Coded data generation or conversion – Digital code to digital code converters
Reexamination Certificate
2000-01-31
2001-08-21
JeanPierre, Peguy (Department: 2819)
Coded data generation or conversion
Digital code to digital code converters
C341S200000
Reexamination Certificate
active
06278385
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to a vector quantizer and a vector quantization method, and more particularly to a vector quantizer and a vector quantization method that permit watermark information to be embedded in a coded bit stream in encoding a digital signal such as an audio signal or an image signal.
2. Prior Art
A vector quantizer which permits watermark information to be embedded in a digital code stream consisting of vector quantized indexes in accordance with embedding position designating information, for example, in a designated bit of a 5-bit index, has been proposed (Japanese Laid-Open Patent Publication No. 10-224342). The proposed vector quantizer is constructed such that it selects a quantization vector that has the least distortion with respect to an input vector from among indexes fixed in this manner, and outputs the index representing the selected quantization vector.
The proposed vector quantizer, however, has a drawback that since the embedding position is fixed, the position where the watermark information has been embedded can be identified using difference data obtained by comparing data that is obtained when encoding is performed with the watermark information embedded, with data that is obtained when encoding is performed with no watermark information embedded, for the same input signal string of finite length.
The proposed vector quantizer has another drawback that since the embedding position is determined without taking into consideration the effect on the quality of the reproduced signal, the influence of the embedded watermark information is exerted upon the entire reproduced signal, so that depending on the embedding position of the watermark information, a user may readily recognize degradation of the reproduced signal due to the addition of watermark information.
SUMMARY OF THE INVENTION
It is a first object of the present invention to provide a vector quantizer and a vector quantization method which do not allow the embedding position of the watermark information to be identified easily, and thus enable unfair use of the vector quantized information to be accurately found out.
It is a second object of the present invention to provide a vector quantizer and a vector quantization method which do not allow a user to readily recognize degradation of the reproduced signal due to addition of watermark information.
To attain the first object, according to a first aspect of the present invention, there is provided a vector quantizer for performing vector quantization on input signal data, comprising a waveform extracting device that performs waveform extraction on the input signal data to output extracted waveform data, and an index generating device that generates index data based on the output extracted waveform data, wherein as index data to be generated based on predetermined extracted waveform data in which no watermark information is to be embedded, out of the output extracted waveform data, the index generating device randomly selects index data corresponding to the predetermined extracted waveform data from a plurality of index data which, when decoded, have amounts of errors falling within a predetermined range with respect to the predetermined extracted waveform data before decoding.
To attain the first object, according to a second aspect of the present invention, there is provided a vector quantizer for performing conjugate structure vector quantization on input signal data, comprising a waveform extracting device that performs waveform extraction on the input signal data to output extracted waveform data, and an index generating device that generates index data based on the output extracted waveform data, wherein as a combination of index data to be generated based on predetermined extracted waveform data in which no watermark information is to be embedded, out of the output extracted waveform data, the index generating device randomly selects a combination of index data corresponding to the predetermined extracted waveform data from a plurality of combinations of index data which, when decoded, have amounts of errors falling within a predetermined range with respect to the predetermined extracted waveform data before decoding.
Preferably, the vector quantizer according to the first or second aspect further comprises a dummy data adding device that adds dummy data to a portion of the extracted waveform data when generating the extracted waveform data.
Preferably, the vector quantizer according to the first aspect comprises a first vector quantizing device to an n-th vector quantizing device (n is an integer not less than 2) for performing multistage vector quantization.
More preferably, at least one of the first vector quantizing device to the n-th vector quantizing device includes a dummy data adding device for adding dummy data to a portion of the extracted waveform data when generating the extracted waveform data.
Also preferably, the extracted waveform data is formed of a predetermined finite number of sample data, and wherein the dummy data adding device divides the input signal data into a plurality of divided data each formed of a predetermined number of sample data smaller than the predetermined finite number of sample data, the dummy data adding device adding the dummy data to each of the plurality of divided data to thereby generate the extracted waveform data formed of the predetermined finite number of sample data.
More preferably, the number of the sample data of the dummy data is different between when the divided data has watermark information embedded therein and when the divided data has no watermark information embedded therein.
The dummy data adding device repeatedly adds the dummy data formed of a predetermined fixed number of sample data to the divided data until the divided data becomes formed of sample data equal in number to the predetermined finite number.
The dummy data adding device may add, as the dummy data, calculated data obtained by performing a predetermined operation based on the divided data and formed of a number of sample data being equal to a difference between a number of sample data of the extracted waveform data and a number of sample data of the divided data.
The calculated data may be calculated by extrapolation operation, or it may be calculated by interpolation operation.
To attain the first object, according to a third aspect of the present invention, there is provided a vector quantization method for performing vector quantization on input signal data, comprising the steps of performing waveform extraction on the input signal data to output extracted waveform data, and generating index data based on the output extracted waveform data, wherein as index data to be generated based on predetermined extracted waveform data in which no watermark information is to be embedded, out of the output extracted waveform data, the index generating step randomly selects index data corresponding to the predetermined extracted waveform data from a plurality of index data which, when decoded, have amounts of errors falling within a predetermined range with respect to the predetermined extracted waveform data before decoding.
To attain the first object, according to a fourth aspect of the present invention, there is provided a vector quantization method for performing conjugate structure vector quantization on input signal data, comprising, comprising the steps of performing waveform extraction on the input signal data to output extracted waveform data, and generating index data based on the output extracted waveform data, wherein as a combination of index data to be generated based on predetermined extracted waveform data in which no watermark information is to be embedded, out of the output extracted waveform data, the index generating step randomly selects a combination of index data corresponding to the predetermined extracted waveform data from a plurality of combinations of index data which, when decoded, have amounts of errors falling within a predetermine
Kobayashi Eiko
Kondo Kazunobu
Yamauchi Ken'ichi
Jean-Pierre Peguy
Jeanglaude Jean Bruner
Pillsbury & Winthrop LLP
Yamaha Corporation
LandOfFree
Vector quantizer and vector quantization method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Vector quantizer and vector quantization method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Vector quantizer and vector quantization method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2511582