Method and apparatus for obtaining an attenuation factor

(54) Title: METHOD AND APPARATUS TO OBTAIN ONE (57) Summary: ATTENUATION FACTOR (30) Unionist Priority: 11/05/2007 cn 200710169618.0 (73) Holder (s): Huawei Technologies CO., LTD.

(72) Inventor (s): Chen Hu, Dongqi Wang, Fengyan Qi, Jianfeng Xu, Jing Wang, Lei Miao, Qing Zhang, Wuzhou Zhan, Yi Yang, Yongfeng Tu, Zhengzhong Du (74) Attorney (s): Dannemann, Siemsen, Bigler & Ipanema Moreira (86) International Order: pct CN2008070807 of 25/04/2008 (87) International Publication: wo 2009 / 059497de 14/05/2009 slOl

A signal changing trend is obtained s! 02

An attenuation factor is obtained according to the trend of change

Descriptive Report of the Invention Patent for METHOD AND APPARATUS TO OBTAIN AN ATTENUATION FACTOR.

TECHNOLOGY FIELD

The present invention relates to the field of signal processing, and particularly to a method and apparatus for obtaining an attenuation factor.

BACKGROUND

Voice data transmission is required to be real-time and reliable in a real-time voice communication system, for example, a Vo IP (Voice over IP) system. Due to the unreliable characteristics of a network system, a data packet may be lost or not reach its destination in time in a procedure for transmitting from a sending end to a receiving end. These two types of situations are considered to be loss of network packet by the receiving end. Network packet loss is inevitable. However, the loss of network packet is one of the most important factors that influence the speech quality of the voice. Therefore, a resistant packet loss cancellation method is necessary to recover the lost data packet in the real-time communication system so that good speech quality is still obtained under the network packet loss situation.

In existing real-time voice communication technology, at the sending end, an encoder divides a broadband voice into a high sub-band and a low sub-band, and uses ADPCM (Adaptive Modulation by Differential Pulse Codes) to encode the two subbands respectively and send them together to the receiving end over the network. At the receiving end, the two subbands are decoded respectively by the ADPCM decoder, and then the final signal is synthesized using a QMF synthesis filter (Mirrored Quadrature Filter).

Different Packet Loss Cancellation (PLC) methods are adopted for two different sub-bands. For a low bandwidth signal, kksob the situation without packet loss, a reconstruction signal is not changed during FADING-CROSSED. Under the packet loss situation, for the first lost frame, the history signal (the history signal is a voice signal before the lost frame in this order document) is analyzed using a short-term indicator and an a long term, and voice classification information is extracted. The lost frame signal is reconstructed using an LPC (linear predictive encoding) based on the interval repetition method, the indicator and the classification information. The ADPCM status will also be updated synchronously until a good frame is found. In addition, not only the signal corresponding to the lost frame needs to be generated, but also a signal section adapted for FADING-CRUZADO needs to be generated. In this way, once a satisfactory frame is received, FADING-CROSSED is performed to process the satisfactory frame signal and the signal section. It is observed that this type of FADING-CROSS occurs only after the receiving end loses a frame and receives the first satisfactory frame.

During the process of carrying out the present invention, the inventor realized at least the following problems in the prior art: The energy of the synthesized signal is controlled using a static self-adaptive attenuation factor in the prior art. Although the defined attenuation factor changes gradually, its attenuation speed, that is, the attenuation factor value, is the same with respect to the same voice classification. However, human voices are varied. If the attenuation factor does not match the characteristic of human voices, an uncomfortable noise will occur in the reconstruction signal, particularly at the end of the fixed vowels. The static self-adaptive attenuation factor may not be adapted to the characteristic of different human voices.

The situation shown in figure 1 is taken as an example, where T_O is the interval period of the historical signal. The upper signal corresponds to an original signal, that is, a schematic diagram in wave form under the situation without packet loss. The lower signal with a dashed line is a signal synthesized according to the prior art. As can be seen from the figure, the synthesized signal does not maintain the same attenuation speed as the original signal. If there is often the same interval repetition, the synthesized signal will produce obvious musical noise so that the difference between the situation of the synthesized signal and the desired situation is large.

SUMMARY

One embodiment of the present invention provides a method and apparatus for obtaining an attenuation factor adapted to obtain a self-adaptive and dynamically adjustable attenuation factor used in synthetic signal processing.

