Data transmission method for uplink based on the IBOC system of psychoacoustic model
Technical field
The present invention relates to wireless telecommunication system, relate in particular to the data transmission method for uplink of a kind of in-band on-channel system.
Background technology
Adopt fixing compound mode that the analog and digital signal frequency spectrum is carried out in-band on-channel (In-Band On-Channel, IBOC) transmission in the HD-Radio broadcasting mixed mode.Yet, owing to the spectral bandwidth of FM analog signal along with programme signal changes, the bandwidth excursion is large.Under very large time probability, the boundary bandwidth of analog fm signal bandwidth regulation analog and digital signal frequency spectrum in the mixed mode.In the case, the real-time change of analog signal bandwidth causes occurring a large amount of idle frequency spectrums.Simultaneously, because the shared spectral bandwidth of analog signal is in real-time change, and the digital signal frequency spectrum is placed on same spectrum position regularly, causes the audio quality in the different time sections that people's ear hears to alter a great deal.
Fig. 1 is the frequency spectrum of HD Radio mixed mode, digital signal is placed on the both sides transmission of simulation FM signal, only use 10 frequency spectrum sub-blocks of each main sideband middle distance central subcarrier distal-most end, and distal-most end is numbered ± 546 reference subcarrier, be called the PM frequency band, altogether comprise 382 subcarriers, the band occupancy scope from-198.402~-129.361KHz and 129.361~198.402KHz.In the frequency spectrum-scope of 129.361~129.361KHz keeps to analog signal, can be monophonic signal or stereophonic signal, also may comprise the subsidiary communications authenticated channel.
In the HD Radio system, transmitting terminal is modulated respectively and is generated analog signal and digital signal, finishes analog and digital signal by synthesizer synthetic, system model such as Fig. 2.At this moment, analog signal and digital signal adopt fixing spectrum mode to make up, and namely digital signal is placed on apart from the about position of 130KHz of carrier wave.
The psychoacoustic model that this patent uses is based on the psychoacoustic model of PEAQ algorithm, as shown in Figure 3.The PEAQ algorithm is by the auditory system of imitation people ear, reference signal and test signal are analyzed objective difference grade (the Objective Difference Grade that draws corresponding to audio quality, ODG), this definition is equal to the SDG in the subjective assessment.
After reference signal and test signal were processed through psychoacoustic model respectively, output separately just can calculate a series of model output parameters (Model Output Variables, MOV) via the sensor model characteristic synthetic.At last, be these MOV Parameter Mappings an objective difference grade output by neural network module.
Psychoacoustic model can convert the time-domain signal of audio frequency to basilar memebrane and represent.Basilar memebrane is positioned at cochlea, and the different frequency composition of sound can excite the excitement of its diverse location.By hair cell this excitement is converted into physiological stimulation again, reaches brain by auditory nerve.The concrete computational process of psychoacoustic model is: for the basic version that uses in this patent, audio signal is transformed into frequency domain by the FFT conversion, then simulate external ear and middle ear to the frequency response of sound by the spectral coefficient weighting, again it is mapped to physiology perception territory.
Sensor model is responsible for signal analysis and comprehensive, and purpose is better to simulate the sense quality of people's ear.
Neural net is responsible for MOV parameter that above two modules are calculated and the MOV parameter is mapped to an objective difference grade by neural net.
Yet, owing to the spectral bandwidth of FM analog signal along with programme signal (such as frequency, amplitude) changes, under very large time probability, the boundary bandwidth of analog fm signal bandwidth regulation analog and digital signal frequency spectrum in the standard.In the case, the real-time change of analog signal bandwidth not only causes a large amount of idle frequency spectrums occurring, and the interior audio quality of different time sections that causes people's ear to be heard alters a great deal.
Summary of the invention
In order to overcome the technical problem that exists in the prior art, the present invention increases digital signal and analog signal self adaptation adjusting module in system, will simulate FM signal and digital signal and join together to process.Frequency spectrum by real-time detection analog fm signal, the information of current analog signal is fed back to digital signal processing module, so that digital signal is carried out the auto-adaptive parameter adjustment, reach and improve the system transmissions ability, and make the different periods that the purpose of identical listening quality be arranged.
