Vector joint encoding/decoding method and vector joint encoder/decoder

• US 8,930,200 B2
• Filed: 07/24/2013
• Issued: 01/06/2015
• Est. Priority Date: 01/26/2011
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: Ma, Fuwei, Zhang, Dejun, Miao, Lei, Qi, Fengyan
• Primary Examiner(s): GUERRA-ERAZO, EDGAR X
• Application Number:
  US13/950,018
• Publication Number:
  US 20130317810A1
• Time in Patent Office: 531 Days
• Field of Search: 704500-504, 704/216, 704/217, 704/219, 704/223, 704/225, 704/229, 704/258, 704/262, 704/263, 704/264
• US Class Current: 704/500
• CPC Class Codes: G01L 19/12 Alarms or signals
• CPC Class Codes: G10L 19/038 Vector quantisation, e.g. TwinVQ audio
• CPC Class Codes: G10L 19/04 using predictive techniques
• CPC Class Codes: G10L 19/107 Sparse pulse excitation, e.g. by using algebraic codebook
• CPC Class Codes: G10L 19/12 the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
• CPC Class Codes: G10L 2019/0005 Multi-stage vector quantisation
• CPC Class Codes: G10L 21/00 Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility G10L19/00 takes precedence

Abstract

A vector joint encoding/decoding method and a vector joint encoder/decoder are provided, more than two vectors are jointly encoded, and an encoding index of at least one vector is split and then combined between different vectors, so that encoding idle spaces of different vectors can be recombined, thereby facilitating saving of encoding bits, and because an encoding index of a vector is split and then shorter split indexes are recombined, thereby facilitating reduction of requirements for the bit width of operating parts in encoding/decoding calculation.

Claims

1. A vector joint encoding method for a voice signal performed by a hardware, comprising:

calculating an encoding index Ind_t of each vector, wherein the each vector is obtained by dividing the voice signal, wherein a subscript t denotes a t^th vector, tε

[0, T−

1], and T is the number of vectors and is an integer greater than or equal to 2;

splitting at least one Ind_t at least once into at least two sections, wherein the splitting at least once is equivalent to splitting the Ind_t into two split indexes Ind_t0 and Ind_t1 according to a set factor SLF_t, wherein the set factor SLF_t is the set split factor SLF for the t^th vector, the SLF_t is a positive integer, the Ind_t0 denotes a serial number of an interval to which the Ind_t belongs, the Ind_t1 denotes a serial number of the Ind_t in the interval to which the Ind_t belongs, the length of the interval is not greater than the SLF_t, and

Ind_t≦

Ind_t0×

SLF_t+Ind_t1;

combining a split index of at least one vector and split indexes or encoding indexes of other vectors to generate a combined index Ind_SLF; and

performing encoding according to the combined index and other uncombined split indexes.

2. The method according to claim 1, wherein the vector is represented as pulse distribution on a track, and the encoding index is an index used for indicating the pulse distribution on the track.

3. The method according to claim 1, wherein the splitting the Ind_t into two split indexes Ind_t0 and Ind_t1 according to a set factor SLF_t is specifically:

Ind_t0=Int(Ind_t/SLF_t), wherein Int( )denotes rounding down to an integer, and

Ind_t1=Ind_t%SLF_t, wherein % denotes taking a remainder.

4. The method according to claim 3, wherein

SLF_t=2^(K_t), wherein K_t is a positive integer, and the combining a split index of at least one vector and split indexes or encoding indexes of other vectors to generate a combined index Ind_SLF is specifically:

for a vector t providing a split index to participate in combination, selecting the Ind_t0 to participate in combination;

or,

SLF_t=Int(Ind_t,max/2^(K_t)), wherein Ind_t,max denotes a maximal value of the Ind_t, and the combining a split index of at least one vector and split indexes or encoding indexes of other vectors to generate a combined index Ind_SLF is;

for the vector t providing a split index to participate in combination, selecting the Ind_t1 to participate in combination.

5. The method according to claim 4, wherein K_t=K_t,max−

2, or K_t=K_t,max−

3, or K_t=K_t,max−

4, wherein the K_t,max is a length of a bit sequence of the Ind_t,max.

6. The method according to claim 1, wherein the combining a split index of at least one vector and split indexes or encoding indexes of other vectors to generate a combined index Ind_SLF is specifically:

for a vector t providing a split index to participate in combination, selecting a split index capable of embodying space occupancy characteristics of the Ind_t to participate in combination, wherein being capable of embodying space occupancy characteristics of the Ind_t refers to that, compared with other split indexes, an occupancy rate of a value range of a selected split index for an encoding space is the closest to an occupancy rate of a value range of the Ind_t for the encoding space.

