert the watermark 2 3 0 into the AC-3 data stream 2 4 0 'the modification unit 4 3 0 can be modified or increased The modified discrete cosine transform coefficient in the discrete cosine transform coefficient group MA5 is modified and the modified discrete cosine transform coefficient group MA5 is replaced by the watermark modified modified cosine transform coefficient group MA5. Following the above example, the fake digital 0011 Or the pseudo-digital 〇1〇〇 can replace the pseudo-digital 〇1〇1 combined with the modified discrete cosine transform coefficient mk (that is, one of the modified discrete cosine transform coefficient groups MA5 modified one of the discrete cosine transform coefficients) Because the watermarked false number is combined with the modified discrete cosine transform coefficient w (i.e., one of the modified discrete discrete cosine transform coefficient groups MA5W having the watermark modified) The false number between the fake numbers 0011 and 0100 (that is, the false value corresponding to the modified discrete cosine transform coefficient wmk having the watermark is 0.4300). The false value corresponding to the false number 11 is _〇.5333 (23 丄 342515, ie -8/15) 'and the false value corresponding to the false number 〇1〇〇 is _ 0.4 (ie, -6/15). In this example, the modifying unit 4 3 selects the fake digital 0100 instead of the fake digital 0011' instead of the original false number 〇1〇1 combined with the modified discrete cosine transform coefficient mk, because the corresponding = the fake digital The false value of -100 is -0.4 is the closest to the desired water = false value -0.430 (^, therefore, the new dummy digit of the watermarked pseudonym WMk corresponding to the modified discrete cosine transform coefficient ww with watermark The meta pattern 0100 replaces the original pseudo bit pattern 〇1〇1. Similarly, each of the modified discrete cosine transform coefficients in the modified discrete cosine transform coefficient group MA5 can be modified in the manner described above. If a new (ie, watermarked) false value is outside the quantified range of the false value (ie, greater than 〇·9333 or less than -0· 9333), then the positive threshold value of 111〇 or 〇 The negative threshold value is selected as the new false number. In addition, although the pseudo-digital system which has modified the discrete cosine transform coefficient in combination with a modified discrete cosine transform coefficient group can be used as described above. The index of the modified discrete cosine transform coefficient is maintained unchanged. The recompression unit 410 constructs a watermark modified discrete cosine transform that recompresses each of the boxes of the ac_3 lean stream 2400. a set of coefficients 560 for transmission. In particular, the recompression unit 404 identifies the location of each modified discrete cosine transform coefficient group within a frame of the AC-3 data stream 2404. So that the corresponding watermarked modified discrete cosine transform coefficient set can be used to replace the tamper-modified phosphorous cosine transform coefficient. To reconstruct a watermarked version of block A, for example, the recompression unit 4 The system can identify the position of the modified discrete cosine transform coefficient group MAO to MA5, and replace the corresponding watermark modified modified cosine transform coefficient group MA0W to MA5W in the corresponding identified position. In the above decompression, modification and recompression process, the AC-3 data stream 240 maintains a compressed digital data stream, and the watermark 2300 is built in or Placed in the AC-3 data stream 240. Therefore, the built-in device 2 1 inserts the watermark 2 300 into the AC-3 data stream 240 without the need for a month & The decompression/compression cycle of the amount of quality of the media content within the AC-3 data stream 240. For the sake of simplicity, the AC-3 data stream 240 is described in conjunction with Figure 5 to include A single channel. However, the methods and apparatus disclosed herein can be applied to a compressed digital data stream having a plurality of audio blocks, such as 5 丨 channels (i.e., 5 full-band channels). As described below. In the example of Figure 7, an uncompressed digital stream 7 〇 can contain a plurality of audio block groups 7丄〇. Each of the audio block groups 7 i can include an audio block combining a plurality of channels 720 and 730, and a plurality of channels 72 and 73 include an example and tr 'one left front channel, one right front channel, one Central channel, a left surround channel, a right surround channel, and a low frequency effect channel (for example, a subwoofer channel). For example, the audio block group AUD0 is a singularity of the audio block A 0 L combined with the left front channel, and a frequency block A0R combined with the right front frequency. Combining the audio block A0C' of the intermediate channel, an audio block ship combined with the left surround channel, an audio block combined with the right surround channel, and an audio block of 25 1342515, .° σ 4 low frequency effect channel AOLFE. Similarly, the audio block group AUD1 includes an audio block A丨L combined with the left front channel, an audio block A1R combined with the right front channel, and a frequency block A1C combined with the intermediate channel, a combination. The audio block A1SL' of the left surround channel is an audio block a sr combined with the right surround channel and an audio block MLFE combined with the low frequency effect channel. Each of the audio frequency blocks combined with one of the specific audio block groups 7 1 可以 can be processed in a manner similar to that described above in connection with Figures 5 and 6. For example, the audio block of the central channel § 1 大致, which is generally shown as A 〇 c, Aic, A2c, and 5, 8 can be converted to produce a compressed data stream. These modified discrete cosine transform coefficient groups 8 2 〇. As described above, each of the modified from the set of cosine transform coefficient sets 8 2 可以 can be made by concatenating - a previous (old) 2 5 6 sample audio block (new) ) 2 5 6 sample audio blocks are formed by a 5 i 2 sample audio block. The modified discrete cosine transform coefficient algorithm can then process the time domain audio block 8 i 〇 (eg, a 〇 c to Α 5 〇 to generate the modified discrete cosine transform coefficient sets (eg, m 〇 C The modified discrete cosine transform coefficient group 8 2 〇 ' according to the dust-reduced digital data stream 8 〇 'the identification unit 4 system identifies a plurality of boxes (not shown in combination with each of the above-mentioned frames) Header information. The header is included in the data stream 8 Q Q (four) reduction ^. For each box... the solution is reduced to 4 2 _ _ _ this 26 1342515 modified discrete a cosine transform coefficient group 8 2 〇 to determine the compression information in combination with the modified discrete cosine transform coefficient group 8 2 。. For example, the metamorphic G contract unit 4 2 0 can be identified by the original compression program The number of bits used to represent the alias of the discrete cosine transform coefficient modified for each of the modified discrete cosine transform coefficient sets 820. Such compressed information can be used as described above The watermark 2 built in conjunction with Figure 6 3 0. The modification unit 4 3 0 is then generated usually for the person 0 (: '', Ding 8 1 (: ', person 1 (:, ', Ding 2 (:,, Ding 2 (:, And 7^3(:, inversely transform the time domain audio block 8 3 0. The time domain audio blocks 8 3 0 contain a set of prior (old) time domain audio blocks (the system represents a first-order block) and a set of current (new) time-domain audio blocks (which are represented as second-order blocks). The corresponding is by the Perncen-Bradley time-domain false signal cancellation technique by way of example. The first-order block and the second-order block are added, and the original time-domain channel block that is compressed to form the AC-3 digital data stream 80 〇 can be reconstructed (ie, the reconstructed time-domain audio block 8 4 For example, the modifying unit 430 may add the time domain audio blocks τ A1C, and TA1C'' to reconstruct the time domain audio block TA1C (ie, TA1CR). Similarly' The modifying unit 430 may add the time domain audio blocks TA2C' and TA2C'' to reconstruct the time domain audio block TA2C (ie, TA2CR) 0 for the watermark source 2 The watermark 2 3 0 is inserted, and the modifying unit 4 3 is connected in series with two adjacent reconstructed outdated domain audio blocks to generate a 51 1 sample audio block (ie, a modifiable time) For example, the modifying unit 430 can serially connect the 27 1342515 reconstructed time domain audio blocks TA1CR A TA2CR, and the reconstructed time domain audio blocks TA1CR and TA2CR Each line is a short area of 2 5 6 samples; ghosts are formed to form a 5 ! 2 sample audio block. The modifying unit 430 adds the watermark 2 3 〇 to the 5 working 2 sample audio blocks formed by the reconstructed time domain audio blocks TA1CR and TA2CR to generate the watermarked time domain audio. Block TA] and according to the watermarked time domain audio block 8 5 ◦, the modifying unit 430 can generate the modified discrete cosine transform coefficient group with watermark 860 〇 For example, the modifying unit The watermarked time domain audio blocks TA1CW and TA2CW may be concatenated to generate the watermark modified modified cosine transform coefficient group M丨cw. The modifying unit 43 modifies the modified discrete cosine transform coefficient group 8 2 by replacing each of the modified discrete cosine transform coefficients with a corresponding watermark modified modified cosine transform coefficient set 860 . For example, the modification unit 430 may replace the original modified discrete cosine transform coefficient group Mic with the watermark modified modified cosine transform coefficient set M1CW. Next, the modifying unit 430 may repeat the above procedure for combining the audio blocks of each channel to insert the watermark 2 3 之中 into the compressed digital data stream 800. Figure 9 is a flow diagram showing the manner in which the exemplary watermark built-in system of Figure 2 can be constructed to embed or insert a watermark into a compressed digital data stream. The exemplary program of Figure 9 can be implemented as any machine accessible instruction using a plurality of different code stored on any combination of machine-accessible media, the machine-accessible medium 28^42515 The system is like a volatile or non-volatile memory (for example, a 妒 妒 其他 其他 其他 其他 其他 其他 其他 , , , , , , , , , , , , 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 In the accessible accessory body, 'L is embodied in 1 machine · Integral, programmable array, - special application mail $路' can erase the programmable read-only memory, a read only (10) &quot A random access memory, a magnetic medium, a light and medium 2 and/or any other suitable media form. Furthermore, the order of the media is shown in Figure 9, and these trips 4 are implemented. Again, the flow chart 9 Q is only set and described in conjunction with the elements of Figures 2 through 5, as an example of a method of constructing a system with a built-in watermark in a compressed digital data stream. In the example of Figure 9, the program begins with the identification unit 4 (Fig. 4) identifying a frame that combines the compressed digital data stream 2 4 (Fig. 2), such as box A ( Figure 5) (Box 9 work). The identified frame may comprise a plurality of modified discrete cosine transform coefficient groups formed by overlapping and concatenating complex vector blocks. According to the ac_3 compression standard, for example, a frame can contain six modified discrete cosine transform coefficient sets (i.e., six "audio blocks"). Furthermore, the identification of Zhuoyuan 4 1 0 (Fig. 4) also identifies the header information associated with the box (block 9 2 0). For example, the identification unit 4 1 can recognize the number of channels combined with the compressed digital data stream 2 4 0. Next, the decompression unit 410 decompresses the plurality of modified discrete cosine transform coefficient groups to determine the I contract information used to generate the compressed digits 29 1342515 in combination with the original compression program (block 9 is discrete) The residual compression unit 4 2 0 identifies each parent of a modified conversion coefficient group, and the parent is a modified cosine transform coefficient index system. Then, the variable cosine transform coefficient is changed: the decompression is changed. Unit 4 2 〇7 4th = reduction standard is grouped 420 (different 4 figure) also determines the sound used to modify the discrete cosine transform coefficient of the number, so that θ: ... shrink standard specified by a suitable The quantified table can be =, > 戍 增大 增大 增大 增大 增大 增大 增大 增大 增大 增大 增大 增大 增大 增大 结合 结合 结合 M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M This is described in more detail below in conjunction with Figure 1. ', Heart Map' This modification procedure begins with the use of this

