CN102355269B - One-dimensional segment code signal rapid decoding method based on GDHT-III domain - Google Patents

One-dimensional segment code signal rapid decoding method based on GDHT-III domain Download PDF

Info

Publication number
CN102355269B
CN102355269B CN 201110211841 CN201110211841A CN102355269B CN 102355269 B CN102355269 B CN 102355269B CN 201110211841 CN201110211841 CN 201110211841 CN 201110211841 A CN201110211841 A CN 201110211841A CN 102355269 B CN102355269 B CN 102355269B
Authority
CN
China
Prior art keywords
iii
gdht
centerdot
sequence
length
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN 201110211841
Other languages
Chinese (zh)
Other versions
CN102355269A (en
Inventor
舒华忠
伍家松
王膂
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southeast University
Original Assignee
Southeast University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southeast University filed Critical Southeast University
Priority to CN 201110211841 priority Critical patent/CN102355269B/en
Publication of CN102355269A publication Critical patent/CN102355269A/en
Application granted granted Critical
Publication of CN102355269B publication Critical patent/CN102355269B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

The invention discloses a one-dimensional segment code signal rapid decoding method based on a GDHT (generalized discrete Hartley transform)-III domain, belonging to the signal processing technology field. GDHT-III domain coefficients {Ak}, {Bk} and {Ck} (k is an integer from 0 to N/3-1) of signal sequences {an}, {bn} and {cn} with a length of N/3 (n is an integer from 0 to N/3-1) are converted into a GDHT-III domain coefficient {Xi} (i is an integer from 0 to N-1) of an original code signal sequence {xm} (m is an integer from 0 to N-1) of a length of N, wherein, calculation of {Xi} is divided into a multiple of 3 output index {X3k}, a multiple of 3 with a reminder 1 output index {X3k+1} and a multiple of 3 with a reminder 2 output index {X3k+2} to carry out calculation respectively, thusGDHT-III transformation times are reduced, and calculation complexity of a decoding process is reduced. Compared with the prior art, the method in the invention has the characteristics of lower complexity, better decoding real-time property, and less signal distortion.