The sending method of the quasi-definite IBOC system transmitting terminal based on psychoacoustic model of the present invention as shown in Figure 4, may further comprise the steps:
The first step, to transmitting terminal simulated audio signal data intercept to be sent, intercepted length is the 4096*n point, is designated as a frame, and wherein n is positive integer, and this signal is reference signal;
Second step, by HD Radio modulation /demodulation software phantom, wherein the digital signal spectrum position is at 130KHZ-200KHZ with this reference signal, obtains the signal after the receiving terminal demodulation, and length is identical with reference signal, and this signal is test signal;
In the 3rd step, according to psychoacoustic model ITU-R BS.1387-1 version, calculate the masking threshold EhS of 109 subbands of present frame reference signal; And the sample noise EbN of 109 subbands of calculating present frame reference signal and test signal; And according to the EhS that calculates and EbN, calculate the NMR of current frame data, wherein NMR=EbN/EhS;
The 4th step, according to the NMR-Fstart rule-of-thumb relation, NMR according to previous step calculating, find corresponding digital signal spectrum position Fstart, carrier number N in the designation number signal OFDM modulation, wherein N=(200KHZ-Fstart)/Δ f rounds, and Δ f presses the regulation in the HDRadio standard, Δ f=363.4HZ, thus obtain according to the feature of analog signal and the digital signal after adjusting in real time;
The 5th step with the modulation of simulated audio signal up-conversion, obtained the FM pumping signal, with FM pumping signal and the digital signal coupling of adjusting in real time, coupled signal was sent.
The NMR variable is as one in 11 MOV variablees in the PEAQ model, the ratio relation between reflection noise and the signal masking threshold.Wherein said masking by noise than the computational methods of NMR is:
The first step is carried out respectively the time-domain windowed operation to the reference of a frame length and the test signal of input, then carries out N
FThe weighted factor according to each frequency of property calculation of external ear and middle ear filter is followed in some DFT conversion, afterwards the result of DFT conversion is carried out frequency domain weighting, and wherein the characteristic of external ear and middle ear filter is:
K is the spectral line count value;
Second step calculates signal difference, and wherein the computational methods of signal difference are:
0≤k≤N
F/ 2, wherein, X
Ref[k] and X
Test[k] is respectively the spectral line value after the DFT conversion of reference signal and test signal, and k is the spectral line count value;
The 3rd step, reference signal and the test signal of frequency domain are transformed into the Bark territory, the transformational relation in its frequency domain and Bark territory is: z=B (k)=7*asinh (k/650), and wherein, k is frequency domain spectral line count value, B (k) represents frequency domain to the transformational relation in Bark territory, z represents the Bark territory, and the data after the conversion are equally divided into 109 Bark territory subbands, finds boundary value corresponding to each Bark territory subband, again frequency domain is changed in the boundary value contravariant, the inverse transformation relational expression is: k=B
-1(z)=650*sinh (z/7), wherein B
-1(z) represent the Bark territory to the transformational relation of frequency domain, find boundary value corresponding to frequency domain, thereby frequency domain data is divided into 109 frequency domain subbands;
The 4th step is respectively with X
RefThe spectral line energy addition that comprises in [k] each frequency domain subband obtains the masking threshold Ehs[m of each frequency domain subband], wherein m is sub band number, m=1 ... 109;
The 5th step is respectively with X
DiffThe spectral line energy addition that comprises in [k] each frequency domain subband obtains the sample noise Ebn[m of each frequency domain subband], wherein m is sub band number, m=1 ... 109;
The 6th goes on foot, and masking threshold and the sample noise weighting of each frequency domain subband is average, calculates the NMR of whole frame data, and its computational methods are:
G[m wherein] be the weighting of masking threshold, expression formula is:
Wherein, the production method of NMR-Fstart rule-of-thumb relation as shown in Figure 5, concrete steps are:
The first step according to HD Radio mixed mode, will be simulated the coupling of FM signal and digital signal, the spectrum position stuck-at-1 30-200KHZ of digital signal wherein, the simulated audio signal after demodulating end obtains demodulation;