Lrn Figure 4) Implementing the modified discrete cosine transform system to convert the second heart to produce an inverse transformed time domain audio block (block 1 0 1 0). In particular, the modification unit * ^ U produces "- ia A afr (the old " capture domain audio block (for example, the mountain., s „ DW table is not a first-order block) and a combination of The resulting (new)%-domain audio block for each of the six sample source-time time-domain audio blocks (which is the table 3 circle) One of the second orders is illustrated in conjunction with Figure 5. For example, the modification: 4 3 0: 糸 can be modified by the discrete cosine transform coefficient group axis 5, and generate TA4 纟 TA5 'by these modifications Over-discrete ΜΒΟ, ΤΒΟ,, produce TA5, and 丁β〇, and by, 糸, - coefficient groups produce 贤 and Τβ1, for discrete cosine transform mother-time domain audio block 30 1342515 'this modification The unit 430 adds the corresponding one-order block and the second-order block to reconstruct the time-domain audio block according to, for example, the Princen-Brad 1 ey time-domain sniffer cancellation technique. The first-order block TA5' and the second-order block TA5'' may be added to reconstruct the time domain The frequency block TA5 (that is, the reconstructed time domain audio block TA5R) 'and the first order block TB0, and the second order block TB0, 'can be added' to reconstruct the time domain audio block TB0 (ie, the reconstructed time domain audio block TB0R). To insert the watermark 2300, the modification unit 430 uses the reconstructed time domain audio block to generate a modifiable time domain audio block ( Block 1 0 3 0 ). 5 Xuan Modification Unit 4 3 0 uses two adjacent reconstructed time domain audio blocks to produce a modifiable 5丄2 sample time domain audio block. For example, The modifying unit 43 can generate a modifiable time domain audio block by concatenating the reconstructed time domain audio blocks tA5r and tb〇R of FIG. 5. Applying an encoding program, such as an example For one or more of the coding sequences described in U.S. Patent Nos. 6,272,176, 6,504,870 and 6,621,881, the edge modifying unit 430 will be from the watermark source 2 2 0, the watermark 2 3 0 is inserted into the repairable time domain audio block (block i 4 〇). For example, the modifying unit 43 can make The watermark 2 3 〇 is inserted into the 5 丄 2 sample time domain audio block by the reconstructed time domain audio blocks TA5R and TB0R to generate the watermarked time domain audio blocks fA^5W and TB0W. According to the watermarked time domain audio block and the & 'Bei-ίΐ»$Modification Unit 4 3 0 system generates the modified discrete cosine to 31 3142515