One embodiment of the present invention provides a method for obtaining the attenuation factor adapted to process the synthesized signal in packet loss cancellation, including:

obtain a tendency to change a signal; and obtain an attenuation factor according to the signal's changing tendency.

An embodiment of the present invention also provides an apparatus for obtaining the attenuation factor, adapted to process the synthesized signal in packet loss cancellation, including:

a changing trend obtaining unit adapted to obtain a changing trend of a signal; and an attenuation factor obtaining unit adapted to obtain an attenuation factor according to the change trend obtained by the change trend obtaining unit.

One embodiment of the present invention also provides a method and an apparatus for obtaining an attenuation factor adapted to smoothly transition the historical data to the last received data.

To achieve the above objective, an embodiment of the invention provides a method for signal processing, adapted to process a synthesized signal in packet loss cancellation, including:

obtain a tendency to change a signal;

obtain an attenuation factor according to the signal's changing tendency; and obtaining a lost frame reconstructed after attenuation according to the attenuation factor.

One embodiment of the present invention also provides a signal processing apparatus, adapted to process a synthesized signal in packet loss cancellation, including the following units:

a changing trend obtaining unit adapted to obtain a changing trend of a signal;

an attenuation factor obtaining unit adapted to obtain an attenuation factor according to the change trend obtained by the change trend obtaining unit; and a lost frame reconstruction unit adapted to obtain a lost frame reconstructed after attenuation according to the attenuation factor.

One embodiment of the present invention also provides a speech decoder adapted to decode the speech signal, including a low band decoding unit, a high band decoding unit and a quadrature mirrored filter unit.

The low band decoding unit is adapted to decode a received low band decode signal, and to compensate for a lost low signal.

The high band decoding unit is adapted to decode a high band decode signal, and to compensate for a lost high band signal.

The quadrature mirrored filter unit is adapted to obtain a final output signal by synthesizing the low band decoding signal and the high band decoding signal.

The low band decode signal unit includes a low band decode signal subunit, an LPC based on the interval repeat subunit, and a cross fading subunit.

The low-band decoding subunit is adapted to decode a received low-band stream signal.

The LPC based on the interval repetition subunit is adapted to generate a synthesized signal corresponding to the lost frame.

The cross-fading subunit is adapted to cross-fade the signal processed by the low-band decoding subunit and synthesized signal corresponding to the lost frame generated by the LPC based on the interval repeat subunit.

The LPC based on the interval repeat subunit includes an analysis module and a signal processing module.

The analysis module is adapted to analyze a history signal and generate a reconstructed lost frame signal.

The signal processing module is adapted to obtain a tendency to change a signal, and to obtain an attenuation factor according to the tendency to change the signal, and to attenuate the reconstructed lost frame signal, and obtain a reconstructed lost frame after attenuation.

One embodiment of the present invention further provides a computer program product. The computer program product comprises computer program codes that allow a computer to perform steps in any of the methods to obtain a mitigation factor in canceling packet loss when computer program codes are executed by the computer.

One embodiment of the present invention further provides a computer accessible storage medium. The storage medium accessible by computer stores computer program codes which enable a computer to perform the steps in any of the methods for obtaining an attenuation factor in packet loss cancellation when the computer program codes are executed by the comput ador .

One embodiment of the present invention further provides a computer program product. The computer program product comprises computer program codes that allow a computer to perform steps in any of the methods for signal processing in canceling packet loss when computer program codes are executed by the computer.

One embodiment of the present invention further provides a computer accessible storage medium. The computer accessible storage medium stores computer program codes that allow a computer to perform steps in any of the methods for signal processing in packet loss cancellation when computer program codes are executed by the computer

Compared with the prior art, the embodiments of the present invention have the following advantages:

A self-adaptive attenuation factor is dynamically adjusted using the trend of changing a historical signal. The smooth transition of the historical data to the last received data is carried out so that the speed of attenuation between the compensated signal and the original signal is kept as constant as possible to adapt the characteristic of different human voices.

