Vector quantization of a time sequential signal by quantizing an error between subframe and interpolated feature vectors
First Claim
1. A vector quantizing method of vector quantizing a time sequential signal divided into consecutive frames, each comprising a plurality of subframes identified by subframe indexes, comprising the steps of:
- extracting first and second feature vectors at least from first and second subframes of a current frame of said time sequential signal;
calculating a quantized frame vector for said current frame by using said first feature vector;
interpolating an interpolated feature vector for said current frame based on interpolation coefficient vectors between said quantized frame vector and another quantized frame vector calculated for a previous frame which is at least one frame prior to said current frame;
calculating an error between said second feature vector and said interpolated feature vector as an error feature vector;
quantizing said error feature vector into a quantized error vector; and
producing quantized subframe vectors for said plurality of subframes by using said subframe indexes, the quantized frame vector for said current frame, said interpolated feature vector, and said quantized error vector.
1 Assignment
0 Petitions
Accused Products
Abstract
For quantization of subframe feature vectors representative of subframes identified by subframe indexes in each frame of a time sequential signal, such as an audio or a video signal, a first feature vector of a first subframe of a current frame is extracted with additional extraction of a second feature vector of a second subframe of the current frame and is quantized by a frame vector quantizer unit into a quantized frame vector for the current frame. Using a quantized frame vector for a previous frame and referring to interpolation coefficient vectors, an interpolator interpolates an interpolated feature vector for the current frame. An error vector is calculated between the interpolated and the second feature vectors and is quantized by an error vector quantizer unit into a quantized error vector. The subframe feature vectors are quantized by using the subframe indexes, the quantized frame vectors, the interpolated feature vector, and the quantized error vector. Preferably, the interpolation coefficient vectors are adjusted according to a mode decided for the current frame by its subframes.
38 Citations
20 Claims
-
1. A vector quantizing method of vector quantizing a time sequential signal divided into consecutive frames, each comprising a plurality of subframes identified by subframe indexes, comprising the steps of:
-
extracting first and second feature vectors at least from first and second subframes of a current frame of said time sequential signal; calculating a quantized frame vector for said current frame by using said first feature vector; interpolating an interpolated feature vector for said current frame based on interpolation coefficient vectors between said quantized frame vector and another quantized frame vector calculated for a previous frame which is at least one frame prior to said current frame; calculating an error between said second feature vector and said interpolated feature vector as an error feature vector; quantizing said error feature vector into a quantized error vector; and producing quantized subframe vectors for said plurality of subframes by using said subframe indexes, the quantized frame vector for said current frame, said interpolated feature vector, and said quantized error vector. - View Dependent Claims (2, 3)
-
-
4. A vector quantizing device for vector quantizing a time sequential signal divided into consecutive frames, each comprising a plurality of subframes identified by subframe indexes, said vector quantizing device comprising:
-
extracting means for extracting first and second feature vectors at least from first and second subframes of a current frame of said time sequential signal; frame vector calculating means for calculating a quantized frame vector for said current frame by using said first feature vector; interpolating means for interpolating an interpolated feature vector for said current frame based on interpolation coefficient vectors between said quantized frame vector and another quantized frame vector calculated for a previous frame which is at least one frame prior to said current frame; error calculating means for calculating an error between said second feature vector and said interpolated feature vector as an error feature vector; error quantizing means for quantizing said error feature vector into a quantized error vector; and quantized subframe vector calculating means for calculating quantized subframe vectors for said plurality of subframes by using said subframe indexes, the quantized frame vector for said current frame, said interpolated feature vector, and said quantized error vector. - View Dependent Claims (5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
Specification