The way it can be handled. The exemplary box processing program 1 begins with the built-in device 2 1 0 reading the header information (block 1 1 1 〇) of the obtained frame (for example, an AC_3 box), and initializing The count of a modified discrete cosine transform coefficient group is 0 (block 1 1 2 〇) ^ In the case where an AC-3 box is processed, each Ac_3 frame contains 6 modified discs with compressed domain data. Cosine conversion coefficient group (for example, MA0, MA1, MA2, MA3, MA4, and MA5 in Fig. 5, which is also referred to as "audio block" in the Ac_3 standard). Therefore, the built-in device 2 1 〇 It is determined whether the modified discrete cosine transform coefficient group has a count of 6 (block 1 1 3 0 ). If the modified discrete cosine transform coefficient has not been equal to 6, thereby indicating that at least one modified discrete cosine transform coefficient set needs to be processed, then the built-in device 2 extracts a modified discrete cosine combined with the frame. Conversion factor) (for example, above in conjunction with Figure 6

The index of the original false number Mk) (block i i 4 〇 ) and the false number (block 1 1 5 0 ). The built-in device 2 calculates a new pseudo-number combined with a code read in block 1 2 2 0 (eg, the new dummy WMk described above in connection with FIG. 6) J6 〇), and modify the original false number (square ghost 1 1 7 0 ) combined with the box according to the new fake number. For example, 5, the original false number system can be replaced by the new pseudo number... The device 2 i is used to increase the count of the modified discrete cosine transform coefficient group by 1 (block i 18〇), and the control system returns to the block. 1 1 3 Q. Although the exemplary procedure of the (4)1 1 diagram is as described above to include six modified discrete cosine transform coefficients 35 (for example, (4) the threshold value of the count of the discrete cosine transform coefficient group is changed to 6), _ is used more 34 1342515 =: The program of the discrete cosine transform coefficient group can be used by the leaf 娄 "" 13 0" If the modified discrete cosine transform coefficient group is at 6, then all the modified detached are processed, 佶锃,c will re-compress the box (block 1 1 g 〇). As mentioned above, many of the methods that are to be displayed in a watermark audio signal, especially those that are known by the system, are known (for example, 'an inaudible method is known. For example, - A known method is described in the patentee's je calculation " ' _ _ _ people (4) patent No. 6, 421, 445. A disclosure is hereby incorporated by reference. In particular, the United States patent As stated in the case of No. 6,421,445, a code ^歹'^ 'a watermark) can be included in the audio signal of 10 at a different frequency = α, which can be used by a detector to use a sequence of audio samples. The #列列列列篇名 The Fourier spectrum analysis of the 1 2 2 8 8 audio transcripts as detailed below. For example, a vertical frequency signal can be transmitted at a rate of 48 kHz. Is sampled to output one; = processed (eg 'using a Fourier transform') to obtain a fairly high resolution (eg, 3.9 Hz) frequency domain of the uncompressed = chirp signal representing ~2 2 8 8 audio samples of audio Sequence. However, according to Jensen: the method revealed by man The encoding process, a sinusoidal code signal having a fixed amplitude across the sequence of the entire audio sample, cannot be accepted because the chord signal can be perceived by the human ear. To satisfy the mask energy; J (ie, ' In order to ensure that the sinusoidal signal information is maintained, the sinusoidal signal is encoded in the entire 12288 audio samples 2, 35 I342515 is synthesized by using a mask energy analysis, and the mask energy analysis is determined in the audio sample. The sinusoidal amplitude of each region within each block (eg, where each block of audio samples can contain 5 i 2 audio samples). Therefore, the sinusoidal waveform of the region can be 丄2 2 8 8 The sequence of audio samples is coherent (in phase) throughout the period. Compared to the method disclosed by Jensen et al., the methods and apparatus described herein can be used to compress digital bits containing the compressed audio signal. The data stream maintains the compressed side 4 during the decompression, modification and recompression process, and the watermark or other code signal is built into a compressed Among the audio signals, the i 2 image shows a case where a watermark such as that disclosed by the person can be inserted into a compressed frequency signal. The exemplary program 丄 2 begins at - The box count becomes 0 (block i 2 i Q). 8 boxes representing 1 2 2 8 8 audio samples per channel, such as m box, can be processed to have one or more code symbols (For example, as shown in Figure i3 - or multiple characters "〇", ''", "s,, and "E") are built into the audio signal. Although the compressed digital data stream is described herein as comprising 1 2 288 audio samples, the compressed digital data stream may have more or fewer audio samples. In particular, the built-in device 2 1 第 (Fig. 2) can read the one from the watermark source: print 2 3 0 ' to inject one or more code symbols into the sequence of the frame" 2 0 The built-in device 2 1 Q can obtain the blocks - (party 1230) 'and proceed to the frame processing operation η" as described above to process the obtained frame. Therefore, the exemplary frame handler Work 36 3642515 0 is over, and the control returns to the block! 2 5 Q to increase the count of the box by 1. The build device 2 i determines whether the count of the box is 8 (block 1 2 6 0 If the count of the box is not 8, then the built-in device 2 1 is returned to obtain another pivot of the sequence, and the exemplary frame processing operation described above in connection with FIG. 1 is repeated. 〇〇, to process another box. Conversely, if the count of the box is 8, the built-in device 2 1 Q returns to the block 1 2 1 Q, to re-initialize the count of the box to 0, and repeat The program 1 2 0 0 to process the frame of another sequence as described above, a code signal (eg, the watermark 2 3 〇) is Built into or inserted into the compressed digital data stream (eg, an AC-3 data stream). As shown in Figure 3, the exemplary table 丄300, the sfL number can contain 1 Q corresponds to a combination of sinusoidal components of frequency indices f 1 to Π 0, which represent 4 _ _ _ _ _ _ _ _ _ _ _ _ _ It is wide that the code ''0' can represent a zero-two carry value, and the code "factory" can represent a 1-two carry value. Furthermore, the code "J-book b can represent a The beginning of the message, and the code "E" can represent the end of the message. Although only the four symbols "more or less code characters" can be used in Figure 13. , the 5 1 2 samples are too φ, and the central frequency index (for example, 10, 1 2, 1 4 , 1 6 , 1 8 , 2 〇 2 2, 2 4 ' 2 6 and 2 8 ) is combined with the compression. A low-resolution frequency domain representation of the 铋 _ 妁 digital data stream, and an index of the frequency of the 256 data samples (eg ' 2 4 〇, 2 8 8 , 3 3 6 3 R / 1 4,4 32,371,342,515 billion