BRIEF DESCRIPTION OF THE DESIGN (S)

Figure 1 is a schematic diagram showing the original signal and the signal synthesized according to the prior art;

figure 2 is a flow chart illustrating a method for obtaining an attenuation factor according to Mode 1 of the present invention;

figure 3 is a schematic diagram illustrating the principles of the encoder;

Figure 4 is a schematic diagram illustrating the module of an LPC based on the interval repeat subunit of the low band decoding unit;

Figure 5 is a schematic diagram showing an output signal after adopting the dynamic attenuation method according to Mode 1 of the present invention;

figures 6A and 6B are schematic diagrams illustrating the structure of the apparatus for obtaining an attenuation factor according to Modality 2 of the present invention;

figure 7 is a schematic diagram illustrating the application scenario of the apparatus for obtaining an attenuation factor according to Mode 2 of the present invention;

figures 8A and 8B are schematic diagrams illustrating the structure of the apparatus for signal processing according to Modality 3 of the present invention;

figure 9 is a schematic diagram illustrating the speech decoder module according to Mode 4 of the present invention;

figure 10 is a schematic diagram illustrating the module of the low-band decoding unit in the voice decoder according to Mode 4 of the present invention;

Figure 11 is a schematic diagram illustrating the LPC module based on the interval repeat subunit according to Mode 4 of the present invention.

DETAILED DESCRIPTION

The present invention will be described in more detail with reference to the drawings and embodiments.

A method is provided for obtaining an attenuation factor in Mode 1 of the present invention, adapted to process the synthesized signal in packet loss cancellation, as shown in Figure 2, including the following steps.

Step s101, a trend for changing a signal is obtained;

Specifically, the trend of change can be expressed in the following parameters: (1) the ratio of the energy of the last periodic interval signal to the energy of the previous periodic signal in the signal; (2) the ratio of the difference between the maximum amplitude value and the minimum amplitude value of the last periodic interval signal to the difference between the maximum amplitude value and the minimum amplitude value of the previous interval periodic signal in the signal.

Step s102, an attenuation factor is obtained according to the trend of change.

The specific processing method of Mode 1 of the present invention will be described together with the specific application scenario.

A method for obtaining an attenuation factor that is adapted to process the synthesized signal in packet loss cancellation is provided in Mode 1 of the present invention.

As shown in figure 3, different PLC methods are adopted for two different sub-bands. The PLC method for the low band part is shown as part 1 in a dashed frame in figure 3. While a dashed frame 2 in figure 3 corresponds to the high band PLC algorithm. For a high band signal, zh (ri) is a high band signal that is finally produced. After obtaining the low band signal zl (ri) and the high band signal zh (n), QMF is executed for the low band signal and the high band signal and a broadband signal finally produced y (n) is synthesized.

Only the low band signal is described in detail below.

Under the situation without loss of frame, the signal ^xi(n), n = 0, ..., 7-1 θ obtained after decoding the current frame received by the low band ADPCM decoder, and the output is ^zi(n), n = 0, ..., Ll _corrSP_Oncurrent picture. In this situation, the reconstruction signal is not changed during FADING-CRUZADO, which is zl [ri \ = xl [n \, n = 0, ..., 7-1, where 7 is the frame length;

Under the situation with loss of frames, with respect to the first lost frame, the historical signal zl (n), n <0 is analyzed using a short-term indicator and a long-term indicator, and voice classification information is extracted . Adopting the above indicators and the classification information, the y / (n) sign is generated using a

LPC based on interval repetition. Perdido the lost frame signal is reconstructed as zl (n) = yl (n), n = 0, ···, 7-1. In addition, the status of ADPCM will also be updated synchronously until a satisfactory picture is obtained. It is observed that not only the signal corresponding to the lost frame needs to be generated, but also a signal of 10ms yl (n), n = £, ···, £ + Ml adapted for FADING-CRUZADO needs to be generated, the M is the number of signal sampling points that are included in the process when calculating the energy. In this way, once a satisfactory frame is received, FADING-CROSSED is performed for x / («), n = Ã, ---, Ã + M-1, and yl (n),« = Z, · ··, /, + Ml. It is observed that this type of FADING-CROSSED occurs only after a loss of frame and when the receiving end receives the first satisfactory frame data.

An LPC based on the interval repetition method in figure 3 is as shown in figure 4.

When the data frame is a satisfactory frame, zl (n) is stored in a buffer for use in the future.

When the first lost frame is found, the final signal yl (n) needs to be synthesized in two steps. In the first, the historical signal zl (ri), n = -297, ---, - 1 is analyzed. Then, the sign yl (n), n = 0, ·, £ -! is synthesized according to the result of the analysis, where L is the frame length of the data frame, that is, the number of sampling points corresponding to a signal frame, Q is the length of the signal that is necessary to analyze the history sign.

