WO2010108445A1 - 通道间延迟估计的方法及其装置和编码器 - Google Patents

通道间延迟估计的方法及其装置和编码器 Download PDF

Info

Publication number
WO2010108445A1
WO2010108445A1 PCT/CN2010/071314 CN2010071314W WO2010108445A1 WO 2010108445 A1 WO2010108445 A1 WO 2010108445A1 CN 2010071314 W CN2010071314 W CN 2010071314W WO 2010108445 A1 WO2010108445 A1 WO 2010108445A1
Authority
WO
WIPO (PCT)
Prior art keywords
correlation function
cross
sound field
field information
cumulative
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.)
Ceased
Application number
PCT/CN2010/071314
Other languages
English (en)
French (fr)
Inventor
吴文海
郎玥
苗磊
刘泽新
胡晨
哈维·米希尔·塔迪
张清
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.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
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 Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP10755445.3A priority Critical patent/EP2413598B1/en
Publication of WO2010108445A1 publication Critical patent/WO2010108445A1/zh
Anticipated expiration legal-status Critical
Priority to US13/245,303 priority patent/US8417473B2/en
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04S—STEREOPHONIC SYSTEMS 
    • H04S1/00—Two-channel systems
    • H04S1/007—Two-channel systems in which the audio signals are in digital form
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/008—Multichannel audio signal coding or decoding using interchannel correlation to reduce redundancy, e.g. joint-stereo, intensity-coding or matrixing
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
    • G10L25/06—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being correlation coefficients
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04S—STEREOPHONIC SYSTEMS 
    • H04S2400/00—Details of stereophonic systems covered by H04S but not provided for in its groups
    • H04S2400/09—Electronic reduction of distortion of stereophonic sound systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04S—STEREOPHONIC SYSTEMS 
    • H04S2420/00—Techniques used stereophonic systems covered by H04S but not provided for in its groups
    • H04S2420/03—Application of parametric coding in stereophonic audio systems

Definitions

  • the present invention relates to communication technologies, and in particular, to a method and apparatus for inter-channel delay estimation, and an encoder. Background technique
  • stereo technology With the development of computer technology, digital signal processing technology, and the development of high-definition television sound systems and home audio-visual systems, stereo technology has been greatly developed. Of course, this also proposes stereo technology, especially codec technology. Higher requirements.
  • the commonly used stereo encoding method is a parametric stereo encoding method.
  • the left and right channel signals are usually not directly encoded, but the left and right channel signals are down-mixed to obtain a downmix signal, and the downmix signal is obtained.
  • Encode Among them, some extra sideband information is added during encoding.
  • the stereo signal can be recovered by the downmix signal and the sideband information.
  • the evaluation of the quality of the stereo signal depends largely on the quality of the downmix signal. That is to say, at the encoding end, the more synchronized the signals of the left and right channels are, the less information will be lost during the downmixing process.
  • the sounding object will be compared with the two microphones recording the left and right channels.
  • There is a distance change or a distance difference which will inevitably result in a complete synchronization between the left and right signals, that is, there will be a certain delay between the left and right signals.
  • a delay estimation method is proposed to improve the quality of the stereo composite signal.
  • the method for estimating the delay includes calculating a cumulative cross-correlation function for the left and right channel signals before the left and right channel signals generate the downmix signal, and using the time corresponding to the maximum value of the cumulative cross-correlation function as the left and right a delay between the channels, encoding the delay, and transmitting the encoded delay to the decoding end, so that the signal is synthesized according to the delay at the decoding end, thereby maintaining the left and right
  • the sound field of the channel signal is stable.
  • the cumulative cross-correlation function is usually used as the basis for the decision. For the square ⁇ ⁇ see convention, when the left channel is ahead of the right channel, the delay is positive, and the delay is negative.
  • the inventors of the present invention have found that in the prior implementation, when the sound field of the left and right channel signals changes, the sound field cannot be well tracked. The change causes the delay between the left and right channels to be incorrectly estimated, resulting in the instability of the synthesized stereo, which degrades the quality of the stereo coding and affects the auditory effect.
  • Embodiments of the present invention provide a method and device for estimating channel delay, so as to correctly estimate the delay between left and right channel signals, and improve the stability of the synthesized stereo sound field.
  • the present invention provides a method for estimating the inter-channel delay, which includes:
  • an embodiment of the present invention provides an apparatus for estimating an inter-channel delay, including: an extracting unit, configured to respectively acquire sound field information of a signal from the cross-correlation function and the cumulative cross-correlation function of a left-right channel synthesized signal;
  • An adjusting unit configured to obtain a cumulative cross-correlation function according to the sound field information respectively obtained by the extracting unit And adjusting the cumulative cross-correlation function by using the adjustment information to obtain an adjusted cumulative cross-correlation function;
  • a delay estimating unit configured to determine that the time corresponding to the maximum value in the cumulative cross-correlation function after the adjustment unit is adjusted is an inter-channel delay.
  • an embodiment of the present invention further provides an encoder, including: an apparatus for estimating delay between channels, and an encoding apparatus, where
  • the apparatus for estimating the inter-channel delay is configured to obtain sound field information of the signal from the cross-correlation function and the cumulative cross-correlation function of the left and right channel synthesized signals, and obtain cumulative cross-correlation function adjustment information according to the separately obtained sound field information, Adjusting the cumulative cross-correlation function by using the adjustment information to obtain an adjusted cumulative cross-correlation function, determining that the time corresponding to the maximum value in the adjusted cumulative cross-correlation function is an inter-channel delay, and the channel is Inter-delay output to the encoding device;
  • the encoding device is configured to encode the received inter-channel delay and transmit the encoded inter-channel delay.
  • the embodiment of the present invention provides a cross-correlation function and a cumulative cross-correlation function between left and right channel signals, and adjusts a cumulative cross-correlation function by extracting sound field information of the signal from the cross-correlation function;
  • the time corresponding to the maximum value in the cumulative cross-correlation function is the estimated delay. That is to say, when the sound field of the left and right channel signals changes, the information of the sound field changes is extracted to correctly estimate the delay between the left and right channel signals, so that the opposite end correctly synthesizes the signals according to the received delay, thereby Improve the stability of the sound field of the synthesized stereo.