, 480, 528, 576, 624; ? R7q, 〆 匕 4 and 6 7 2 ) are combined with a high resolution frequency domain representation of the compressed digital data stream.

As described above, the ##-one code character can use the 10 combined noise index shown in the table 3 〇 〇 :. To the sine component of "〇. If not in the exemplary table 1 300, each of the frequency indices η to η 具有 has one or each of the 1 2 28 8 sample center frequency indices or adjacent to the i 2 28 8 The unique frequency value of each of the sample center frequency indices. For example, a code signal for injecting or constructing the code symbol 包含 contains 10 values corresponding to the frequency index 2 3 7, 2 8 9, 3 3 9 ' 3. 8 3, 4 2 9, respectively. 4 8 1,5 3 1,5 7 5 '62 1 and 6 7 3 sinusoidal components. Similarly, one for injecting or constructing the code symbol "Γ" code signal contains i 〇 one corresponding to the frequency index 2 3 9 , 2 9 1 , 3 3 7 , 3 8 work, 4 3 丄, 4 83, 529, 573, 623 and 675 sinusoidal components.

Each of the 10 sinusoidal components of the combined frequency index f 1 to f 可以 can be synthesized in the time domain using the methods and apparatus described herein. For example, the code signal used to inject or build the code "〇" can contain sinusoids cl(k), c2(k), c3(k), C4(k), C5(k), C6( k), c7(k)' c8(k)' C9(k) and ci〇(k). In particular, the first sinusoidal c 1 (k) can be synthesized into a sequence of samples in the time domain as follows: for k = 0 to 12287 ' cl(k) = cos(2TT * 237k) / 12288. In addition, the sample value in a 51 1 sample audio block combined with the first sine cl (k) can be calculated as follows: For m=〇 to and p=〇 to 46, 38 !342515 clp( m)=w(in)cos[27r *237k*(p*256 + m) /12288] where w(m) is a window function used in the above AC_3 compression standard. The modified discrete cosine transform of the clp(m) contains a number of modified T-cosine transform coefficient values (e.g., 2 5 6 real numbers). In this example, the modified discrete cosine transform coefficient values of the 5 1 2 sample frequency index 9, the work order, and the η may have large amplitudes because. The system is combined with the central frequency index of the 1 2 2 8 8 samples, and it should be 1 〇 at the center frequency of the 512 samples. The modified discrete cosine transform coefficient values of the f-sample frequency index are coincident with the 5 1 2 sample frequency index 9, and the discrete cosine transform coefficient value can be ignored. The modified discrete cosine transform (four) value is modified. This factor can be divided by the factor of ~(4)'· Q=512/4 = 128, where 5丄2 is the number of normalized samples as follows. Therefore, the time domain cosine wave system of the unit vibration structure where the i 2 2 8 j is combined with each block can be: 丨 24 〇 frequency index 1 - - unit cell amplitude solid sample center, as a Combining a central frequency with a frequency of approximately 9 3 6 = cosine transform coefficients, 1', the central frequency is 5 1 2 at the 13th. 'Reference of the discrete duty conversion coefficient (as shown for c, p(m), the code frequency is $ in combination with the frequency of the code "〇" f (for example, corresponding to the 12 sample central frequency index iQ relative rate Value) causes the 5 index 9 and ... words to have the highest final change; f 12 sample frequencies 'because 512 samples of the central frequency t discrete cosine transform amplitude frequency sheep index" corresponds to the 122 39 8 sample t center frequency索丨5丨2 4 0 is close to the (1) "sample center frequency"; = "so" is similar to the code frequency index 2 8 9 similarly, the pair has 512 sample frequencies in the same;;;; The ίί2 series can be modified to have a large discrete cosine transforming vibration 1 2 and i 3 orders. The drinking frequency index 2 8 9 can be: wide discrete cosine transform system frequency cable bow 1 2 and; ^ β - μ The result is that the 忒5 1 2 samples t-d has a modified central deviation from the central cosine transform of the 5 1 2 samples, because the central-style Γ 112 corresponds to the 12 2 8 8 central frequency index 2 8 8 Η 2 8 8 sample U 2 8 8 sample center frequency cable W89 is close to this, corresponding to the code frequency (4) 3 3 9 < ° in another In order to generate a modified discrete cosine having a large value in the frequency; 3 series; 5; the conversion coefficient. The modified frequency discrete index cosine to the central frequency index of the code frequency index Q ... Samples because of the modified discrete cosine transform amplitude in 5 1 2 samples > 8 sample center frequency index = index 14 corresponds to the ... 8 near the 1 2 2 8 8 sample too a and 6 hai code frequency index 3 3 9 The one-to-one _ index 3 3 6 ° of the connector "0" is based on the index chord wave combined with the code, and the unit amplitude of the mother of the real m can be modified for the range from 9 to ^ Discrete cosine transform coefficient system. The "2 sample frequency two (four) frequency index is calculated 1 3,1 5i 7 , } Some: such as 9,1 1, can be subject to two adjacent = rate = 2 3, 2 5 And ", is affected by the drought index, because none of these 5 i 2 40 1342515 8 sample frequency indexes directly correspond to the sample center frequency index (for example, 2 4 0, 2 8 8 , 3 3 β , 384 '432, 480 '528, 576, 624 or 6 7 2 ). Therefore, each of the compressed digital streams Among the 5 丄 2 sample audio blocks, the modified discrete cosine transform coefficient can be calculated as described below to represent the code signal. The compressed AC_3 data stream, for example, each铊3 frame contains 6 modified discrete cosine transform coefficients (example - 5 map of MAG, MAI, MA2, MA3, MA4, ma5 & ma6) 4": string conversion coefficient group ' and each modified discrete cosine transform Should be in a 512 sample audio block. As described above in conjunction with 5th and 6th: 逑,: a modified discrete cosine transform coefficient system table = (repair 2 'where χ " 'for the finger]. The false ... is a fake step size, and the value : Number product. The number of steps of the number is large... The index (4) Value 2 = Quantization step size... 'Refer to the table = Table 6 0 〇, like 丄a — mouth ~ must read, -4/i original leave The value is seven coffees (ie, 5) b". The integer step size is large. The integer value Nk is -4. Η is 1 / 1 5, and in order to inject - code signals into the compression (4) to 29 The false number system is determined. ^^ stream, for the modified discrete cosine transform coefficient & but the corresponding cosine transform coefficient amplitude C9, C丨. and c;: eight m丨丨 of the modified discrete. The 512 is difficult: two, respectively - °.3, " and 0.2 ^ 丄 d samples of the central frequency cable 1 of the code modified from 41 1342515 political cosine transform coefficient vibration Gi i is the minimum amplitude (for example, The absolute value of an absolute value of 〇. 2 is modified by the discrete cosine transform. The code is modified to change the value of L from the political cosine to the number k of the sniper sk used to obtain the repairs. The values of the cosine transform coefficients m9, ~ and the claws are respectively (-[5, 4., and 1 · 〇), because the code modified the discrete cosine transform coefficient amplitude hi has a small amplitude. In particular, the 'false integer i & adds i to normalize the modified discrete cosine transform coefficient and the value of m". Therefore, the integer values Ng and N! are modified as follows, because The step size of the alias is: sn: Ng is modified to N9+(_15*heart)/% and Ni0b is changed to N1G + (4. 0*Sn)/S1Q. Therefore, the modified integer integer value is modified. The N1Q and N!! systems can be used to generate corresponding modified modified discrete cosine transform coefficients _9,_1() and " respectively. As described in conjunction with Figures 5 and 6, the modified modified discrete The cosine transform coefficients wmg, wm" and wmn can replace the original modified discrete cosine transform coefficients, and other modified discrete cosine transform coefficients can be modified in a similar manner. In fact, for any modified discrete cosine Conversion factor and L' the biggest change is limited by its false integer value Limits and lower limits. Refer to Figure 6, for example, the upper and lower limits of the indications in Table 6 are -〇. 9333 and +〇. 9333. Therefore, the block is covered by the energy system. Used to determine the amplitude of the code, and the code character can be built into an audio signal by modifying 8 consecutive frames of the compressed digital tributary stream ' without the coding of the above method and device The modified decompression of the discrete cosine transform coefficients is implemented during the process. Figure 14 is a block diagram of an exemplary processor system 2 that can be used to implement the method 42 1342515 and apparatus disclosed herein. The processor system 20 can be a desktop computer, a laptop computer, a notebook computer, a personal digital assistant, a server ' _ internet device or any other form of computing Device. The processor system 2 〇 〇 示 shown in Fig. 14 includes a wafer group 2 0 1 〇 which includes a memory controller 2 〇 1 2 and a wheel input/output controller 2 〇 14 . Typically, memory and input/output are provided for general purpose and/or special purpose registers, processors 2 0 2 存取 are accessed or implemented using one or more processors. The Yuli Technology Department can be used to implement the process. The system includes a cacher 2 〇2 first layer integrated cacher (L1), an L2) and a second layer integrated cacher ( Structure to store data. Management functions, timers, etc. This process replaces the embodiment of the 2 2 2 〇 one chip set and a plurality of them, which can be used in the other 2 0 2 0.