The LPC module based on interval repetition specifically includes the following parts.

(1) An analysis of LP (Linear Prediction)

The short-term analysis filter A (z) and the synthesis filter MA (z) are Linear Prediction (LP) filters based on the P order. The LP analysis filter is defined as:

A (z) = l + a_t z ~ '+ a₂ z ~² -l ----- ha_P z ~^P

Through the LP analysis of the historical signal zl (ri), n = -Q, ---, - 1 with filter A (z), a residual signal and (n), n = -Q, ---, -1 corresponding to the historical signal z / (n), n = -Q, ---, - l is obtained:

P e (n) = zl (ri) + ^ a_ízl (n-í), n = (2) A historical signal analysis

The lost signal is compensated for by an interval repeat method. So first, a period of interval T_O corresponding to the historical sign zl (n), «= -Q, - -, - 1 needs to be estimated. The steps are as follows: zl (n) is pre-processed to remove a useless low-frequency ingredient in an LTP (long-term prediction) analysis, and the T interval period_Q z / (') can be obtained by LTP analysis. Voice classification is obtained despite combining a signal classification module after obtaining the T interval period_O.

The voice classifications are as shown in table 1 below:

Table 1 Voice classifications

<td>Classification Name</td><td>Explanation</td>

<td>TRANSIENT</td><td>for voices with great energy variation (for example, stops)</td>

<td>UNVOICED</td><td>for speechless signals</td>

<td>VUV_TRANSITiON</td><td>for a transition between voice signals and without voice</td>

<td>WEAKLY_VOICED</td><td>for weak voice signals (for example, vowels initial or final)</td>

<td>VOICED</td><td>voice signals (for example, fixed vowels)</td>

(3) One interval repetition

An interval repeat module is adapted to estimate a residual LP signal and (n), n = of a lost frame. Before interval repetition is performed, if the voice classification is not VOICED, the following formula is adopted to limit the amplitude of a sample:

e (laughs) = min (maxje (n -T_O + z ') |), | e («) | j xsign (e (n) \ n = —T₀, · -) where, f 1 if x&gt; 0 sign (x) = &lt; _n [- 1 if x &lt;0

If the voice classification is VOICED, oe (n), n = 0, ···, £ -!

residual signal corresponding to the lost signal is obtained by adopting a step of repeating the residual signal corresponding to the signal of the last interval period in the signal of a recently received satisfactory frame, which is

e (n) = e (n - T₀)

With respect to other voice classifications, to avoid that the periodicity of the generated signal is too intense (in relation to the signal without voice, if the periodicity is very intense, you may hear some uncomfortable noise such as a noise), the residual signal and (n), n = 0, ---, / - 1 corresponding to the lost signal is generated using the following formula:

and (n) = _Ç(«-R₀+ (- l) ”)

In addition to generating the residual signal corresponding to the lost frame, the residual signals e (n), n = Ã, - -, £ + N- from additional samples N continue to be generated to generate a signal adapted to FADING CROSSED, in order to guarantee the smooth combination of the lost frame and the first satisfactory frame after the lost frame.

(4) An LP analysis

After generating the residual signal and (n) corresponding to the lost frame and the FADING-CROSSED, a yl reconstruction lost frame signal_pre(n) 'n = 0, ---, / - 1 is obtained using the following formula:

yl_pre(n) = e (ri) - ^^ ylÇn - /) i '= l where the residual signal e (n), n = 0, ---, / - 1 is the residual signal obtained from the interval repeat steps above .

Also, yt_pre(n), n = /, ---, Z + N -1 with N samples adapted for FADING-CRUZADO are generated using the formula above.

(5) An adaptive silencing

To perform a smooth energy transition, before running QMF with the high band signal, the low band signal also needs to perform FADING-CROSSED, the rules are shown as the following table:

<td colspan="2" rowspan="2"></td><td colspan="2">Current framework</td>

<td>Unsatisfactory condition</td><td>Satisfactory picture</td>

<td>Frame</td><td>Frame</td><td>zl (ri) = yl (n),</td><td>zl (n) = —-— xl (n) + (1 ---—) &gt;&gt; / («) ' Nl Nl η = 0, ···, Ν-1</td>

<td>previous</td><td>unsatisfactory</td><td>n = 0, ···, £ - 1</td><td>and</td>

<td></td><td></td><td></td><td>zl (n) = xl (n), η = N, ---, /, - 1</td>

<td></td><td>Frame satisfactory</td><td>zl (n) = yl (n), n = 0, ·· -, L - 1</td><td>zl (n) = xl (n), n = 0, ·· -, L -1</td>

In the table above, zl (n) is a signal finally produced corresponding to the current frame; xl (n) is the sign of the satisfactory frame corresponding to the current frame; yl (n) is a synthesized signal corresponding to the current frame at the same time, where L is the frame length, N is the number of samples that perform FADING-CROSS.