Description

A kind of one dimension segment encoding signal fast decoding method based on the GDHT-III territory
Technical field
The present invention relates to a kind of signal decoding method, relate in particular to a kind of one dimension segment encoding signal fast decoding method based on the GDHT-III territory, belong to the signal processing technology field.
Background technology
Encoding and decoding are extremely important parts in the Digital Signal Processing, and coding refers to an input signal is converted to code, and this code is being beneficial to transmission or storing of optimised mistake, and decoding then is the reverse procedure of coding.Encoding-decoding process is finished by coding and decoding device usually.Common signal cataloged procedure generally includes time domain direct transform, quantification, entropy this plurality of processes of encoding, and decode procedure comprises anti-entropy coding, inverse quantization and frequency domain inverse conversion.
Discrete Hartley transform (Discrete Hartley Transform:DHT) is a kind of important mathematical tool in the Digital Signal Processing, and it can describe the relation of the time-domain and frequency-domain of discrete signal, and important status is arranged in Digital Signal Processing.As the expansion of DHT, the discrete Hartley transform of broad sense (Generalized Discrete Hartley Transform:GDHT) can be applied to field more widely.GDHT has four kinds of forms, is respectively that GDHT-I(is DHT), GDHT-II, GDHT-III and GDHT-IV.The real number Properties of Some Mapping that GDHT kernel function itself is intrinsic makes it be very suitable for handling the real signal sequence.
List entries { x m, m=0,1 ..., the GDHT-III of N – 1 is defined as
X i = GDH T N III { x m } = 1 N Σ m = 0 N - 1 x m cas πm ( 2 i + 1 ) N , i = 0,1 , . . . , N - 1 ,
Its inverse transformation (IGDHT-III) is defined as
x m = IGDH T N III { X i } = Σ i = 0 N - 1 X i cas πk ( 2 m + 1 ) N , m = 0,1 , . . . , N - 1 ,
Wherein N is sequence length and cas α=cos α+sin α. we ignore the normalization factor 1/N in the formula (1) in the following discussion, because it is a constant division calculation.Such as: for N=2 l, l 〉=2, its needs are done simple shifting function.
GDHT-III(and IGDHT-III) as DFT(and IDFT) a kind of the replacement at the compression coding and decoding of handling real signal then have its special advantages: the first, its kernel function is real function, when being input as real signal, only need carry out real arithmetic, therefore have lower computation complexity than DFT; The second, its positive and negative conversion has identical form, therefore can realize positive and negative conversion with same module.
In existing decoding method based on the GDHT-III conversion, need the signal { x that sends mLength is long usually, sends so need carry out segment encoding to signal, wherein a kind of common situation is with { x mBe divided into three sections { a n, { b nAnd { c n, i.e. a n=x n, b n=x N+N/3, c n=x N+2N/3, n=0,1 ..., N/3 – 1.At first with { a n, { b nAnd { c nConversion obtains its corresponding coefficient in transform domain { A through GDHT-III respectively k, { B kAnd { C k, then to these coefficients quantize, obtain after the processing such as entropy coding coefficient A ' k, B ' kAnd C ' kBe sent to receiving terminal.In when decoding, at first to the coefficient that receives A ' k, B ' kAnd C ' kCarry out the coefficient { A that processing such as anti-entropy coding and inverse quantization are restored respectively k, { B kAnd { C k, the key issue that wherein relates to is how to pass through { A k, { B kAnd { C kCalculate { X i({ X wherein iBe { x nLength be the coefficient of the GDHT-III of N)? because the encoding and decoding of signal are quite high to the requirement of real-time, so under the situation of ensuring the quality of products, require complexity more low more good.Existing method is to be the GDFT-II domain coefficient { A of N/3 with the length of importing earlier k, { B kAnd { C kGain time domain by the IGDHT-III contravariant respectively and obtain original time-domain signal { a n, { b nAnd { c n, then these three sequence tandem compounds are become { x m, computational length is the sequence { x of N again mThe coefficient { X of GDHT-III i.Can know that thus the IGDHT-III that it is N/3 that traditional method need be calculated three length and the GDHT-III that length is N need higher computation complexity, thereby influence the real-time of decoding to a certain extent.
Summary of the invention