2 'The system can use a second layer to integrate L3) and or the cacher (other suitable as the traditional technology, the memory controller 2 〇χ 2 system actually rotate the processor 2 Q 2 ◦ can pass through - One bus 2 ◦ 4 ◦ 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 一个 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性 挥发性The 3 2 system can be implemented by synchronous dynamic random access memory, ram coffee moving away from memory, or any other form of dynamic random access memory. The non-volatile memory 2 Q 3 4 system can be used. Flash memory, read-only memory 'electronic erasable, program-only memory and /

43 1342515 Any other form of desired memory component. ° Xuan treatment system 2 〇 〇 〇 also contains a connection with the bus 2 〇 4 0 interface j. v * 屯 2 2 〇 5 〇.泫Interface circuitry 2〇5〇 can use any form of well-known interface standard and real-time; such as the Ethernet interface' - a universal serial bus with a third-generation input/output interface and/or any other suitable Interface form. Or the input device 2 0 6 0 is connected to the interface circuit 2 ) — ) The device 2 〇 6 0 allows a user to input data and commands into the processor 2 〇 2 0 . For example, the input device 26 0 can be a keyboard, a mouse, a touch sensor, a track pad, a trackball, an indicator device and/or a voice recognition system. 2 0 7 〇 is also connected to the interface circuit 2 Ή ° For example, 'the output device 2 0 7 0 can be seen in the media table (such as a light-emitting display, a liquid crystal display, a cathode ray g display Benefits, _ a printer and / or leftover implementation. Therefore 'the interface circuit ◦ 50 typically most importantly contains a graphics driver card. / Processing & System 2 Q system also contains - or more A large number of storage devices are used to store software and data. The examples of such a large number of storage devices are floppy disk and floppy disk drive, hard disk drive, (3) and its a player and digital video player. Ground & 曰 first disc and digital video disc player. The interface circuit 2 〇5 〇 also data machine or a network interface card contains a communication device, such as a, to facilitate exchange with a 44 U42515 computer through a network data The processor system 2 Q "and the communication link can be any form of network connection, such as a beta network connection 'a digital subscriber loop' ~ a telephone line, a 乂太话 system, a Coaxial cable, etc. τ is in a conventional manner, picking up the wheeling device 2〇6〇, out device 2〇7〇, the mass storage device 2 0 80 and/or the network: typically Controlled by the input/output controller 2, the input/wheeling controller 2 is configured to pass the processor 2 through the bus 2040 and the interface circuit 2〇5〇. And into the device 2 0 6 0, the output device 2 ◦ 7 〇, the mass storage heart 2 0 80 and/or the function of the network communication. Although shown in Figure 1 $ #I# π - t The system is shown as a separate block within the processor system 2 〇0 0, and the functions implemented by some of the blocks may be integrated into a single semiconductor circuit or two or more may be used. Implemented by a separate integrated circuit. For example, although the memory controller 2 0 1 9 BM k , , u丄2忒The input/output controller 2〇丄4 h is not a separate block in the wafer phase, 2·1 〇1 ,, the memory controller 2〇12 and #在>/4^^^ The imaginary input/output controller 2 0 1 4 can be integrated into a single-semiconductor circuit. The methods and apparatus disclosed herein are particularly suitable for use with data streams implemented according to the C 3 privacy. Using 'however, the methods and apparatus disclosed herein may be applied to other digital audio coding techniques. This disclosure is made for an exemplary television system that 'should be understood' is a system that can be easily disclosed Applied to many 45 4* 4*I342515 other media systems. Thus, although this disclosure is illustrative of exemplary systems and sequences, the disclosed examples are not the only way to implement such a system. Although certain exemplary manufacturing methods, apparatus, and articles have been described herein, the scope of the patent is not limited thereto. Conversely, this patent covers all manufacturing methods, equipment, and articles that are within the scope of the patent, whether within the meaning of the word or when the law of equality falls behind. For example, although this disclosure is illustrative of the most important components including an exemplary system of software executed on a hardware, it should be noted that such systems are merely exemplary and should not be considered as limiting. . In particular, it is believed that any or all of the hardware and soft systems disclosed herein can be used in a separate dedicated hardware, as a separate dedicated blade, as a separate dedicated software or as a hardware, body and/or software. One group is actually embodied. [Simple description of the diagram] (1) The first part of the drawing is an exemplary medium. The block diagram of the 1 琛 k 糸 表示 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第Block diagram 〒 Figure 3 is a block diagram representation of an uncompressed digital data stream combined with the second watermark built-in system; ', the fourth image of f life can be used for the real steam bar Z The block diagram of an exemplary built-in device of the non-compressed built-in system of Fig. ♦ shows that the water 5th figure shows the demonstration of the built-in device combined with the 'Figure' 46 1342515 Sexual compressed digital data stream; Figure 6 shows an exemplary quantitative look-up table that can be used for real-time d-draw watermarking built-in systems; Figure 7 shows one that can be used 2^~ Do not put the watermark in the continuation · Another data stream that is compressed and then processed. · The uncompressed digits of the woman's uncompressed digits. The exemplary compressed digital data stream of the civil/constant digital stream. ;

Figure 9 shows the 图 笳 第 笳 | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印 水印The modified program of the figure can be implemented in the form; the 11th figure shows a form in which a data frame can be processed; the 12th figure shows that a watermark can be built into a compressed digital data stream. One form,

Figure 13 shows an exemplary code frequency index table that can be used to implement the exemplary watermark built-in system of Figure 2; Figure 14 is a graph that can be used to implement Figure 2. An exemplary watermark built-in system - a block diagram representation of an exemplary processor system. (ii) Component Representation Symbol 10 0 Broadcasting System 110 Service Provider 112 Cable TV Service Provider 47 1342515 114 Radio Frequency Television Service Provider 116 Satellite TV Service Provider 12 0 Television 12 5 Remote Control Device 13 0 Receiving Device 14 0 Measuring Device 15 0 viewing area

160 User, Home Member 17 0 Network 180 Data Collection Device 2 0 0 Watermark Built-in System 2 10 Built-in Device 220 Watermark Source 2 3 0 Watermark Information 2 4 0 Compressed Digital Data Stream

2 5 0 Compressed digital data stream with watermark 3 0 0 Uncompressed digital data stream 3 10 Time domain audio block 3 2 0 Modified discrete cosine transform coefficient group 4 10 Identification unit 420 Decompression unit 4 3 0 Modification Unit 440 Recompression Unit 5 10 Box 48 1342515 5 2 0 Modified Discrete Cosine Transform Coefficient Group 5 3 0 Time Domain Audio Block 5 4 0 Time Domain Audio Block 5 5 0 Time Domain Audio Block with Watermark 5 6 0 Modified discrete cosine transform coefficient group with watermark 6 0 0 Quantization look-up table 7 0 0 Uncompressed digital data stream 7 10 Audio block group 7 2 0 Channel 7 3 0 Channel 8 0 0 Compressed digital data stream 8 10 Central channel 8 2 0 modified discrete cosine transform coefficient group 8 3 0 inverse transform time domain audio block 8 4 0 reconstructed time domain audio block 8 5 0 time domain audio block with watermark 8 6 0 with watermark modification Over Discrete Cosine Transform Coefficient Set 2 0 0 0 Processor System 2 0 10 Chipset 2 0 12 Memory Controller 2 0 14 Input/Output Controller 2 0 2 0 Processor 2 0 2 2 Cache 2 0 3 0 Main memory 49 1342515 2032 2080 mass storage device 2034 made of non-volatile memory device 2070 memory 2060 input output device interface circuit 2040 bus 2050

50

Claims (1)