  • FIG. 1 is a flowchart of a method for estimating an inter-channel delay according to an embodiment of the present invention
  • FIG. 2 is a flowchart of a method for estimating inter-channel delay in Embodiment 1 of the present invention
  • FIG. 3 is a flowchart of adjusting cumulative cross-correlation function using sound field information according to Embodiment 1 of the present invention
  • FIG. 4 is another embodiment of Embodiment 1 of the present invention
  • FIG. 5 is a flowchart of another method for adjusting a cumulative cross-correlation function by using sound field information according to Embodiment 1 of the present invention.
  • FIG. 6 is a flow chart showing an application example of determining a sound field category according to a ratio in Embodiment 1 of the present invention
  • FIG. 7 is a flowchart of a method for estimating inter-channel delay according to Embodiment 2 of the present invention
  • FIG. 8 is a flowchart of adjusting cumulative cross-correlation function by using sound field information according to Embodiment 2 of the present invention
  • FIG. 9 is another embodiment of Embodiment 2 of the present invention
  • FIG. 10 is a flow chart showing another method for adjusting a cumulative cross-correlation function using sound field information in Embodiment 1 of the present invention.
  • 11 is a flowchart of a method for estimating delay between channels in 3 according to an embodiment of the present invention.
  • FIG. 13 is a schematic structural diagram of an apparatus for estimating inter-channel delay according to an embodiment of the present invention
  • FIG. 14 is a schematic structural diagram of an encoder according to an embodiment of the present invention. detailed description
  • FIG. 1 is a flowchart of a method for estimating inter-channel delay according to an embodiment of the present invention. the method includes:
  • Step 101 Determine a cross-correlation function and a cumulative cross-correlation function of the left and right channel signals; this step is an optional step in this embodiment. among them,
  • d is the delay, is a constant, n is the number of samples, is a variable; r is the signal of the right channel; 1 is the signal of the left channel;
  • a _ccf ⁇ d a _ ccf(d) * + ccf(d) a ⁇ 0
  • Step 102 Acquire sound field information of the signal from the cross-correlation function of the left and right channel signals and the cumulative cross-correlation function;
  • the sound field information of the current frame cross-correlation function may be extracted from the cross-correlation function, and the sound field information of the cumulative cross-correlation function before the current frame cross-correlation function is extracted from the cumulative cross-correlation function; short-term mutuals may also be extracted from the cross-correlation function
  • the sound field information of the correlation function and the sound field information for extracting the long-term cumulative cross-correlation function from the cumulative cross-correlation function are not limited in this embodiment.
  • Step 103 Obtain cumulative cross-correlation function adjustment information according to the separately obtained sound field information, and adjust the cumulative cross-correlation function by using the adjustment information to obtain an adjusted cumulative cross-correlation function;
  • the weighting coefficient of the cumulative cross-correlation function may be determined according to the extracted different sound field information, and the cumulative cross-correlation function may be adjusted by using the weighting coefficient to obtain an adjusted cumulative cross-correlation function;
  • the basis of the weighting coefficient is obtained by multiplying the value corresponding to the extracted signal type to determine the weighting coefficient of the cumulative cross-correlation function; and the corresponding sound field may be determined according to different sound field information of the current frame cross-correlation function and the cumulative cross-correlation function. Determining whether the corresponding sound field categories are the same, and setting a weighting coefficient of the cumulative cross-correlation function according to the judgment result;
  • the weighting coefficient adjusts the cumulative cross-correlation function to obtain an adjusted cumulative cross-correlation function.
  • Step 104 Determine that a time corresponding to a maximum value in the adjusted cumulative cross-correlation function is an inter-channel delay.
  • the method further includes: determining whether the sound field information changes, and if yes, performing step 103; otherwise, ending the process. That is to say, in the embodiment, the inter-channel delay is estimated, the current frame cross-correlation function and the sound field information of the previous cumulative cross-correlation function are extracted first, and the weighted coefficients of the cumulative cross-correlation function are calculated according to the extracted sound field information, and modified.
  • the weighting factor is used to adjust the cumulative cross-correlation function to estimate the delay between the left and right channel signals when the sound field changes. That is, the present embodiment extracts the information of the sound field change, and adjusts the delay estimation between the left and right channels according to the change of the sound field information.
  • the method comprises: adaptively weighting the cumulative cross-correlation function of the delay estimation according to the change of the sound field information of the current frame cross-correlation function and the sound field information of the cumulative cross-correlation function, or synchronizing the sound field information according to the short-term cross-correlation function
  • the cumulative cross-correlation function of the delay estimation correlation function is adaptively weighted by the change of the sound field information of the cumulative cross-correlation function, or the cumulative cross-correlation function of the delay estimation correlation function is adaptively weighted according to the extracted signal type and the extracted sound field information. Adjustment, so as to correctly estimate the delay between the left and right channel signals, and send the delay, so that the receiving end correctly synthesizes the signals according to the received delay, thereby improving the stability of the synthesized stereo sound field.
  • FIG. 2 it is a flowchart of a method for estimating inter-channel delay according to Embodiment 1 of the present invention, where the method includes:
  • Step 201 Perform windowing processing on the two signals of the left and right channels, and output the signal processed by the windowing, which is an optional step;
  • Step 202 Find a cross-correlation function of the left and right channel signals after windowing processing; the specific process of finding the cross-correlation function is detailed in the above formula, and details are not described herein again.
  • Step 203 Find the cumulative cross-correlation function; the specific process of accumulating the cross-correlation function is detailed in the above formula, and details are not described herein again.
  • Step 204 Extract sound field information of a current frame cross-correlation function from the cross-correlation function; specifically:
  • the first partial delay correlation time current frame cross-correlation function is defined as a current frame cross-correlation function with a delay greater than or equal to 0; and the second partial delay timing cumulative cross-correlation function is defined as a cumulative cross-correlation function with a delay less than or equal to zero.
  • the operation is exemplified by division and subtraction, and the first sound field information value includes a first ratio or a first difference value, where the second sound field information value includes the following The second ratio or the second difference, but is not limited thereto.
  • a preferred way to extract the sound field information is to first determine the sum of the current frame cross-correlation functions with a delay greater than or equal to 0, and then determine the sum of the current frame cross-correlation functions with a delay less than or equal to 0, and finally find a delay greater than or equal to zero.