With respect to different voice classifications, the signal energy in yl_pre(n) is controlled before executing FADING-CRUZADO according to the coefficient corresponding to each sample. The value of the coefficient changes according to different voice classifications and the situation of packet loss.

In detail, in the case where the last two periodic interval signals in the received history signal is the original signal as shown in figure 5, the self-adaptive dynamic attenuation factor is dynamically adjusted according to the changing trend of the last two interval periods. on the history sign. The detailed adjustment method includes the following steps:

Step s201, the signal change trend is obtained.

The trend of signal change can be expressed by the ratio of the energy of the last interval period signal to the energy of the previous interval period signal in the signal, that is, the energy Ei and S₂of the last two signs of interval periods in the history signal, and the ratio of the two energies is calculated.

T,_t

E, = Vx /²(-z)

Ε₂ = Σχ1²(-ί-Τ₀)

<img file="BRPI0808765A2_D0001.tif" />

\ Ε₂

Ε_χ is the energy of the last interval period signal, E₂ is the energy of the previous interval period signal, and T_Q is the interval period corresponding to the history signal.

Optionally, the sign change trend can be expressed by the reason for the peak-valley differences of the last two interval periods in the history signal.

P_y = max (x / (z)) - min (x / (y ')) (i, f) = -T_O ,...,-1

P₂ = max (x / (z)) - min (x / (y)) (z, j) = -2T₀ , ..., - (T₀ +1) where, Ρ_λ is the difference between the maximum amplitude value and the minimum amplitude value of the last periodic interval signal, P₂ is the difference between the maximum amplitude value and the minimum amplitude value of the periodic signal from the previous interval, and the ratio is calculated as:

r = A

P.

In step s202, the synthesized signal is dynamically attenuated according to the change trend obtained from the signal.

The calculation formula is shown as follows: yl (n) = yl_pre(«) * (1 - C * (zz +1)) n = 0, .., N -1 where, yl_pre(ri) is the lost frame reconstruction signal, N is the length of the synthesized signal, and C is the self-adaptive attenuation coefficient whose value is:

T ^JO

Under the situation of the attenuation factor lC * («+ l) &lt;0, it is necessary to adjust lC * (n + l) = 0, to avoid showing a situation where the attenuation factor corresponding to the samples is negative.

In particular, to avoid the situation where the amplitude value corresponding to a sample is exceeded under the situation of R&gt; 1, the synthesized signal is dynamically attenuated using the formula of step s202 in the present modality that takes into account only the situation of R &lt;1.

In particular, to avoid the situation where the attenuation speed of the signal with less energy is very fast, only under the situation where £ exceeds a certain limit value, the synthesized signal is dynamically attenuated using the formula of step s202 in the present mode.

In particular, to prevent the attenuation speed of the synthesized signal from being too fast, especially under the situation of continuous loss of frame, an upper limit value is adjusted for the attenuation coefficient C. When C * («+ l) exceeds a limit value, the attenuation coefficient is set as the upper limit value.

In particular, under the situation of a bad network environment and continuous loss of frame, a certain condition can be established to avoid a very fast attenuation speed. For example, one can take into account that when the number of frames lost exceeds a fixed number, for example, two frames; or when the signal corresponding to the lost frame exceeds a fixed length, for example, 20ms; or in at least one of the conditions above the current attenuation coefficient 1 ~ C * («+1) reaches a fixed threshold value, the attenuation coefficient C needs to be adjusted to avoid the very fast attenuation speed that can result in the situation where the output signal is muted.

For example, under situation sampling at a frequency of 8k and the frame length of 40 samples, the number of frames lost can be adjusted to 4, and after the attenuation factor ^1_C * (n + l) t_Omanif less than 0.9, the attenuation coefficient C is adjusted to be a lower value. The rule of adjusting the smallest value is as follows.