1342515 - Month 4曰Revised replacement page picking, patent application scope: 1 method of built-in watermark' includes: identifying one or more frames combined with a compressed digital data stream; decompressing the one or more frames Each of which identifies a plurality of sets of conversion coefficients: and modifying the plurality of sets of conversion coefficients to construct the watermark by: determining a watermarked conversion combining one of the plurality of converted coefficient sets having a watermark a false number of the coefficient; and replacing the false number of a corresponding one of the plurality of conversion coefficient groups with a false number combined with the watermark conversion coefficient. 2. The method of claim 2, wherein the modifying the plurality of conversion coefficient groups comprises: α-a watermark conversion coefficient group replacing at least one of the plurality of conversion coefficient groups. 3. The method of claiming patent scope, wherein the combination is determined A pseudo-digital system having a watermark conversion coefficient of a plurality of watermark conversion coefficient groups includes: selecting a code signal frequency to be the conversion coefficient according to the data to be built in; the code signal of the conversion coefficient The frequency determines a masking energy that will be encoded as a rate; the amplitude of the number is selected according to the mask energy; and the conversion for the watermark determines the combination of the watermarked conversion coefficients according to the amplitude False 51 1342515 妒 妒 巧 修正 修正 替换 修正 4 4 4 4 4 4 4 4 4 4 4 4 4 4 β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β The method of claim 3, wherein the code signal frequency comprises one or more positive chords, and wherein each sinusoidal component has a frequency according to a desired code. The method of claim 1, wherein modifying the plurality of conversion coefficient groups comprises: generating a plurality of times according to the plurality of conversion coefficient groups An audio block; generating a plurality of reconstructed audio blocks according to the plurality of time domain audio blocks, and generating a plurality of watermarked audio blocks according to the plurality of reconstructed audio blocks. The method of claim 6, wherein generating the plurality of time domain audio blocks comprises generating a first time domain audio block and a second audio time domain block in combination with the original audio block. The method of claim 6, wherein the generating the plurality of reconstructed audio blocks according to the plurality of time domain audio blocks comprises: according to a first time domain audio block and a second audio time domain block. Generating a reconstructed time domain audio block corresponding to an original audio block. 9. The method of claim 8, wherein the first time domain audio block and the second audio time domain block are And generating the first reconstructed time domain audio block corresponding to the first audio block includes adding the first and second time domain audio blocks. 52 1342515 Playing/correcting replacement page 1 0 · Application The method of claim 6, wherein generating the plurality of watermarked audio blocks based on the plurality of reconstructed audio blocks comprises: generating a modifiable time domain based on the plurality of reconstructed audio blocks An audio block; and generating a first watermarked audio block and a second watermarked audio block according to the modifiable time domain audio block and the watermark. 1 1 *as claimed in claim 1 The method of claim, wherein generating the modifiable time domain audio block according to the plurality of reconstructed audio blocks comprises concatenating a first reconstructed audio block and a second reconstructed audio block To form an audio block of 5 i 2 samples. 12. The method of claim 1, wherein the plurality of conversion systems are modified, the group includes modifying a plurality of conversion coefficient groups according to i watermarks Conversion factor group. The method of claim 1, wherein the modifying the plurality of conversion coefficient groups comprises generating the plurality of audio blocks according to the first watermark and the second watermarked audio block. Converted coefficient group with watermark And a pseudo number corresponding to each of the second watermarked audio blocks in the compression coefficient group of the compressed watermark of the reduced digital data stream. The changed coefficient group includes determining that each of the plurality of combinations has a combination according to the pressure information 53 1342515 Office February 1 Revision Replacement Page 1 5 · If the application for the first quarter of the patent scope has been fixed, the method of each of the multiple conversion coefficient groups is r ^ 3 or more modified Separated cosine transform coefficients. 16 If the scope of the patent is ί j§ 夕*·, t *}., to the method of the 丄 丄 item, where the compressed digital data stream is based on the sound m 1 solid 9 frequency standard compression. 1 7 'If Shen Qing's patent scope is 1st 苜 苜 + +4., 礼 乐 丄 之 ' 其中 其中 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识 辨识The framework of the rebel poor camera includes an audio block that identifies at least one of the plurality of audio channels. One Shishen. The method of item 1 of the patent scope of the month of 'decompressing the:: each of the plurality of blocks to identify a plurality of sets of conversion coefficients includes identification, "the compressed information of the compressed digital data stream. 八 九 . The method of claiming the scope of the patent, wherein the method further recompresses the one or the community according to the plurality of watermark conversion coefficient groups. The method of applying for a patent, wherein the watermark system Ό 5 - one media source and one media program. 2 1 . A device for built-in watermarks, comprising: - an identifier for identifying one or more combined digital data a box of flows; a 'decompressor' for decompressing each of the one or more frames to identify a plurality of sets of conversion coefficients; and Μ L L L, which is used for Determining a combination of a plurality of watermarked conversion coefficients having one of the watermark coefficients, and combining the imaginary, /^7卩 conversion coefficient of the false number to replace a combined complex number of 54 1342515 _ years "Monthly 1曰Replacement a false number of a corresponding conversion coefficient of one of the conversion coefficient groups of the page, to modify the plurality of conversion coefficient groups to construct the watermark. 2 2. The device of claim 21, wherein the modifier system Selecting a code signal frequency to encode into a plurality of converted coefficient groups according to the data to be built; determining a masking energy that is to be encoded into a plurality of converted coefficient groups of the code signal frequency; Masking the energy and selecting an amplitude for the watermarked conversion coefficient; and determining the false number of the watermarked conversion coefficient according to the amplitude. 2 3. The device of claim 2, wherein the code The signal frequency includes a frequency corresponding to a plurality of high-resolution frequency domain representations. 2 4. The device of claim 2, wherein the code frequency includes one or more sinusoidal components. And wherein each of the sinusoidal lines has a frequency according to a desired code. 2 5. As in the device of claim 2, the modification is made to ^ A plurality of reconstructed audio blocks are generated to generate a plurality of time domain audio blocks, a plurality of reconstructed audio blocks are generated, and a plurality of audio blocks having a watermark of φ φ are generated. 〃 2 6 . The device of item 5, wherein the modifier generates a first time domain audio zone and a second time domain audio zone combined with an original audio block. 2, 7. The device of claim 25, The modifier generates a reconstructed time domain audio corresponding to an original audio block according to the -first-time domain audio block and - (4) two-audio time domain = 55 1342515. Day correction replacement page _ ____ 1 block. 2 8. The device of claim 27, wherein the modifier is added to the first and second time domain audio blocks. 2 9. The device of claim 25, wherein the modifier generates a modifiable time domain audio block based on the plurality of reconstructed audio blocks; and based on the modifiable time domain audio The block and the watermark generate a first watermarked audio block and a second watermarked audio block. 