  • the sum of the current frame cross-correlation function is divided by the sum of the current frame cross-correlation function with a delay less than or equal to 0, and the obtained ratio is referred to as a first ratio, and the first ratio is the sound field information for extracting the current frame cross-correlation function.
  • Another way to extract the sound field information is to first determine the sum of the current frame cross-correlation functions with a delay greater than or equal to 0, and then determine the sum of the current frame cross-correlation functions with a delay less than or equal to 0, and finally find the current delay with a delay greater than or equal to 0.
  • the sum of the frame cross-correlation function is subtracted from the sum of the current frame cross-correlation function with a delay less than or equal to 0, and the obtained difference is referred to as a first difference, and the first difference is a sound field information value for extracting a cross-correlation function of the current frame. .
  • Step 205 After delaying the cumulative cross-correlation function by one frame or multiple frames (not limited in this embodiment), extracting the sound field information of the delayed cumulative cross-correlation function, for example, if the current frame is N frames, the cumulative cross-correlation function That is, the sound field information of the cumulative cross-correlation function of the past N-1 frame; one way of extracting is: first determining the sum of the cumulative cross-correlation functions with the delay greater than or equal to 0, and then determining the sum of the cumulative cross-correlation functions with the delay less than or equal to 0, Finally, the sum of the cumulative cross-correlation function with the delay greater than or equal to 0 is divided by the sum of the cumulative cross-correlation functions with the delay less than or equal to 0, and the obtained ratio is called the second ratio, and the second ratio is the extracted cumulative cross-correlation function.
  • Sound field information Another way to extract is: first determine the sum of the cumulative cross-correlation functions with a delay greater than or equal to 0, then determine the sum of the cumulative cross-correlation functions with a delay less than or equal to 0, and finally find the cumulative cross-correlation function with a delay greater than or equal to zero. And subtracting from the sum of the cumulative cross-correlation functions with a delay less than or equal to 0, the obtained difference is referred to as a second difference, and the second difference is the sound field information for extracting the cumulative cross-correlation function.
  • Step 206 Calculate a weighting coefficient of the cumulative cross-correlation function by using the sound field information extracted to the change;
  • Step 207 Adjust the cumulative cross-correlation function according to the weighting coefficient of the cumulative cross-correlation function; the specific adjustment process is to calculate the cumulative cross-correlation function by using the calculated weighting coefficient as the adjustment weighting coefficient, and the specific implementation process thereof See Figure 3 and Figure 4 for details.
  • Step 208 Search for the time corresponding to the maximum value in the cumulative cross-correlation function, and the time is the estimated delay.
  • Step 209 Determine whether the change of the delay is valid compared with the original delay. If yes, go to step 210; otherwise, go to step 211; the judgment is based on: comparing the determined delay with the original delay, if it is determined The delay is full of conditions, it is valid, otherwise it is invalid. Step 210: Output the delay; Step 211: Output the original delay.
  • the weighting coefficients of the different cumulative cross-correlation functions adjust the cumulative cross-correlation function according to the weighting coefficients to track the change of the upper sound field, thereby estimating a relatively accurate delay.
  • FIG. 3 is a flowchart of adjusting a cumulative cross-correlation function by using sound field information in Embodiment 1 of the present invention.
  • the current frame cross-correlation function is Cc /(").
  • Step 301 Find a delay of the current frame cross-correlation function with a delay greater than or equal to zero, which is equal to zero.
  • cur _ ratio Ccf ( ⁇ )
  • the ratio of the sum of the current frame cross-correlation functions of I 2 Ccf (n) ( cur- ratio ): ⁇ " .
  • cur_ratio can be limited to a certain range, such as ⁇ min,max>, where the values of min and max can be set empirically, or the value of min can be set to 0, and The value of max can be set to infinity, which is not limited in this embodiment.
  • the purpose of setting ⁇ min,max> is to prevent the cur-ratio from being too large or too small.
  • Step 302 Find a cumulative cross-correlation function whose delay is greater than or equal to zero and a cumulative cross-correlation function whose delay is less than or equal to zero.
  • Step 303 Calculate a weighting coefficient of the cumulative cross-correlation function according to the obtained cur-ratio and prev-ratio; one way is: obtaining a weighting coefficient of the cumulative cross-correlation function by the following formula, / ⁇ Limited to this.
  • Step 304 Perform weighting operation on the cumulative cross-correlation function by using the weighting coefficient to obtain a weighted cumulative cross-correlation function, that is, the weighted cross-correlation function can better track the change of the upper sound field.
  • This embodiment provides a form of the cumulative cross-correlation function, that is, the cumulative cross-correlation function of the inter-channel delay is the sum of the cumulative cross-correlation function multiplied by one weighting coefficient and the cross-correlation function of the current frame, but is not limited to the above.
  • the cumulative cross-correlation function specifically:
  • Step 401 Find a sum of a cross-correlation function of a current frame with a delay greater than or equal to zero and a delay of less than or equal to zero. The difference between the sum of the current frame cross-correlation functions, the difference being referred to as the first difference;
  • Step 402 Find a difference between a sum of cumulative cross-correlation functions with a delay greater than or equal to zero and a cumulative cross-correlation function with a delay less than or equal to zero, the difference being referred to as a second difference;
  • Step 403 Find a first difference and a second The absolute value of the difference between the differences, the force weight coefficient of the cumulative cross-correlation function is obtained;
  • Step 404 Perform weighting operation on the cumulative cross-correlation function by using the weighting coefficient to obtain weighting After the cumulative cross-correlation function.
  • 5 is a flowchart of adjusting the cumulative cross-correlation function by using sound field information in Embodiment 1 of the present invention.
  • the delay of the current frame cross-correlation function is greater than or equal to zero and the delay is less than or equal to zero.
  • determining the class of the sound field according to the ratio of the two that is, determining whether the sound field categories corresponding to the ratios of the two are the same, and setting the weighting coefficient of the cumulative cross-correlation function according to the judgment result; adjusting the accumulation by using the set weighting coefficient Cross-correlation function.
  • the current frame cross-correlation function is still /(") - ⁇ n ⁇ T, T>0 ⁇ for the cumulative cross-correlation function before the current frame is still " c - Cc ("), - T ⁇ n ⁇ T, T>0 ⁇ .
  • the cross-correlation function is a cross-correlation function including normalization, but is not limited thereto.