Technical problem to be solved by this invention is to overcome the technical problem of the existing existing computation complexity height of one dimension segment encoding signal decoding method based on the GDHT-III territory, decoding real-time difference, a kind of one dimension segment encoding signal fast decoding method based on the GDHT-III territory is provided, this method has lower computation complexity, and the decoding real-time is better.
The present invention is by the following technical solutions:
A kind of one dimension segment encoding signal fast decoding method based on the GDHT-III territory, described segment encoding signal is by being that the original signal sequence of N is divided into the burst that three segment length are N/3 with length, respectively this three segment signals sequence is carried out the GDHT-III conversion then and obtain its corresponding GDHT-III domain coefficient, at last to these three groups of GDHT-III domain coefficients quantize respectively, entropy coding handles and obtains, described fast decoding method may further comprise the steps:
Step 1, to the segment encoding signal carry out anti-entropy coding, inverse quantization is handled three groups of GDHT-III domain coefficients that are restored;
Step 2, establish three groups of GDHT-III domain coefficients that step 1 obtains and be respectively { A k, { B kAnd { C k, k=0,1 ...,
N/3 – 1 is respectively according to the following formula sequence of calculation { X 3k, { X 3k+1, { X 3k+2, k=0 wherein, 1 ..., N/3 – 1:
X 3k+1=A k-B k+C k,k=0,1,…,N/3–1
X 3 k = 1 2 ( Y k + Z k ) , k = 0,1,2 · · · · · · , N / 3 - 1 ,
X 3 k + 2 = 1 2 ( Y k - Z k ) , k = 0,1,2 · · · · · · , N / 3 - 1
Wherein,
Y k = GDH T N / 3 III { IGDH T N / 3 III ( 2 A k + B k - C k ) cos θ n + 3 IGDH T N / 3 III ( B k + C k ) sin θ n } ,
Z N / 3 - 1 - k = GDH T N / 3 III { - IGDH T N / 3 III ( 2 A k + B k - C k ) sin θ n + 3 IGDH T N / 3 III ( B k + C k ) cos θ n } ,
In the formula,
Figure GDA00003349345400035
With
Figure GDA00003349345400036
Represent that respectively be forward and the oppositely GDHT-III conversion of N/3 to the burst in the bracket as length, θ n=2 π n/N are twiddle factors;
Step 3, with sequence { X 3k, { X 3k+1, { X 3k+2In element successively tandem compound obtain sequence { X i; K=0 wherein, 1 ..., N/3 – 1; I=0,1 ..., N – 1; Sequence { X iBe the GDHT-III domain coefficient that length is the original signal sequence of N.
Compared to existing technology, the computation complexity that the inventive method has is lower, and the real-time of decoding is better.The inventive method also has distorted signals still less, and this is because as a rule, the step that and then can quantize after the signal process GDHT-III conversion, and utilize IGDHT-III can cause the distortion of signal with the coefficient after quantizing.Signal after the distortion carries out the GDHT-III conversion again will make error further strengthen.Therefore, in order to reduce the distortion of signal, our will try one's best number of less IGDHT-III and GDHT-III.Traditional method need be carried out three IGDHT-III and three GDHT-III, and the inventive method only needs twice IGDHT-III and twice GDHT-III.Therefore the inventive method has distorted signals still less.
Description of drawings
Fig. 1 carries out the schematic flow sheet of segment encoding for existing method;
Fig. 2 carries out the schematic flow sheet of segmentation decoding for existing method;
Fig. 3 is the signal flow graph of the fast decoding method based on the GDHT-III conversion of the present invention; Wherein with the numeral transmission factor (being equivalent to multiplier) by the line segment of arrow, " ⊕ " represents adder." anti-pleat " expression is with burst { Z N/3-1-kConvert { Z to k, k=0,1 ...., N/3-1.
Embodiment
Below in conjunction with accompanying drawing technical scheme of the present invention is elaborated:
Fig. 1 has shown the flow process of traditional segment encoding, the signal { x that at first will send mBe divided into three sections { a n, { b nAnd { c n, i.e. a n=x n, b n=x N+N/3, c n=x N+2N/3, n=0,1 ..., N/3 – 1, and respectively to { a n, { b nAnd { c nCarry out the GDHT-III conversion and obtain its corresponding coefficient { A k, { B kAnd { C k, k=0,1 ..., N/3 – 1, then to these coefficients quantize, entropy coding obtains coefficient after handling A ' k, B ' kAnd C ' k, it is sent to receiving terminal or is stored in the medium.
Fig. 2 has shown that conventional method carries out the flow process of segmentation decoding, at first to the coefficient that receives A ' k, B ' kAnd C ' kCarry out the coefficient { A that anti-entropy coding and inverse quantization processing are restored respectively k, { B kAnd { C k, and these coefficients are obtained original time-domain signal { a by IGDHT-III respectively n, { b nAnd { c n, then these three sequence tandem compounds are become { x m, computational length is the sequence { x of N again nThe coefficient { X of GDHT-III i.When adopting conventional method, if input { x mBe real signal, its computation complexity is