Second step according to the PEAQ model, calculates the overall noise masking ratio of simulated audio signal, be designated as NMR_ total, wherein transmitting terminal simulated audio signal to be sent is reference signal, is designated as Ref_ total, simulated audio signal after the demodulation is test signal, is designated as Test_ total;
The 3rd step, Test_ always intercepts respectively to the Ref_ summation, wherein the initial sampled point of i frame data is (4096*n-2048) * i, stopping sampled point is (4096*n) * (i+1)-2048*i, signal after the intercepting is respectively as reference signal and the test signal of i frame, be designated as Ref_i and Test_i, wherein n is positive integer;
The 4th step, according to the PEAQ model, respectively as reference signal and the test signal of i frame, calculate the masking by noise ratio of i frame audio signal with Ref_i and Test_i, be designated as NMR_i;
The 5th step, the end of digital signal frequency spectrum is fixed in 200KHZ remains unchanged, according to NMR_i and the total size of NMR_, adjust the original position p of digital signal frequency spectrum, wherein the initial value of p is 130KHZ; If NMR_i<NMR_ is total, 1KHZ then moves to left p, generate new digital signal, behind simulation FM signal coupling, obtain the simulated audio signal of current i frame at demodulating end, calculate the NMR_i of this moment, and again differentiate with the total size of NMR, until NMR_i〉NMR_ is total, and the minimum p value when record NMR_i<NMR_ is total is designated as Fstart_i; In like manner, if NMR_i〉NMR_ is total, the 1KHZ that then p moved to right, until NMR_i<NMR_ is total, the minimum p value when record NMR_i<NMR_ is total;
The 6th step, compile the original NMR_i value of every frame data and the Fstart_i value of determining at last, set up the relational expression between NMR and the Fstart, i.e. the NMR-Fstart curve.
Can be further understood by following detailed description and accompanying drawings about advantage of the present invention and method.
Description of drawings
Accompanying drawing described herein is used to provide a further understanding of the present invention, consists of the application's a part, and illustrative examples of the present invention and explanation thereof are used for explaining the present invention, do not consist of improper restriction of the present invention.In the accompanying drawings:
Fig. 1 is the spectrogram of mixed mode (Hybrid);
Fig. 2 is the HD Radio system diagram of the U.S.;
Fig. 3 is psychoacoustic model PEAQ algorithm block diagram;
Fig. 4 be the present invention draft HD Radio system diagram;
Fig. 5 is the production method figure of NMR-Fstart rule-of-thumb relation.
Embodiment
The below is to elaborating preferred embodiment of the present invention by reference to the accompanying drawings, thereby so that advantages and features of the invention can be easier to be it will be appreciated by those skilled in the art that protection scope of the present invention made more explicit defining.
Fig. 4 shows the transmission block diagram according to the IBOC system transmitting terminal based on psychoacoustic model of the present invention.With reference to Fig. 4, may further comprise the steps based on the sending method of the IBOC system transmitting terminal of psychoacoustic model:
The first step, to transmitting terminal simulated audio signal data intercept to be sent, intercepted length is the 4096*n point, is designated as a frame, and wherein n is positive integer, and this signal is reference signal;
Second step, by HDRadio modulation /demodulation software phantom, wherein the digital signal spectrum position is at 130KHZ-200KHZ with this reference signal, obtains the signal after the receiving terminal demodulation, and length is identical with reference signal, and this signal is test signal;
In the 3rd step, according to psychoacoustic model ITU-R BS.1387-1 version, calculate the masking threshold EhS of 109 subbands of present frame reference signal; And the sample noise EbN of 109 subbands of calculating present frame reference signal and test signal; And according to the EhS that calculates and EbN, calculate the NMR of current frame data, wherein NMR=EbN/EhS;
The 4th step, according to the NMR-Fstart rule-of-thumb relation, NMR according to previous step calculating, find corresponding digital signal spectrum position Fstart, carrier number N in the designation number signal OFDM modulation, wherein N=(200KHZ-Fstart)/Δ f rounds, and Δ f presses the regulation in the HDRadio standard, Δ f=363.4HZ, thus obtain according to the feature of analog signal and the digital signal after adjusting in real time;
The 5th step with the modulation of simulated audio signal up-conversion, obtained the FM pumping signal, with FM pumping signal and the digital signal coupling of adjusting in real time, coupled signal was sent.