  • the flag of the corresponding sound field category is 1; otherwise, it is determined whether the ratio cur_ratio is greater than or equal to the second threshold, and if so, the flag of the sound field category corresponding to the ratio cur_ratio is set to 0; otherwise, the flag of the sound field category corresponding to the ratio cur_rat io is set.
  • Step 503 Calculate the cumulative cross-correlation of the sum of the cumulative cross-correlation function with the delay greater than or equal to zero and the delay less than or equal to zero.
  • the ratio of the sum of the functions prev-ratio :
  • the ratio prev-ratio determines the sound field class corresponding to the cumulative cross-correlation function, and is marked with prev-flag; the process of determining and the process in step 503 Similarly, the method includes: determining whether the ratio prev-ratio is greater than a preset first threshold; if yes, setting a flag of the sound field category corresponding to the ratio prev-ratio to 1; otherwise, determining whether the ratio prev_ratio is greater than or equal to the second width Value, if yes, the flag of the sound field category corresponding to the ratio prev-ratio is 0; otherwise, the flag of the sound field category corresponding to the set value prev-ratio is -1, wherein the second threshold is less than the first threshold.
  • the specific implementation process is shown in Figure 5.
  • Step 505 Determine whether the sound field category corresponding to the ratio cur_ratio is the same as the sound field category corresponding to the ratio prev-ratio. If yes, perform steps 506 and 508; otherwise, perform steps 507 and 508; Step 506: Set the cumulative cross-correlation function The weighting coefficient of the method is 1; Step 507: The weighting coefficient of the cumulative cross-correlation function is set to be less than 1, and is generally set to 0.85, and may be other values less than 1. This embodiment is not limited.
  • Step 508 Adjust the cumulative cross-correlation function according to the set weighting coefficient. That is, the weighted coefficients are used to weight the cumulative cross-correlation function so that the weighted cumulative cross-correlation function can better track the change of the upper sound field.
  • This embodiment provides a cumulative cross-correlation function form but is not limited to this kind of cumulative cross-correlation function, that is, the cumulative cross-correlation function of the inter-channel delay is a cumulative cross-correlation function multiplied by a weighting coefficient and cross-correlation with the current frame. The sum of the functions,
  • the ratio rat io may be a ratio cur_rat io or a ratio prev-rat io, which is not limited in this embodiment.
  • this embodiment is similar to the implementation process of Embodiment 1, and the differences include: extracting short-term cross-correlation from a cross-correlation function
  • the sound field information of the function, and the sound field information of the long-term cumulative cross-correlation function are extracted from the cumulative cross-correlation function, and then the weighting coefficients of the cumulative cross-correlation function are calculated according to the extracted different sound field information.
  • the short-term mutual function and the long-term cumulative cross-correlation function are a relative concept, for example,
  • Step 701 Two left and right channels The road signals are separately windowed and output a windowed signal, which is an optional step;
  • Step 702 Find a cross-correlation function of the left and right channel signals after the windowing process
  • Step 703 Find the cumulative cross-correlation function; wherein the specific implementation process of Step 702 and Step 703 is detailed in Embodiment 1, and no longer Praise
  • Step 704 Extract sound field information of the short-term cross-correlation function from the cross-correlation function; Includes:
  • the fourth sound field information value is obtained by operating on the sum of the cumulative delay cross-correlation function of the third partial delay time and the cumulative cross-correlation function of the fourth partial delay time.
  • the third partial delay time short-time cross-correlation function is defined as a short-term cross-correlation function with a delay greater than or equal to 0; and the fourth partial delay time long-time cumulative cross-correlation function is defined as a long-term accumulation with a delay less than or equal to zero.
  • Cross-correlation function is exemplified by division and subtraction, and the third sound field information value includes a third ratio value or a third difference value, where the fourth sound field information value includes the following fourth The ratio or the fourth difference, but is not limited to this.
  • a preferred method for extracting sound field information is to determine a ratio of a sum of short-term cross-correlation functions with a delay greater than or equal to 0 and a sum of short-term cross-correlation functions with a delay less than or equal to 0. For convenience of description, it is referred to as a third ratio.
  • the third ratio is a sound field information that extracts a short-term cross-correlation function.
  • Another way to extract the sound field information is to determine the sum of the short-term cross-correlation functions with a delay greater than or equal to 0, and the sum of the short-term cross-correlation functions with a delay less than or equal to 0, and the short-term cross-correlation with the delay greater than or equal to 0.
  • Step 705 After delaying the cumulative cross-correlation function by one frame or multiple frames (in this embodiment, delaying one frame as an example), extracting sound field information of the long-term cumulative cross-correlation function accumulated before the delayed short-term cross-correlation function;
  • One way of extracting is to determine the ratio of the sum of the long-term cumulative cross-correlation functions with a delay greater than or equal to 0 to the sum of the long-term cumulative cross-correlation functions with a delay less than or equal to zero.
  • the fourth ratio is a sound field information for extracting a long-term cumulative cross-correlation function.
  • Another way to extract is: Determine the sum of long-term cumulative cross-correlation functions with a delay greater than or equal to 0, And a sum of long-term cumulative cross-correlation functions with a delay less than or equal to 0; subtracting the sum of the long-term cumulative cross-correlation functions with a delay greater than or equal to 0 and the sum of long-term cumulative cross-correlation functions with a delay less than or equal to 0, resulting in a difference It is called the fourth difference, and the fourth difference is the sound field information for extracting the long-term cumulative cross-correlation function.
  • Step 706 Calculate weighting coefficients of the cumulative cross-correlation function by using the extracted sound field information;
  • Step 707 Adjust the cumulative cross-correlation function according to the weighting coefficient of the cumulative cross-correlation function; the specific process includes: The specific implementation process is shown in FIG. 8, FIG. 9, and FIG. Step 708: Search for a time corresponding to a maximum value in the cumulative cross-correlation function, where the time is an estimated delay;
  • Step 709 Determine whether the change of the delay is valid with the original extended comparison. If yes, execute step 710: Otherwise, return to step 711. The basis for the judgment is: Compare the determined delay with the original extension, if the determined delay meets the condition, it is valid, otherwise it is invalid. Step 710: Output the delay; Step 711: Output the original delay.