30. The device of claim 29, wherein the modification _ is connected in series with a first reconstructed audio block and a second reconstructed audio block to form a 5.2 sample. Audio block. The device of claim 3, wherein the modifier determines, according to the compressed information of the compressed digital data stream, each of the plurality of watermarking conversion coefficient groups each having a watermark. A false number of the coefficient of the watermark. The apparatus of claim 21, wherein the modifier modifies the plurality of sets of call conversion coefficients based on the plurality of sets of conversion coefficients having a watermark. 3 3. The device of claim 3, wherein the modifier generates a plurality of watermarked conversion coefficients according to a first watermarked audio block and a second watermarked audio block. One of the groups. 3 4. The apparatus of claim 3, wherein the modified person replaces one of the plurality of conversion coefficient groups with one of a plurality of watermark conversion coefficient groups. 56 1342515 Π 蜱 J!.. 分 3 5 The equipment of item 2 of the patent scope ‘where a plurality of conversion factors··· and #自系包含- or a plurality of modified separated conversion coefficients. The apparatus of claim 21, wherein the compressed digital lean stream is compressed according to an audio compression standard. 3 7. The device of claim 2, wherein the identification unit identifies an audio block that combines a plurality of audio channels. B one. 3 8. The apparatus of claim 2, wherein the decompressing soap element identifies compression information that is combined with the compressed digital data stream. 3 9. The device of claim 2, wherein the watermark comprises a watermark combined with one of a media source and a media program. 4 0. The device of claim 2, wherein the step further comprises a frame recompressor to recompress the one or more blocks based on the plurality of watermarked coefficient groups. 4 1 · A machine-accessible medium having a plurality of instructions that, when executed, cause a machine to: identify one or more frames that combine a compressed digital data stream; decompress the one or Each of the plurality of blocks to identify a plurality of sets of conversion coefficients; and modifying the plurality of sets of conversion coefficients to construct the watermark by: determining a watermark combining one of the plurality of sets of conversion coefficients having a watermark a false number of the conversion coefficient; and a false number combined with the watermark conversion coefficient is substituted for a corresponding conversion coefficient of one of the plurality of conversion coefficient groups. 57 1342515 Digital Shenqing Patent Range Item 41 Accessible media, when the ==! command is executed, the instructions cause the machine to modify the plurality of reloaded rents 5 by the following steps to have a watermark conversion coefficient, and then replace one of the multiple conversions At least one of the coefficient groups.苴中土: This machine can access the media in accordance with item 41 of the patent scope, in which “when some instructions are executed, the instructions are guided by the steps to determine the conclusion of the 葙 彻 彻 彻 俄 俄 俄 俄... " a false number with a watermark conversion coefficient of one of a plurality of watermark conversion coefficient groups: Select: Erma signal frequency 'to be a plurality of conversion coefficient groups according to the resources to be built; 1 Decide - combine a masking energy to be encoded into a plurality of transcoding frequencies; < multiple codes are selected according to the masking energy for the amplitude of the watermarked number; and π &# determines the combination according to the amplitude The conversion coefficient code of the watermark 44. The machine-accessible medium of claim 43 of the patent application, the handle § 〖 〖 (m) contains a frequency corresponding to one of the high-resolution representations of Wei. or 45. The machine-accessible medium of claim 43, wherein the code signal frequency comprises one or more sinusoidal components, and each of the sinusoidal components has a frequency according to a desired code. Please access the media on page 41 of the patent scope ' 58 ^ 342515 When the instructions are executed, the instructions cause the machine to modify the plurality of conversion coefficient groups by the following dreams: generating a plurality of time domain audio blocks according to the plurality of conversion coefficient groups, according to the A plurality of time domain audio blocks are generated to generate a plurality of reconstructed sound blocks, and a plurality of watermarked audio blocks are generated according to the plurality of reconstructed audio blocks. 4 7. The machine-accessible medium of claim 46, wherein when the instructions are executed, the instructions cause the machine to generate a plurality of time domain audio blocks by the following steps: generating a combination— A first time domain audio block and a second audio time domain block of the original audio block. 4 8 - The machine-accessible medium of claim 46, wherein when the instructions are executed, the instructions cause the machine to be based on the plurality of time domain audio blocks by the following steps A first time domain audio block and a second audio time domain block generate a reconstructed time domain audio block corresponding to the original audio block. 4 9 - The machine-accessible medium of claim 4, wherein when the instructions are executed, the instructions cause the machine to be generated by adding the first and second time domain audio blocks Corresponding to the reconstructed time domain audio block of the original audio block. 5. The machine-accessible medium of claim 46, wherein when the instructions are executed, the instructions cause the machine to generate the plurality of watermarked audio blocks by the following steps: Generating a modified video block according to the plurality of reconstructed audio blocks, and modifying the page time domain audio block; and generating the time domain audio block according to the edge and the watermark - The first - the watermarked audio block and the second watermarked audio block. 5 1. If the machine-accessible media of the scope of claim 5, /, g, and the second finger are executed, the instructions cause the machine to generate the modifiable time domain by the following steps: Audio block: A block of reconstructed audio blocks and a second reconstructed audio block are connected in series to form an audio block of 512 samples. b 2. The machine-accessible medium of claim 46, wherein when the instructions are executed, the instructions cause the machine to generate a plurality of sets of conversion coefficients having a watermark by the following steps: One of a plurality of sets of converted coefficients having a watermark is generated based on the two first watermarked audio blocks and one second watermarked audio block. 5 3. If the machine-accessible medium of claim 52 is applied, the instructions cause the machine to set a number of watermarked coefficient groups by: The root of the compressed digital data stream invites, ° 3 m "" 1 钿 Λ Λ 疋 个 个 个 个 个 个 个 个 个 个 个 结合 结合 结合 结合 结合 结合 结合 结合 结合 结合 结合 结合 。 。 。 。 。 。 。 。 。 。 。 。 。 Wm relies on the machine-accessible medium of the patent pillow of the 'I" patent, in which each of the plurality of conversion coefficient groups has a separate cosine conversion coefficient. 'Xi Yi tampering 5 5 ' Please refer to the machine of the scope of the 4th item of the patent, in which the compressed digital data is chopped by the '+ machine according to an audio compression standard and 60!342515) fT year 13⁄4 month bucket correction replacement page j is compressed. 5 6 . Applying for a patent, Fan Ganbei, the machine can be accessed and accessed. 8* Among them, when the instructions are taken by the stick, the media is identified, and the following steps are used to identify the one or more-machine error box: identification combined with plural Audio 枓 枓 k, ^ Λ 至少 频道 频道 频道 频道 频道 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 Each of the plurality of boxes decompresses the - or 裣, "* 4 Wei conversion coefficient groups: (4) the compression information of the digital data stream combined with the pressure & past 5 8 . Article 41 of the scope of patents, when these instructions are executed, The sub-media, and ... the "causes the machine to recompress the - or multiple cabinets based on the plurality of watermarked conversion system' arrays. 59. If the machine-accessible medium of the "Scope of Patent Application" is applied, the t-distal watermark is combined with one media source and one media format. 6 ◦ • The machine can be saved as claimed in item 4 of the patent application scope. Take the media, the system includes a programmable closed array, a special application integrated circuit: an erasable programmable read-only memory, a read-only memory, a random access memory, a wall media And one of the optical media. Pick up, schema: as the next page 61