Methods of selecting an encoding mode
View Patent ↗Information needed to generate an encoded version of a matrix of j transform coefficients c0 . . . cj−1 representing a block of digital video data is determined by forming a numerical signature σ of the matrix and looking up a value for the information needed in an array employing the signature σ as an index to the array.
1. A method, comprising:
a) forming, by a computing machine that performs compression coding of video signals, a numerical signature σ of a matrix of j transform coefficients c 0 . . . c j−1 representing a block of digital video data; and
b) looking up, by the computing machine that performs compression coding of video signals, a number of bits R equivalent to a length of an encoded version of the matrix in an array employing the numerical signature σ as an index to the array.
2. The method according to claim 1 , comprising calculating the numerical signature σ, by the computing machine that performs compression coding of video signals, using a group of the coefficients in accordance with a procedure that returns a unique value of the numerical signature a for each unique set of coefficient values.
3. The method according to claim 1 , wherein a) comprises:
a1) linearly ordering the coefficients c 0 , c 1 , . . . , c j−1
a2) selecting a group of m sequentially ordered coefficients cu through c v where u is less than v and v equals u+m−1,
a3) mapping each coefficient c i in the group to an unsigned integer representation r i ,
a4) weighting each integer representation r i to form a weighted representation w i , and
a5) combining the weighted representations w u . . . w v to form the numerical signature σ.
4. The method according to claim 3 , wherein a1) comprises ordering the coefficients according to a zig-zag scan.
5. The method according to claim 3 , wherein a1) comprises ordering the coefficients according to a left-to-right-top-to-bottom scan.
6. The method according to claim 3 , wherein each coefficient c i in the group is limited to a predefined range of possible values and the predefined range of possible values for coefficient c i is greater than the predefined range of possible values for coefficient c i+1 .
7. The method according to claim 3 , wherein coefficients c v+i . . . c j−1 equal zero.
8. The method according to claim 3 , wherein c i is represented by a maximum number of bits b i and a3) comprises:
calculating r i for each coefficient c i in the group according to the equation:
r i =c i +2 bi−1 −1.
9. A method according to claim 3 , wherein a4) comprises:
calculating w i for each coefficient c i in the group according to the equation:
w k =2 tk r k ,
where t k =0 for k=u and
t
k
=
∑
i
=
0
k
-
1
b
i
for
k
>
u
.
10. The method according to claim 3 , wherein a5) comprises summing the weighted representations w u . . . w v to calculate numerical signature σ.
11. The method according to claim 3 , wherein a) further comprises:
receiving an identifier associated with the matrix, and, after a5),
storing the numerical signature σ in a holding table using the matrix's identifier as an index.
12. The method according to claim 11 , wherein the array comprises a plurality of data tables and the method further comprises, between a) and b):
receiving a communication from a requestor indicating a data table, and
receiving the identifier from the requestor, b) further comprises:
using the identifier as an index to the holding table to retrieve the numerical signature σ, and
reading a value of the information needed from the indicated data table, and, after b), the method further comprises:
returning the value of the information needed to the requestor.
13. The method according to claim 1 , wherein the array comprises multiple data tables and b) comprises:
receiving a communication from a requestor indicating a data table,
reading the number of bits R from the indicated table, and
returning the number of bits R to the requestor.
14. The method according to claim 1 , wherein, prior to b), the method comprises calculating the numerical signature σ and the number of bits R for every matrix in which the value of each coefficient falls within a predetermined range.
15. The method according to claim 14 , wherein prior to b), the method comprises creating the array, the array contains the number of bits R indexed by the calculated numerical signature σ values.
16. The method according to claim 1 , comprising performing a) in parallel with calculating transform coefficients c 0 . . . c j−1 .
17. A method, comprising:
a) forming, by a computing machine that performs compression coding of video signals, a numerical signature σof a matrix of j transform coefficients c 0 . . . c j−1 representing a block of digital video data; and
b) looking up, by the computing machine that performs compression coding of video signals, a sequence of bits S equivalent to an encoded version of the matrix of transform coefficients in an array employing the numerical signature σ as an index to the array.
18. A method, comprising:
a) forming, by a computing machine that performs compression coding of video signals, a numerical signature σ of a matrix of transform coefficients c 0 . . . c j− 1 representing a block of digital video data;
b) looking up, by the computing machine that performs compression coding of video signals, at least one value for a parameter R in a coding mode array using the numerical signature σ as an index to the array, the at least one value of the parameter R being equal to a size, in bits, of an encoded version of the matrix encoded with one of a plurality of coding modes;
c) using, by the computing machine, the at least one value of parameter R as an input to a mode selection function J to determine a corresponding value of J;
d) selecting, by the computing machine, a coding mode from the plurality of coding modes corresponding to the smallest value of J; and
e) selecting, by the computing machine, a sequence of bits S equivalent to an encoded version of the matrix of coefficients from a bit stream array, the sequences of bits S being uniquely identified by the numerical signature σand the selected coding mode.
19. The method according to claim 18 , comprising calculating the numerical signature σ, by the computing machine that performs compression coding of video signals, using a group of the coefficients in accordance with a procedure that returns a unique value of the numerical signature σ for each unique set of coefficient values and wherein a) comprises:
a1) linearly ordering the coefficients c 0 , c 1 , . . . , c j−1 ,
a2) selecting a group of m sequentially ordered coefficients c u through c v where u is less than v and v equals u+m−1,
a3) mapping each coefficient c i in the group to an unsigned integer representation r i ,
a4) weighting each integer representation r i to form a weighted representation w i , and
a5) combining the weighted representations w u . . . w v to form the numerical signature σ.
20. The method according to claim 18 , wherein prior to a), the method comprises calculating the numerical signature σ value and a value of the parameter R and the sequence of bits S for each of the plurality of coding modes for every matrix in which the value of each coefficient falls within a predetermined range.
21. The method according to claim 20 , wherein the coding mode array comprises multiple coding mode tables and the method comprises, prior to b), creating said coding mode tables wherein each coding mode table corresponds to a given coding mode and stores the calculated R value for each value of the numerical signature σ for that coding mode.
22. The method according to claim 21 , wherein for every unique combination of the numerical signature σ and coding mode there is a single S value and the method comprises, prior to a), storing each calculated S value in a bit stream table.
23. The method according to claim 18 , comprising performing a) in parallel with calculating the matrix of transform coefficients c 0 . . . c j−1 .