  • FIG. 8 is a flowchart of adjusting a cumulative cross-correlation function by using sound field information according to Embodiment 2 of the present invention.
  • a sum of delays greater than or equal to zero in a short-term cross-correlation function is obtained.
  • determining, by the ratio of the two the type of the sound field is determined according to the ratio of the delay, that is, determining whether the sound field categories corresponding to the ratios of the two are the same, and setting the weighting coefficient of the cumulative cross-correlation function according to the determination result;
  • the weighting factor adjusts the cumulative cross-correlation function.
  • the short-term cross-correlation function takes "- cc / 2 ( -T ⁇ d ⁇ T, T>o
  • the cumulative cross-correlation function to a cumulative length of the cross-correlation function - cc f l (d), ⁇ T ⁇ d ⁇ T, T> 0 ⁇ embodiment; comprises:
  • Step 801 Find a ratio (cur_ratio) of the sum of the short-term cross-correlation functions with a delay greater than or equal to zero and the sum of the short-term cross-correlation functions with a delay less than or equal to zero.
  • the ratio is referred to as a third ratio, specifically com.
  • Step 803 finding the long-term cumulative cross-correlation function with a delay greater than or equal to zero And the ratio of the sum of the long-term cumulative cross-correlation functions with a delay less than or equal to zero (pr ⁇ Y_ratio):
  • the ratio prev ratio V a ccf ⁇ (d)l V a ccf2(d)
  • the ratio prev-ratio determines the sound field class corresponding to the cumulative cross-correlation function, J, marked with prev-flag; its determination process and
  • the process in step 703 is similar, and specifically includes: determining whether the fourth ratio is greater than a preset first threshold; if yes, setting a flag of the sound field category corresponding to the fourth ratio to 1; otherwise continuing to determine whether the fourth ratio is greater than Is equal to the second threshold, if yes, the flag of the sound field category corresponding to the fourth ratio is set to 0; otherwise, the flag of the sound field category corresponding to the fourth ratio is set to -1, wherein the fourth threshold is less than the third value.
  • Step 805 Determine whether the sound field category corresponding to the ratio cur_ratio is the same as the sound field category corresponding to the ratio prev-ratio. If yes, perform steps 806 and 808; otherwise, perform steps 807 and 808; Step 806: Set the cumulative cross-correlation function The weighting coefficient of the method is 1; Step 807: The weighting coefficient of the cumulative cross-correlation function is set to be less than 1, and is generally set to 0.85, and may be other values less than 1. This embodiment is not limited. Step 808: Adjust the cumulative cross-correlation function according to the set weighting coefficient.
  • the two signals of the left and right channels are respectively windowed, and the cross-correlation function between the two signals is obtained; the sound field information of the short-term cross-correlation function is extracted from the cross-correlation function and the past N - 1 frame accumulates the sound field information of the long-term cumulative cross-correlation function, adjusts the weighting coefficients of the relevant cumulative cross-correlation function by extracting different sound field information; searches for the value of the largest cumulative cross-correlation function from the cumulative cross-correlation function, and seeks a corresponding time of the maximum value in the cumulative cross-correlation function, the time being an estimated delay; then determining whether the delay is a valid delay, and if so, outputting the delayed output, so that the receiving end is based on the received delay Synthesize the signal correctly to improve the stability of the synthesized stereo sound field.
  • FIG. 9 is another flowchart of adjusting cumulative cross-correlation function using sound field information in Embodiment 2 of the present invention, wherein, in this embodiment, the short-term cross-correlation function is " -cc / 2 (° -
  • a_ccf ⁇ d) ⁇ _ ⁇ / ⁇ ⁇ >0 is taken as an example; specifically: Step 901: Find a delay greater than or equal to zero
  • Cur_ratio The ratio of the sum of the short-term cross-correlation functions to the sum of the short-term cross-correlation functions with a delay of less than or equal to zero (cur_ratio), for convenience of description, is called the third ratio, specifically public.
  • the weighting factor but not limited to the formula, is:
  • Lk+ba is the weighting coefficient of the cumulative cross-correlation function, and k and b are constants.
  • Step 904 Perform weighting operation on the cumulative cross-correlation function by using the weighting coefficient to obtain a weighted cumulative cross-correlation function.
  • the form of its cumulative cross-correlation function is detailed above.
  • Step 1001 Find a sum of a short-term cross-correlation function with a delay greater than or equal to zero and a delay less than or equal to zero. The difference between the sum of the short-term cross-correlation functions, the difference is called the third difference;
  • Step 1002 Find the sum of the long-term cumulative cross-correlation function with a delay greater than or equal to zero and the long-term cumulative cross-correlation function with a delay less than or equal to zero The difference between and, the difference is called the fourth difference;
  • the weighting coefficient of the cumulative cross-correlation function is obtained; of course, the formula for calculating the weighting coefficient is not limited thereto, and may be calculated by other formulas.
  • Step 1004 Perform weighting operation on the cumulative cross-correlation function by using the weighting coefficient to obtain The weighted cumulative cross-correlation function.
  • Embodiment 3 is also referred to FIG. 11 , which is a flowchart of a method for estimating the delay between channels in 3 according to an embodiment of the present invention. The method includes:
  • Step 111 The two signals of the left and right channels are respectively windowed, and the signal processed by the window is output, and the step is an optional step;
  • Step 112 Find a cross-correlation function of the left and right channel signals after windowing processing;
  • Step 113 Find the cumulative cross-correlation function;
  • Step 112 and step 113 The specific implementation process of step 112 and step 113 is described in detail in Embodiment 1, and is not further described herein.
  • Step 114 Extracting a signal type from the cross-correlation function, and sound field information of a current frame or a short-term cross-correlation function.
  • the specific implementation process of extracting the sound field information of the current frame or the short-term cross-correlation function from the cross-correlation function is as described above, and details are not described herein again.
  • the process of extracting the signal type from the cross-correlation function includes: the signal type can be collected from the cross-correlation function.
  • the specific acquisition process is well known to those skilled in the art, and details are not described herein.
  • Step 115 After delaying the cumulative cross-correlation function by one or more frames (in this embodiment, delaying one frame as an example), extracting the cumulative cross-correlation function or the cumulative cross-correlation function before the delayed current frame cross-correlation function
  • the sound field information of the cumulative cross-correlation function is accumulated, and the specific implementation process is described above, and will not be described here.