M T ( N ) = 6 M III ( N / 3 ) + 4 N / 3 = Nlo g 2 ( N / 3 ) + N / 3 A T ( N ) = 6 A III ( N / 3 ) + 8 N / 3 = 3 Nlo g 2 ( N / 3 ) - N / 3 + 24 N = 3 × 2 l , l ≥ 2 .
M wherein III(N) and A III(N) be respectively that computational length is the needed multiplication number of real number GDHT-III and the addition number of N.
Fig. 3 has provided the specific implementation flow graph that carries out the decoding of N point real number signal with the inventive method, and wherein input is that length is the real number signal { a of N/3 n, { b nAnd { c nGDHT-III domain coefficient { A k, { B kAnd { C k; Output is that length is the real number signal { x of N mGDHT-III domain coefficient { X i, i=0,1 ..., pass through { X among N – 1 figure 3k, { X 3k+1And { X 3k+2K=0,1 ..., 1, three part of N/3 – is expressed.
When adopting the inventive method to decode, the length GDHT-III that is N the GDHT-III that to be decomposed into three length be N/3 is calculated, be about to output { X iCalculating be divided into 3 multiple output index { X 3k, 3 multiple is surplused 1 output index { X 3k+1And 3 multiple surplus 2 output index { X 3k+2Three parts calculate respectively.
Index part { X 3k+1Obtain according to following formula
X 3 k + 1 = Σ m = 0 N - 1 x m cas πm ( 2 ( 3 k + 1 ) + 1 ) N
= Σ n = 0 N / 3 - 1 a n cas πn ( 2 k + 1 ) N / 3 - Σ n = 0 N / 3 - 1 b n cas πn ( 2 k + 1 ) N / 3 , k = 0,1 , . . . , N / 3 - 1 .
+ Σ n = 0 N / 3 - 1 c n cas πn ( 2 k + 1 ) N / 3
= A k - B k + C k
Index part { X 3kAnd { X 3k+1Calculate according to following two steps:
The first step: structure also calculates intermediate quantity Y kAnd Z k, k=0,1 ..., N/3 – 1 order
Y k=X 3k+X 3k+2,k=0,1,…,N/3–1
Z k=X 3k-X 3k+2,k=0,1,…,N/3–1
Y then kAnd Z kObtain according to following formula
Y k = Σ m = 0 N - 1 x m cas πm ( 6 k + 1 ) N + Σ m = 0 N - 1 x m cas πm ( 6 k + 5 ) n
= Σ m = 0 N - 1 2 x m cos 2 πm N cas πm ( 2 k + 1 ) N / 3
= Σ m = 0 N / 3 - 1 { ( 2 a n + b n - c n ) cos θ n + 3 ( b n + c n ) sin θ n } cas πm ( 2 k + 1 ) N / 3
= GDH T N / 3 III { ( 2 a n + b n - c n ) cos θ n + 3 ( b n + c n ) sin θ n }
= GDHT N / 3 III { IGDH T N / 3 III ( 2 A k + B k - C k ) cos θ n + 3 IGDH T N / 3 III ( B k + C k ) sin θ n }
Z N / 3 - 1 - k = Σ m = 0 N - 1 x m cas πm ( 6 ( N / 3 - 1 - k ) + 1 ) N - Σ m = 0 N - 1 x m cas πm ( 6 ( N / 3 - 1 - k ) + 5 ) N
= - Σ m = 0 N - 1 2 x m sin 2 πm N cas πm ( 2 k + 1 ) N / 3
= Σ m = 0 N / 3 - 1 { - ( 2 a n + b n - c n ) sin θ n + 3 ( b n + c n ) cos θ n } cas πm ( 2 k + 1 ) N / 3
= GDH T N / 3 III { - ( 2 a n + b n - c n ) sin θ n + 3 ( b n + c n ) cos θ n }
= GDHT N / 3 III { - IGDH T N / 3 III ( 2 A k + B k - C k ) sin θ n + 3 IGDH T N / 3 III ( B k + C k ) cos θ n }
K=0 wherein, 1 ..., N/3 – 1,
Figure GDA000033493454000511
With
Figure GDA000033493454000512
Represent that respectively be forward and the oppositely GDHT-III conversion of N/3 to the burst in the bracket as length, θ n=2 π n/N are twiddle factors.
Second step: by intermediate quantity Y kAnd Z k, computation index part { X 3kAnd { X 3k+2}
{ X 3kAnd { X 3k+2Can be obtained by following formula respectively
X 3 k = 1 2 ( Y k + Z k ) , k=0,1,…,N/3–1
X 3 k + 2 = 1 2 ( Y k - Z k ) ,
With sequence { X 3k, { X 3k+1, { X 3k+2In element successively tandem compound can obtain the GDHT-III domain coefficient { X that length is the original signal sequence of N i.
When adopting the inventive method decoding, if input { x mBe real signal, its computation complexity is:
M P ( N ) = 4 M III ( N / 3 ) + 4 N / 3 = ( 2 N / 3 ) lo g 2 ( N / 3 ) + 2 N / 3 A P ( N ) = 4 A III ( N / 3 ) + 3 N = 2 Nlo g 2 ( N / 3 ) + N + 16 , N = 3 × 2 l , l ≥ 2 .
M wherein III(N) and A III(N) be respectively that computational length is the needed multiplication number of real number GDHT-III and the addition number of N.
Following table 1 has shown the computation complexity contrast (being input as real number signal) when adopting the inventive method and adopting the conventional method decoding.
Table 1
As can be seen from Table 1, coding/decoding method of the present invention is more effective than conventional method.For the real number input signal, when sequence length N was increased to 192 from 12, the inventive method had been saved 17% to 26% computation complexity than conventional method.Simultaneously, because the present invention has used the GDHT-III/IGDHT-III conversion of less number of times, therefore has distorted signals still less.