Hypothetically, it is predicted that the current attenuation coefficient is C and the attenuation factor value is V, and the attenuation factor V can be attenuated to 0 after the V / C samples. While a more desired situation is one where the factor attenuation V should be attenuated to 0 after M (M / C) samples. Then, the attenuation coefficient C is adjusted to:

As shown in figure 5, the upper signal is the original signal; the intermediate signal is the synthesized signal. As seen in the figure, although the signal has some degree of attenuation, the signal still remains sound intensive. If the duration is too long, the signal can be shown as a musical noise, especially at the end of the sound. The lower signal is the signal after using dynamic attenuation in the mode of the present invention, which can be very similar to the original signal.

According to the method provided by the modality mentioned above, the self-adaptive attenuation factor is dynamically adjusted using the trend of changing the historical signal, so that the smooth transition from the historical data to the last received data can be performed. The attenuation speed is kept as constant as possible between the compensated signal and the original signal to adapt the characteristic of varied human voices.

An apparatus for obtaining an attenuation factor is provided in Modality 2 of the present invention, adapted to process the synthesized signal in packet loss cancellation, including:

a changing trend obtaining unit 10, adapted to obtain a changing trend of a signal;

an attenuation factor 20 unit, adapted to obtain an attenuation factor according to the change trend obtained by the change trend unit 10.

The attenuation factor obtaining unit 20 additionally includes: an attenuation coefficient obtaining unit 21, adapted to generate the attenuation coefficient according to the change trend obtained by the obtaining change trend unit 10; an attenuation factor 22 subunit, adapted to obtain an attenuation factor according to the attenuation coefficient generated by the attenuation factor 21 subunit. The attenuation factor 20 unit also includes: an attenuation coefficient adjustment subunit 23, adapted to adjust the attenuation coefficient value obtained by the attenuation coefficient subunit to a certain value under certain conditions that include at least least one of the following characteristics: if the value of the attenuation coefficient exceeds an upper limit value; if there is a situation of continuous loss of staff; and if the attenuation speed is too fast.

The method for obtaining an attenuation factor in the above modality is the same method for obtaining an attenuation factor in the method modalities.

In detail, the changing trend obtained by the changing trend obtaining unit 10 can be expressed in the following parameters: (1) the ratio of the energy of the last periodic interval signal to the energy of the previous periodic signal in the signal; (2) the ratio of the difference between the maximum amplitude value and the minimum amplitude value of the last periodic interval signal to the difference between the maximum amplitude value and the minimum amplitude value of the previous interval periodic signal in the signal.

When the trend of change is expressed in the energy ratio in (1), the structure of the device to obtain an attenuation factor is as shown in figure 6A. The unit for obtaining a trend of change 10 also includes:

an energy acquisition subunit 11 adapted to obtain the energy of the last periodic interval signal and the energy of the previous periodic signal;

an energy ratio obtaining subunit 12 adapted to obtain the energy ratio of the last periodic interval signal to the energy of the previous interval periodic signal obtained by the energy acquisition subunit 11 and uses the reason to show the changing trend of the signal.

When the trend of change is expressed in the amplitude difference ratio in (2), the structure of the device to obtain an attenuation factor is as shown in figure 6B. The unit for obtaining a trend of change 10 also includes:

a subunit for obtaining amplitude difference 13, a17 adapted to obtain the difference between the maximum amplitude value and the minimum amplitude value of the last periodic interval signal, and the difference between the maximum amplitude value and the minimum amplitude value the previous interval periodic signal;

a subunit for obtaining amplitude difference ratio 14, adapted to obtain the ratio of the difference between the maximum amplitude value and the minimum amplitude value of the last periodic interval signal for the difference between the maximum amplitude value and the value of minimum amplitude of the periodic signal of previous interval, and use the reason to show the trend of changing the signal.

A schematic diagram illustrating the application scenario of the device to obtain an attenuation factor according to the Modality 2 of the present invention is as shown in figure 7. The self-adaptive attenuation factor is dynamically adjusted using the historical signal changing trend .

Using the device provided by the aforementioned modality, the self-adaptive attenuation factor is dynamically adjusted using the trend of changing the historical signal so that the smooth transition of the historical data to the last received data is carried out. The attenuation speed is kept as constant as possible between the compensated signal and the original signal to adapt the characteristic of varied human voices.