  • Step 116 Calculate a weighting coefficient of the cumulative cross-correlation function by using the extracted sound field information
  • Step 117 Adjust a cumulative cross-correlation function according to the weighting coefficient of the cumulative cross-correlation function;
  • Step 118 Search for a time corresponding to a maximum value in the cumulative cross-correlation function, where the time is an estimate Delay;
  • Step 119 Determine whether the original delay of the change of the delay is valid, if yes, perform step 120: Otherwise, perform step 121; step 120: output the delay; step 121: output the original delay.
  • step 119 by extracting the signal type, and the sound field information of the current frame or the short-term cross-correlation function, and the cumulative cross-correlation function before the current frame cross-correlation function or the sound field information of the long-term cumulative cross-correlation function, Whether the sound field of the channel changes, and calculating weighting coefficients of different cumulative cross-correlation functions according to the changed sound field information, and adjusting the cumulative cross-correlation function according to the weighting coefficient, so as to track the change of the upper sound field, thereby estimating the comparison Accurate delay.
  • the embodiment of the present invention further provides an apparatus for estimating the inter-channel delay, which is shown in FIG. 13 and includes: an extracting unit 131, an adjusting unit 132, and a delay estimating unit 133, where the extracting The unit 131 is configured to obtain sound field information of the signal from the cross-correlation function and the cumulative cross-correlation function of the left and right channel signals respectively.
  • the adjusting unit 132 is configured to obtain a cumulative cross-correlation function according to the sound field information respectively acquired by the extracting unit. Adjusting the information, adjusting the cumulative cross-correlation function by using the adjustment information to obtain an adjusted cumulative cross-correlation function; the delay estimating unit 133 is configured to determine a maximum of the cumulative cross-correlation functions after the adjustment unit adjusts The time corresponding to the value is the inter-channel delay.
  • the extracting unit specifically includes: a first extracting unit and a second extracting unit, where the first extracting unit is configured to extract sound field information from a cross-correlation function of left and right channel signals; and the second extracting unit is configured to The sound field information is extracted from the cumulative cross-correlation function before the cross-correlation function extracted by the first extraction unit.
  • the first extracting unit includes: a first calculating unit and a first determining unit, where the first calculating unit is configured to calculate a sum of a cross-correlation function of a first partial delay time cross-correlation function and a second partial delay time; a first determining unit, configured to calculate, by the first calculating unit, a sum of a first partial delay time cross-correlation function and a second partial delay time cross-correlation function to obtain a first sound field information value;
  • the second extracting unit includes: a second calculating unit and a second determining unit, configured to calculate a sum of a sum of a first partial delay time cumulative cross-correlation function and a second partial delay time cumulative cross-correlation function
  • the second determining unit is configured to calculate, by the second calculating unit, a sum of a first partial delay time cumulative cross-correlation function and a second partial delay time cumulative cross-correlation function to obtain a second sound field information value.
  • the adjusting unit includes: a first coefficient calculating unit and a first adjusting unit, wherein the first coefficient calculating unit is configured to calculate a cumulative cross-correlation function according to the first sound field information value and the second sound field information value a weighting coefficient; the first adjusting unit, configured to adjust the cumulative cross-correlation function by using a weighting coefficient calculated by the first coefficient calculating unit, to obtain an adjusted cumulative cross-correlation function.
  • the adjustment unit includes: a first sound field category determining unit, a first determining unit, a first setting unit, and a second adjusting unit, wherein the first sound field category determining unit is configured to perform according to the first and the first Determining, by the first determining unit, the first sound field information value and the second sound field information value, the corresponding sound field type; the first determining unit, configured to determine the sound field corresponding to the first sound field information value and the second sound field information value Whether the categories are the same, and the result of the determination is sent; the first setting unit is configured to set a weighting coefficient different from the cumulative cross-correlation function according to the determination result sent by the first determining unit; and the second adjusting unit is configured to use the first setting The cumulative cross-correlation function is adjusted by different weighting coefficients of the unit settings to obtain an adjusted cumulative cross-correlation function.
  • the cross-correlation function of the first extraction unit from the left and right channel signals includes: a current frame cross-correlation function
  • the cumulative cross-correlation function before the cross-correlation function extracted by the second extraction unit from the first extraction unit includes: a cumulative cross-correlation function before the frame cross-correlation function
  • the cross-correlation function of the first extraction unit from the left and right channel signals includes: a short-term cross-correlation function
  • the cumulative cross-correlation function before the cross-correlation function extracted by the second extraction unit from the first extraction unit includes: a long-term cross-correlation function.
  • the first extracting unit further includes: a third extracting unit, configured to extract a signal type from a cross-correlation function of the left and right channel signals; the adjusting unit further includes :
  • the device may further include: a determining unit, configured to determine whether the sound field information changes, and send the determination result of the change to the adjusting unit.
  • the device for estimating the inter-channel delay may be integrated in the encoder, or may be integrated into the multi-user positioning device for communication, and may be integrated into the multi-source source determining device, which is not limited in the embodiment of the present invention.
  • the following cross-correlation functions respectively use the current frame cross-correlation function and the short-term cross-correlation function; the cumulative cross-correlation function respectively takes the cumulative cross-correlation function and the long-term cross-correlation function before the current frame cross-correlation function.
  • the extracting unit specifically includes: a current frame extracting unit and a cumulative extracting unit, where the current frame extracting unit is configured to cross-correlate signals from left and right channels Extracting sound field information of the current frame cross-correlation function in the function; the cumulative extracting unit, configured to accumulate cross-correlation function from the current frame cross-correlation function The number is extracted from the sound field information.
  • the current frame extracting unit includes: a first calculating unit and a first determining unit, where the first calculating unit is configured to calculate a sum of a current frame cross-correlation function of the first partial delay time and a current frame of the second partial delay time a sum of correlation functions, such as a sum of a sum of a current frame cross-correlation function with a delay greater than or equal to 0 and a current frame cross-correlation function with a delay less than or equal to 0; the first determining unit, configured to calculate a first portion of the first computing unit The sum of the current frame cross-correlation function at the delay time and the sum of the current frame cross-correlation function at the second-part delay time are obtained to obtain a first sound field information value, such as a sum of a current frame cross-correlation function for determining that the delay is greater than or equal to zero.