Claims (1)

1. one dimension segment encoding signal fast decoding method based on the GDHT-III territory, described segment encoding signal is by being that the original signal sequence of N is divided into the burst that three segment length are N/3 with length, respectively this three segment signals sequence is carried out the GDHT-III conversion then and obtain its corresponding GDHT-III domain coefficient, at last these three groups of GDHT-III domain coefficients are quantized respectively, the entropy coding is handled and is obtained, it is characterized in that, described fast decoding method may further comprise the steps: step 1, the segment encoding signal is carried out anti-entropy coding, inverse quantization is handled, three groups of GDHT-III domain coefficients that are restored;
Step 2, establish three groups of GDHT-III domain coefficients that step 1 obtains and be respectively { A k, { B kAnd { C k, k=0,1 ..., N/3 – 1 is respectively according to the following formula sequence of calculation { X 3k, { X 3k+1, { X 3k+2, k=0 wherein, 1 ..., N/3 – 1:X 3k+1=A k-B k+ C k, k=0,1 ..., N/3 – 1
X 3 k = 1 2 ( Y k + Z k ) , k = 0,1,2 · · · · · · , N / 3 - 1 ,
X 3 k + 2 = 1 2 ( Y k - Z k ) , k = 0,1,2 · · · · · · , N / 3 - 1
Wherein, Z kBy with following Z N/3-1-kAnti-pleat is converted to,
Y k = GDH T N / 3 III { IGDH T N / 3 III ( 2 A k + B k - C k ) cos θ n + 3 IGDH T N / 3 III ( B k + C k ) sin θ n } ,
Z N / 3 - 1 - k = GDH T N / 3 III { - IGDH T N / 3 III ( 2 A k + B k - C k ) sin θ n + 3 IGDH T N / 3 III ( B k + C k ) cos θ n } ,
In the formula,
Figure FDA00003349345300013
With
Figure FDA00003349345300016
Represent that respectively be forward and the oppositely GDHT-III conversion of N/3 to the burst in the bracket as length, θ n=2 π n/N are twiddle factors;
Step 3, with sequence { X 3k, { X 3k+1, { X 3k+2In element successively tandem compound obtain sequence { X i; K=0 wherein, 1 ..., N/3 – 1; I=0,1 ..., N – 1; Sequence { X iBe the GDHT-III domain coefficient that length is the original signal sequence of N.
CN 201110211841 2011-07-27 2011-07-27 One-dimensional segment code signal rapid decoding method based on GDHT-III domain Expired - Fee Related CN102355269B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110211841 CN102355269B (en) 2011-07-27 2011-07-27 One-dimensional segment code signal rapid decoding method based on GDHT-III domain