An apparatus for signal processing in Mode 3 of the present invention is provided, adapted to process the synthesized signal in packet loss cancellation, as shown in figure 8A and figure 8B. Based on Modality 2, a lost frame reconstruction unit 30 related to an attenuation factor obtaining unit is added. The lost frame reconstruction unit 30 obtains a lost reconstructed frame after attenuation according to the attenuation factor obtained by the attenuation factor 20 unit.

Using the device provided by the aforementioned modality, the self-adaptive attenuation factor is dynamically adjusted using the trend of changing the historical signal, and a lost frame reconstructed after the attenuation is obtained according to the attenuation factor, so that a smooth transition from historical data to the last received data be carried out. The attenuation speed is kept as constant as possible between the compensated signal and the original signal to adapt to the characteristic of varied human voices.

A voice decoder by Mode 4 of the present invention is provided, as shown in figure 9. The speech decoder includes: a high-band decoding unit 40 is adapted to decode a received high-band decode signal and to compensate for a lost high-band signal; a low band decoding unit 50 is adapted to decode a received low band decode signal and to compensate for a lost low band signal; and a quadrature mirrored filter unit 60 is adapted to obtain a final output signal synthesizing the low band decode signal and the high band decode signal. The high band decoding unit 40 decodes the high band flow signal received by the receiving end, and synthesizes the lost high band signal. The low band decoding unit 50 decodes the low band flow signal received by the receiving end and synthesizes the lost low band signal. The mirrored quadrature filter unit 60 obtains the final decoding signal by synthesizing the low band decoding signal sent by the low band decoding unit 50 and the high band decoding signal sent by the high band decoding unit 40.

For the low-band decoding unit 50, as shown in figure 10, the following units are included. An LPC based on the interval repeat subunit 51 which is adapted to generate a synthesized signal corresponding to the lost frame, a low band decoding subunit 52 which is adapted to decode a received low band stream signal, and a cross-fading subunit 53 which is adapted to cross-fade the signal decoded by the low-band decoding subunit and the synthesized signal corresponding to the lost frame generated by the LPC based on the interval repeat subunit.

The low band decoding subunit 52 decodes the received low band flow signal. The LPC based on the interval repeat subunit 51 generates the synthesized signal by executing an LPC on the lost low band signal. And finally, the cross-fading subunit 53 cross-fades the signal processed by the low-band decoding subunit 52 and the synthesized signal to obtain a final decoding signal after lost frame compensation.

The LPC based on the interval repeat subunit 51, as shown in figure 10, further includes an analysis module 511 and a signal processing module 512. Analysis module 511 analyzes a history signal, and generates a reconstructed lost frame signal; the signal processing module 512 obtains a signal changing tendency, and obtains an attenuation factor according to the signal changing tendency, and attenuates the reconstructed lost frame signal, and obtains a reconstructed lost frame after attenuation .

The signal processing module 512 further includes an attenuation factor obtaining unit 5121 and a lost frame reconstruction unit 5122. The attenuation factor obtaining unit 5121 obtains a tendency to change a signal, and obtains an attenuation factor according to the tendency to change; the lost frame reconstruction unit 5122 attenuates the lost frame signal reconstructed according to the attenuation factor, and obtains a lost frame reconstructed after the attenuation. The signal processing module 512 includes two structures corresponding to the schematic diagrams which illustrate the structure of the apparatus for signal processing in figures 8A and 8B, respectively.

The attenuation factor obtaining unit 5121 includes two structures corresponding to the schematic diagrams that illustrate the structure of the apparatus to obtain an attenuation factor in figures 6A and 6B, respectively. The specific functions and means of implementing the above modules and units can refer to the content revealed in the method modalities. Unnecessary details will not be repeated here.

By describing the modalities mentioned above, those skilled in the art can clearly understand that the present invention can be realized depending on the software plus the general and necessary hardware platform, and certainly can also be accomplished by hardware. However, in most situations, the trainer is a preferable modality. Based on such an understanding, the essence or part that contributes to the prior art in the technical scheme of the present invention can be expressed in the form of a software product that is stored in a storage medium, and the software product includes some instructions to instruct a device to perform the modalities of the present invention.

Although the illustration and description of the present description are determined with reference to the modalities of this, it should be assessed by those skilled in the art that various changes in shapes and details can be made without abandoning the scope of the description.