7. The method according to claim 1, wherein the splitting the Ind_t into two split indexes Ind_t0 and Ind_t1 according to a set factor SLF_t is specifically:

selecting a value of bits of the Ind_t as the Ind_t0, the bits being located at a set position, and selecting a value of bits at a remaining position as the Ind_t1, wherein SLF_t=2^(K_t), the number of the bits at the set position is K_t0, the number of the bits at the remaining position is K_t, K_t0+K_t=K_t,max, K_t,max is a length of a bit sequence of Ind_t,max, and the Ind_t,max denotes a maximal value of the Ind_t.

8. The method according to claim 7, wherein combining split indexes from at least two vectors to generate a combined index Ind_SLF is specifically:

for a vector t providing a split index to participate in combination, selecting, in a bit sequence with a length being K_t,max, a split index of the Ind_t to participate in combination, wherein the split index at least comprises a value of the highest two bits.

9. The method according to claim 1, wherein the performing encoding according to the combined index and other uncombined split indexes is:

splitting the combined index into T1 recombined indexes Ind_t0′

according to a set value range, wherein T1 is less than or equal to the number of vectors generating the combined index, a value range of at least one Ind_t0′

is greater than a value range of the split index of a corresponding vector t, wherein the split index participates in combination, and a value range of at least one Ind_t0′

is less than the value range of the split index of the corresponding vector t, wherein the split index participates in combination; and

respectively combining each recombined index and an uncombined split index of a corresponding vector and then performing encoding, and if a vector without being allocated a recombined index exists, encoding an uncombined split index of the vector.

10. The method according to claim 9, wherein the splitting the combined index into T1 recombined indexes Ind_t0′

according to a set value range is specifically;

splitting a total length K_SLF of the bit sequence of the combined index into T1 sections according to a set length, wherein a value of each section corresponds to one Ind_t0′

, the K_SLF is the length of the bit sequence of an Ind_SLF,max, and the Ind_SLF,max denotes a maximal value of the Ind_SLF.

11. The method according to claim 10, wherein

the splitting the Ind_t into two split indexes Ind_t0 and Ind_t1 according to a set factor SLF_t is specifically:

in the bit sequence with the length being K_t,max, selecting a value of K_t0 bits, starting from the highest bit, of the Ind_t, as the Ind_t0, and selecting a value of remaining bits as the Ind_t1, wherein SLF_t=2^(K_t), K_t0+K_t=K_t,max, the K_t,max is the length of the bit sequence of the Ind_t,max, and the Ind_t,max denotes a maximal value of the Ind_t; and

the combining a split index of at least one vector and split indexes or encoding indexes of other vectors to generate a combined index Ind_SLF is specifically;

for a vector t needs to provide a split index, selecting the Ind_t0 to participate in combination; and

the splitting the total length K_SLF of the bit sequence of the combined index into T1 sections according to a set length is specifically;

splitting the K_SLF according to a K_t0 value used by the vector t generating the combined index, wherein the number of bits split by each Ind_t0'"'"' correspondingly is less than or equal to the K_t0 value used by the corresponding vector t.

12. The method according to claim 1, wherein the performing encoding according to the combined index and other uncombined split indexes is:

comparing the combined index Ind_SLF and adjusting a threshold value THR, wherein,

THR≦

2^(K_SLF)−

Ind_SLF,max,

the K_SLF is the length of the bit sequence of the Ind_SLF,max, and the Ind_SLF,max denotes a maximal value of the Ind_SLF;

if the Ind_SLF is less than the THR, encoding the Ind_SLF by using a first number of encoding bits;

otherwise, encoding the Ind_SLF added with an offset value THR₀ by using a second number of encoding bits, wherein THR≦

THR₀≦

2^(K_SLF)−

Ind_SLF,max, the first number is less than the second number, the second number is less than or equal to the K_SLF, and the first number and the second number are both positive integers; and

encoding other uncombined split indexes.