  • a first sound field information value such as a sum of a current frame cross-correlation function for determining that the delay is greater than
  • the ratio of the sum of the current frame cross-correlation functions with a delay less than or equal to 0 is referred to as a first ratio; or, the sum of the current frame cross-correlation function for determining the delay greater than or equal to 0 and the current frame cross-correlation function with a delay less than or equal to 0
  • the cumulative extraction unit includes: a second calculating unit and a second determining unit, wherein the second calculating unit is configured to calculate a sum of a first partial delay time cumulative cross-correlation function and a second partial delay time cumulative cross-correlation function And, for example, a sum of a cumulative cross-correlation function for calculating a delay greater than or equal to 0 and a cumulative cross-correlation function having a delay less than or equal to 0; the second determining unit configured to calculate a first partial delay time accumulation for the second computing unit The sum of the cross-correlation function is operated with the sum of the second partial delay time cumulative cross-correlation function to obtain a second sound field information value, such as a cumulative inter-correlation function for determining that the delay is greater than or equal to 0 and a cumulative mutual with a delay of less than or equal to zero.
  • a second sound field information value such as a cumulative inter-correlation function for determining that the delay is greater than or equal to 0 and a cumulative mutual
  • the ratio of the sum of the correlation functions is called the second ratio; or the difference between the sum of the cumulative cross-correlation function for determining the delay greater than or equal to 0 and the sum of the cumulative cross-correlation functions with the delay less than or equal to 0 is called the second difference. .
  • the adjustment unit includes: a coefficient calculation unit and a first adjustment unit, and/or a first sound field category determining unit, a first determining unit, a first setting unit, and a second adjusting unit, wherein the coefficient calculating unit is used by Calculating a weighting coefficient of the cumulative cross-correlation function according to the first sound field information value and the second sound field information value, such as calculating a cumulative cross-correlation function according to the first ratio and the second ratio determined by the first and second determining units a weighting coefficient; the first adjusting unit, configured to adjust the cumulative cross-correlation function by using a weighting coefficient calculated by the coefficient calculating unit, to obtain an adjusted cumulative cross-correlation function;
  • the first sound field category determining unit is configured to determine a corresponding sound field category according to the first sound field information value determined by the first and second determining units and the second sound field information value, for example, according to the first Determining, by the first determining unit, the first ratio and the second ratio, the corresponding sound field category; the
  • the short-time extracting unit includes: a third calculating unit and a third determining unit, wherein the third calculating unit is configured to calculate a third-part delay time short-time cross-correlation function and a short-term cross-correlation with the fourth part delay time
  • the sum of the functions such as the sum of the sum of the short-term cross-correlation functions for calculating the delay greater than or equal to 0 and the short-term cross-correlation function with the delay less than or equal to 0; the third determining unit for shortizing the delay time of the third portion
  • the sum of the time cross-correlation function and the sum of the fourth-part delay time short-term cross-correlation function operate to obtain a third sound field information value, such as a sum of a short-term cross-correlation function
  • the ratio of the sum of the short-term cross-correlation functions is called the third ratio; or the difference between the sum of the short-term cross-correlation functions for determining the delay greater than or equal to 0 and the sum of the short-term cross-correlation functions with the delay less than or equal to 0,
  • the third time difference; the long time accumulation extraction unit includes: a fourth calculation unit and a fourth determination unit, the fourth meter An arithmetic unit, configured to calculate a sum of a cumulative cross-correlation function of a third-part delay time cumulative time-correlation function and a fourth-part delay time-time cumulative cross-correlation function, such as a long-term cumulative cross-correlation function for calculating a delay greater than or equal to zero And a sum of a long-term cumulative cross-correlation function with a delay less than or equal to 0; the fourth determining unit, configured to accumulate cross-correlation with the long-term cumulative cross-correlation function of the third-part delay time and the fourth-
  • the adjustment unit includes: a second sound field category determining unit, a second determining unit, a second setting unit, and a third adjusting unit, wherein the second sound field category determining unit is configured to determine according to the third and fourth And determining, by the unit, the third sound field information value and the fourth sound field information value, corresponding to the sound field category, for example, determining the sound field category corresponding to the third ratio determined by the third and fourth determining units and the fourth ratio;
  • the second determining unit is configured to determine whether the sound field category corresponding to the third sound field information value and the fourth sound field information value are the same, and send a determination result, such as a sound field corresponding to the third ratio and the fourth ratio.
  • the embodiment of the present invention further provides an encoder 14, which is shown in FIG. 14 and includes: an inter-channel delay estimation device 141 and an encoding device 142, and the inter-channel delay estimation device 141 is configured to determine left and right channels.
  • the encoding device is configured to encode the received inter-channel delay and transmit the encoded inter-channel delay.
  • the function of each unit in the apparatus for estimating the delay between channels For the implementation process, refer to the corresponding implementation process in the above method, and details are not described herein again. It can be seen from the above embodiment that when the sound field of the left and right channel signals changes, the information of the sound field change can be extracted from the cross correlation function and the cumulative cross correlation function between the left and right channel signals, and the information of the sound field change can be correctly estimated. The delay between the left and right channel signals, so that the opposite end correctly synthesizes the signals according to the received delay, thereby improving the stability of the synthesized stereo sound field.
  • the present invention can be implemented by means of software plus a necessary general hardware platform, and of course, can also be through hardware, but in many cases, the former is a better implementation. the way.
  • the technical solution of the present invention may be embodied in the form of a software product in essence or in the form of a software product, which may be stored in a storage medium such as a ROM/RAM or a disk. , CD, etc., including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.)