Wherein, described masking by noise may further comprise the steps than the computational methods of NMR:
The first step is carried out respectively the time-domain windowed operation to the reference of a frame length and the test signal of input, then carries out N
FThe weighted factor according to each frequency of property calculation of external ear and middle ear filter is followed in some DFT conversion, afterwards the result of DFT conversion is carried out frequency domain weighting, and wherein the characteristic of external ear and middle ear filter is:
K is the spectral line count value;
Second step calculates signal difference, and wherein the computational methods of signal difference are:
0≤k≤N
F/ 2, wherein, X
Ref[k] and X
Test[k] is respectively the spectral line value after the DFT conversion of reference signal and test signal, and k is the spectral line count value;
The 3rd step, reference signal and the test signal of frequency domain are transformed into the Bark territory, the transformational relation in its frequency domain and Bark territory is: z=B (k)=7*asinh (k/650), and wherein, k is frequency domain spectral line count value, B (k) represents frequency domain to the transformational relation in Bark territory, z represents the Bark territory, and the data after the conversion are equally divided into 109 Bark territory subbands, finds boundary value corresponding to each Bark territory subband, again frequency domain is changed in the boundary value contravariant, the inverse transformation relational expression is: k=B
-1(z)=650*sinh (z/7), wherein B
-1(z) represent the Bark territory to the transformational relation of frequency domain, find boundary value corresponding to frequency domain, thereby frequency domain data is divided into 109 frequency domain subbands;
The 4th step is respectively with X
RefThe spectral line energy addition that comprises in [k] each frequency domain subband obtains the masking threshold Ehs[m of each frequency domain subband], wherein m is sub band number, m=1 ... 109;
The 5th step is respectively with X
DiffThe spectral line energy addition that comprises in [k] each frequency domain subband obtains the sample noise Ebn[m of each frequency domain subband], wherein m is sub band number, m=1 ... 109;
The 6th goes on foot, and masking threshold and the sample noise weighting of each frequency domain subband is average, calculates the NMR of whole frame data, and its computational methods are:
G[m wherein] be the weighting of masking threshold, expression formula is:
Wherein, the generation step of NMR-fstart rule-of-thumb relation is:
The first step according to HD Radio mixed mode, will be simulated the coupling of FM signal and digital signal, the spectrum position stuck-at-1 30-200KHZ of digital signal wherein, the simulated audio signal after demodulating end obtains demodulation;
Second step according to the PEAQ model, calculates the overall noise masking ratio of simulated audio signal, is designated as NMR total, and wherein transmitting terminal simulated audio signal to be sent is reference signal, is designated as Ref total, and the simulated audio signal after the demodulation is test signal, is designated as Test total;
The 3rd step, Test always intercepts respectively to the Ref summation, wherein the initial sampled point of i frame data is (4096*n-2048) * i, stopping sampled point is (4096*n) * (i+1)-2048*i, signal after the intercepting is respectively as reference signal and the test signal of i frame, be designated as Ref_i and Test_i, wherein n is positive integer;
The 4th step, according to the PEAQ model, respectively as reference signal and the test signal of i frame, calculate the masking by noise ratio of i frame audio signal with Ref_i and Test_i, be designated as NMR_i;
The 5th step, the end of digital signal frequency spectrum is fixed in 200KHZ remains unchanged, according to NMR_i and the total size of NMR_, adjust the original position p of digital signal frequency spectrum, wherein the initial value of p is 130KHZ; If NMR_i<NMR_ is total, 1KHZ then moves to left p, generate new digital signal, behind simulation FM signal coupling, obtain the simulated audio signal of current i frame at demodulating end, calculate the NMR_i of this moment, and again differentiate with the total size of NMR, until NMR_i〉NMR_ is total, and the minimum p value when record NMR_i<NMR_ is total is designated as Fstart_i; In like manner, if NMR_i〉NMR_ is total, the 1KHZ that then p moved to right, until NMR_i<NMR_ is total, the minimum p value when record NMR_i<NMR_ is total;
The 6th step, compile the original NMR_i value of every frame data and the Fstart_i value of determining at last, set up the relational expression between NMR and the Fstart, i.e. the NMR-Fstart curve.
The above; it only is one of the specific embodiment of the present invention; but protection scope of the present invention is not limited to this; any those of ordinary skill in the art are in the disclosed technical scope of the present invention; variation or the replacement that can expect without creative work all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range that claims were limited.