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110211841 CN102355269B (en) 2011-07-27 2011-07-27 One-dimensional segment code signal rapid decoding method based on GDHT-III domain

Publications (2)

Publication Number Publication Date
CN102355269A CN102355269A (en) 2012-02-15
CN102355269B true CN102355269B (en) 2013-09-25

Family

ID=45578774

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110211841 Expired - Fee Related CN102355269B (en) 2011-07-27 2011-07-27 One-dimensional segment code signal rapid decoding method based on GDHT-III domain

Country Status (1)

Country Link
CN (1) CN102355269B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070201574A1 (en) * 2006-02-24 2007-08-30 Chih-Feng Wu Discrete multi-tone system having DHT-based frequency-domain equalizer
CN102036075A (en) * 2010-12-29 2011-04-27 东南大学 Image and digital video coding and decoding methods

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070201574A1 (en) * 2006-02-24 2007-08-30 Chih-Feng Wu Discrete multi-tone system having DHT-based frequency-domain equalizer
CN102036075A (en) * 2010-12-29 2011-04-27 东南大学 Image and digital video coding and decoding methods

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Generalized Discrete Hartley Transforms;Neng-Chung Hu等;《IEEE TRANSACTIONS ON SIGNAL PROCESSING》;19921231;第40卷(第12期);2931-2940 *
Neng-Chung Hu等.Generalized Discrete Hartley Transforms.《IEEE TRANSACTIONS ON SIGNAL PROCESSING》.1992,第40卷(第12期),2931-2940.
成礼智.离散Hartley变换(DHT)及其快速算法.《数值计算与计算机应用》.1988,(第3期),162-168. *

Also Published As

Publication number Publication date
CN102355269A (en) 2012-02-15

Similar Documents

Publication Publication Date Title
CN107832837B (en) Convolutional neural network compression method and decompression method based on compressed sensing principle
CN103201950B (en) The combined process estimator with variable tap delay line in power amplifier digital pre-distortion
EP3518235B1 (en) Method and apparatus for generating from a coefficient domain representation of hoa signals a mixed spatial/ coefficient domain representation of said hoa signals
CN1860527B (en) Device and method for processing a signal with a sequence of discrete values
CN100416553C (en) Device and method for conversion into a transformed representation or for inversely converting the transformed representation.
CN102132494B (en) Method and apparatus of communication
CN101197988B (en) Amount-of-compressed data control method and image data compressing apparatus
CN103262164A (en) Cross product enhanced subband block based harmonic transposition
CN102334159A (en) Encoder, decoder, and method therefor
CN106506430A (en) A kind of new algorithm of the compensation peak-to-average force ratio non-linear distortion based on compressed sensing technology
CN104584122A (en) Linear prediction based audio coding using improved probability distribution estimation
CN102036075B (en) Image and digital video coding and decoding methods
CN102685501A (en) Fixed-point wavelet transform method for joint photographic experts group 2000 (JPEG2000) image compression
CN102084594A (en) Method for treating digital data
CN102355269B (en) One-dimensional segment code signal rapid decoding method based on GDHT-III domain
CN101960515B (en) Method and apparatus for transforming between different filter bank domains
CN102132342B (en) Method for updating an encoder by filter interpolation
CN101572092B (en) Method and device for searching constant codebook excitations at encoding and decoding ends
CN103299307A (en) Estimator for estimating a probability distribution of a quantization index
CN102362231A (en) Analog computation using numerical representations with uncertainty
CN100570597C (en) Digital signal is transformed to the method for frequency field and reciprocal transformation thereof from time domain
CN101944235B (en) Image compression method based on fractional fourier transform
CN102103859B (en) Methods and devices for coding and decoding digital audio signals
CN102215406B (en) DCT (Discrete Cosine Transformation)-based fast decoding method of segmented encoded signal
CN112911289B (en) DCT/IDCT transformation optimization method and system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20130925

Termination date: 20160727

CF01 Termination of patent right due to non-payment of annual fee