13. A vector joint pulse encoding method for a voice signal performed by a hardware, comprising:

grouping vectors participating in joint encoding, wherein each group at least comprises two vectors;

in each group, calculating an encoding index of each vector, wherein the each vector is obtained by dividing the voice signal;

splitting at least two encoding indexes;

splitting each encoding index at least once into at least two sections, wherein splitting at least once is equivalent to splitting an encoding index into two one-stage split indexes according to a set one-stage factor, one one-stage split index denotes serial numbers of several intervals with a length not greater than a set value, the other one-stage split index denotes a serial number of the encoding index in an interval to which the encoding index belongs; and

combining one-stage split indexes from at least two vectors to generate a one-stage combined index;

starting from m=2, repeating the following operation of generating an m-stage combined index until m=M;

splitting at least two (m−

1)-stage combined indexes, wherein mε

[2, M], M is an integer greater than or equal to 2;

splitting each (m−

1)-stage combined index at least once into at least two sections, wherein splitting at least once is equivalent to splitting an (m−

1)-stage combined index into two m-stage split indexes according to a set m-stage factor; and

combining m-stage split indexes from at least two (m−

1)-stage combined indexes to generate an m-stage combined index; and

encoding according to the M-stage combined index and other uncombined one-stage to (M−

1)-stage split indexes.

14. The method according to claim 13, wherein the one-stage split index used for generating the one-stage combined index is a value intercepted from a set bit of a corresponding encoding index starting from the highest bit;

and the m-stage split index used for generating the m-stage combined index is a value intercepted from a set bit of a corresponding (m−

1)-stage combined index starting from the highest bit.

15. A vector joint pulse decoding method for a voice signal performed by a hardware, comprising:

acquiring a joint code, and acquiring, from the joint code, a combined index and an uncombined split index corresponding to a vector, wherein the vector is obtained by dividing the voice signal;

splitting the combined index into split indexes corresponding to the vector, or splitting the combined index into a split index and an encoding index that correspond to the vector;

for each vector participating in splitting the encoding index, splicing, according to a split manner of an encoding end, a split index not participating in combination and a split index participating in combination of the vector to generate the encoding index of the vector; and

reestablishing the vector according to the encoding index of the vector for each vector.

16. The method according to claim 15, wherein the acquiring, from the joint code, a combined index and an uncombined split index corresponding to each vector is:

extracting, from the joint code, a code corresponding to each vector;

splitting a recombined index and an uncombined split index from the code of each vector; and

if a code comprising no recombined index exists, directly obtaining an uncombined split index corresponding to a corresponding vector; and

splicing all recombined indexes into a combined index according to the split manner of the encoding end.

17. A vector joint pulse decoding method for a voice signal performed by a hardware, comprising:

acquiring a joint code, and acquiring, from the joint code, an M-stage combined index, an uncombined m-stage split index corresponding to each (m−

1)-stage combined index, and an uncombined one-stage split index corresponding to each vector, wherein M is an integer greater than or equal to 2, and mε

[2, M], wherein the each vector is obtained by dividing the voice signal;

starting from m=M, repeating the following operation of generating an (m−

1)-stage combined index until m=2;

splitting each m-stage combined index into an m-stage split index corresponding to each (m−

1)-stage combined index for generating the m-stage combined index; and

for each (m−

1)-stage combined index, splicing, according to a split manner of an encoding end, an m-stage split index not participating in combination and an m-stage split index participating in combination of the (m−

1)-stage combined index to generate the (m−

1)-stage combined index according to a split manner of an encoding end;

splitting each one-stage combined index into a one-stage split index corresponding to each vector in a vector group for generating the one-stage combined index; and

for each vector in each vector group, splicing, according to the split manner of the encoding end, a one-stage split index not participating in combination and a one-stage split index participating in combination of the vector to generate an encoding index; and

reestablishing the vector according to the encoding index for each vector in each vector group.

18. A vector joint pulse encoder for encoding a voice signal, comprising:

an encoding index calculation unit, configured to calculate an encoding index Ind_t of each vector, wherein the each vector is obtained by dividing the voice signal, wherein a subscript t denotes a t^th vector, tε

[0, T−

1], and T is an integer greater than or equal to 2;

a vector index splitting unit, configured to split at least one Ind_t at least once into at least two sections, wherein splitting at least once is equivalent to splitting the Ind_t into two split indexes Ind_t0 and Ind_t1 according to a set factor SLF_t, wherein the set factor SLF_t is the set split factor SLF for the t^th vector, the SLF_t is a positive integer, the Ind_t0 denotes a serial number of an interval to which the Ind_t belongs, the Ind_t1 denotes a serial number of the Ind_t in the interval to which the Ind_t belongs, a length of the interval is not greater than the SLF_t, and Ind_t≦

Ind_t0×

SLF_t+Ind_t1;

an index combination unit, configured to combine a split index of at least one vector and split indexes or encoding indexes of other vectors to generate a combined index Ind_SLF; and

an encoding unit, configured to perform encoding according to the combined index and other uncombined split indexes.

19. The encoder according to claim 18, wherein the encoding unit comprises:

a recombining sub-unit, configured to split the combined index into T1 recombined indexes Ind_t0′

according to a set value range, wherein T1 is less than or equal to the number of vectors generating the combined index, a value range of at least one Ind_t0′

is greater than a value range of the split index, participating in combination, of a corresponding vector t, and a value range of at least one Ind_t0′

is less than the value range of the split index, participating in combination, of the corresponding vector t; and

an encoding sub-unit, configured to combine each recombined index and an uncombined split index of a corresponding vector respectively and then perform encoding, and if a vector without being allocated a recombined index exists, encode an uncombined split index of the vector.

20. A vector joint pulse encoder for encoding a voice signal, comprising:

a vector index calculation unit, configured to group vectors participating in joint encoding, wherein each group at least comprises two vectors;

in each group, calculate an encoding index of each vector, wherein the each vector is obtained by dividing the voice signal;

split at least two encoding indexes;

split each encoding index at least once into at least two sections, wherein splitting at least once is equivalent to splitting an encoding index into two one-stage split indexes according to a set one-stage factor, one one-stage split index denotes serial numbers of several intervals with a length not greater than a set value, the other one-stage split index denotes a serial number of the encoding index in an interval to which the encoding index belongs; and

combine one-stage split indexes from at least two vectors to generate a one-stage combined index;

a cascade calculation unit, configured to;

starting from m=2, repeat the following operation of generating an m-stage combined index until m=M;

split at least two (m−

1)-stage combined indexes, wherein mε

[2, M], M is an integer greater than or equal to 2;

split each (m−

1)-stage combined index at least once into at least two sections, wherein splitting at least once is equivalent to splitting an (m−

1)-stage combined index into two m-stage split indexes according to a set m-stage factor; and

combine m-stage split indexes from at least two (m−

1)-stage combined indexes to generate an m-stage combined index; and

an encoding unit, configured to perform encode according to the M-stage combined index and other uncombined one-stage to (M−

1)-stage split indexes.

21. A vector joint pulse decoder for decoding a voice signal, comprising:

a decoding unit, configured to acquire a joint code, and acquire, from the joint code, a combined index and an uncombined split index corresponding to a vector, wherein the vector is obtained by dividing the voice signal;

an index splitting unit, configured to split the combined index into split indexes corresponding to the vector, or split the combined index into a split index and an encoding index that correspond to the vector;

a vector index reestablishing unit, configured to;

for each vector participating in splitting the encoding index, splice, according to a split manner of an encoding end, a split index not participating in combination and a split index participating in combination of the vector to generate the encoding index of the vector; and

a vector reestablishing unit, configured to reestablish the vector according to the encoding index of the vector for each vector.

22. The decoder according to claim 21, wherein the decoding unit comprises:

a decoding sub-unit, configured to extract, from the joint code, a code corresponding to each vector;

split a recombined index and an uncombined split index from the code of each vector; and

if a code comprising no recombined index exists, directly obtain an uncombined split index corresponding to a corresponding vector; and

a splicing sub-unit, configured to splice all recombined indexes into a combined index according to the split manner of the encoding end.

23. A vector joint pulse decoder for decoding a voice signal, comprising:

a decoding unit, configured to acquire a joint code, and acquire, from the joint code, an M-stage combined index, an uncombined m-stage split index corresponding to each (m−

1)-stage combined index, and an uncombined one-stage split index corresponding to each vector, wherein the each vector is obtained by dividing the voice signal, wherein M is an integer greater than or equal to 2, and mε

[2, M];

a cascade recovering unit, configured to;

starting from m=M, repeat the following operation of generating an (m−

1)-stage combined index until m=2;

split each m-stage combined index into an m-stage split index corresponding to each (m−

1)-stage combined index for generating the m-stage combined index; and

for each (m−

1)-stage combined index, splice, according to a split manner of an encoding end, an m-stage split index not participating in combination and an m-stage split index participating in combination of the (m−

1)-stage combined index to generate the (m−

1)-stage combined index;

a vector index reestablishing unit, configured to split each one-stage combined index into a one-stage split index corresponding to each vector in a vector group for generating the one-stage combined index; and

for each vector in each vector group, splice, according to the split manner of the encoding end, a one-stage split index not participating in combination and a one-stage split index participating in combination of the vector to generate an encoding index; and

a vector reestablishing unit, configured to reestablish the vector according to the encoding index for each vector in each vector group.