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Acoustics & Sound (AREA)
  • Signal Processing (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Stereophonic System (AREA)

Abstract

本发明实施例提供一种通道间延迟估计的方法及其设备和编码器，所述方法包括：从左右声道合成信号的互相关函数及累积互相关函数中分别获取信号的声场信息；根据所述分别获取的声场信息获得累积互相关函数调整信息；采用所述调整信息对所述累积互相关函数进行调整，得到调整后的累积互相关函数；确定所述调整后的累积互相关函数中最大值对应的时间为通道间延迟。 本发明实施例以正确估计出左右声道信号之间的延迟，提高合成立体声的声场的稳定性。

Description

通道间延迟估计的方法及其装置和编码器
本申请要求于 2009 年 3 月 25 日提交中国专利局， 申请号为 200910129492.3 , 发明名称为"通道间延迟估计的方法及其装置和编码器 "的中 国专利申请的优先权， 其全部内容通过引用结合在本申请中。 技术领域
本发明涉及通信技术，特别涉及一种通道间延迟估计的方法及其装置和编 码器。 背景技术
随着计算机技术、数字信号处理技术的发展， 以及由于发展高清晰度电视 声系统、 家用视听系统的需要， 使立体声技术有较大的发展， 当然， 这也对立 体声技术尤其编解码技术提出了较高的要求。
常用的立体声编码方法为参数立体声编码的方法，在参数立体声编码方法 中， 通常不直接对左右声道信号进行编码， 而是将左右声道信号进行下混， 得 到下混信号， 对下混信号进行编码。 其中， 在编码时添加了一些额外的边带信 息。 在解码端， 可以通过下混信号和所述边带信息来恢复立体声信号。 而评价 立体声信号质量的好坏， 很大程度上取决于下混信号的质量。 也就是说， 在编 码端， 左右两个声道的信号越同步， 在下混的过程中损失的信息就会越少， 而 在通常情况，发声物体相对录制左右声道的两个麦克来说会有距离的变动或者 距离差，这样必然会造成左右两路信号之间不能完全同步， 即左右两路信号间 会存在一定的延时。 为了能保持左右声道的信号同步，从而提出了延迟估计方 法， 以便于提升立体声合成信号的质量。
目前，现有技术提供的估计延迟的方法包括在左右声道信号生成下混信号 之前，对左右声道信号求出累积互相关函数， 利用所述累计互相关函数的最大 值对应的时间作为左右声道之间的延迟，对该延迟进行编码， 并将该编码后的 延迟发送到解码端， 以便在解码端根据该延迟进行信号的合成，从而保持左右 声道信号的声场稳定。 在实际的应用中， 为了使左右声道的延迟保持稳定， 通 常将累积互相关函数作为判决的依据。 为方^ ^见约定， 当左声道相对右声道 在前的时候， 延迟为正， 反之延迟为负。
但是， 在上述方法中， 当左右声道信号的声场发生变化的时候， 比如从一 个方向转换到另外一个方向的时候，估计的延迟的正负就发生了变化， 而现有 技术中却不能很好跟踪声场的这种变化， 也就是说， 当声场发生变化时， 由于 累积互相关函数不能感应到该变化，将会导致延迟估计错误， 而解码端在根据 错误的延迟进行信号合成时， 信号的声场会出现不稳定的情况。
由此可见， 在对现有技术的研究和实践过程中， 本发明的发明人发现， 现 有的实现方式中，在左右声道信号的声场发生变化时， 由于不能很好跟踪声场 的这种变化，造成了不能正确估计出左右声道之间的延迟，从而导致合成立体 声的不稳定性， P争低了立体声编码质量， 影响听觉效果。
发明内容
本发明实施例提供一种通道间延迟估计的方法及设备，以正确估计出左右 声道信号之间的延迟， 提高合成立体声的声场的稳定性。 为解决上述技术问题， 本发明是实施例提供一种通道间延迟估计的方法， 包括：
从左右声道合成信号的互相关函数及累积互相关函数中分别获取信号的 声场信息；
根据所述分别获取的声场信息获得累积互相关函数调整信息；
采用所述调整信息对所述累积互相关函数进行调整，得到调整后的累积互 目关函数；
确定所述调整后的累积互相关函数中最大值对应的时间为通道间延迟。 相应地， 本发明实施例提供一种通道间延迟估计的装置， 包括： 提取单元，用于从左右声道合成信号的所述互相关函数及累积互相关函数 中分别获取信号的声场信息；
调整单元，用于根据提取单元分别获取的所述声场信息获得累积互相关函 数调整信息， 采用所述调整信息对所述累积互相关函数进行调整，得到调整后 的累积互相关函数；
延迟估计单元，用于确定调整单元调整后的所述累积互相关函数中最大值 对应的时间为通道间延迟。
相应地， 本发明实施例还提供一种编码器， 包括： 通道间延迟估计的装置 和编码装置， 其中
所述通道间延迟估计的装置，用于从左右声道合成信号的互相关函数及累 积互相关函数中分别获取信号的声场信息，根据所述分别获取的声场信息获得 累积互相关函数调整信息， 采用所述调整信息对所述累积互相关函数进行调 整，得到调整后的累积互相关函数，确定所述调整后的累积互相关函数中最大 值对应的时间为通道间延迟， 并将所述通道间延迟输出至编码装置；
所述编码装置， 用于对接收到的通道间延迟进行编码， 并发送编码后的通 道间延迟。
由上述技术方案可知，本发明实施例提供确定左右声道信号间的互相关函 数及累积互相关函数；利用从所述互相关函数中提取信号的声场信息来调整累 积互相关函数； 并确定所述累积互相关函数中最大值对应的时间为估计的延 迟。也就是说，在左右声道信号的声场发生变化时，通过提取声场变化的信息， 以正确估计出左右声道信号间的延迟，以便于对端根据接收到的延迟正确对信 号进行合成， 从而提高合成立体声的声场的稳定性。 附图说明
图 1为本发明实施例中通道间延迟估计的方法的流程图；
图 2为本发明实施例 1中通道间延迟估计的方法的流程图； 图 3为本发明实施例 1中利用声场信息调整累积互相关函数的流程图； 图 4为本发明实施例 1 中另一利用声场信息调整累积互相关函数的流程 图；
图 5为本发明实施例 1中又一种利用声场信息调整累积互相关函数的流程 图；
图 6为本发明实施例 1 中才据比值判断声场类别的一种应用实例的流程 图；
图 7为本发明实施例 2中通道间延迟估计的方法的流程图； 图 8为本发明实施例 2中利用声场信息调整累积互相关函数的流程图； 图 9为本发明实施例 2中另一种利用声场信息调整累积互相关函数的流程 图；
图 10为本发明实施例 1中又一利用声场信息调整累积互相关函数的流程 图；
图 11为本发明实施例中 3中通道间延迟估计的方法的流程图；
号延迟估计的对比示意图；
图 13为本发明实施例中一种通道间延迟估计的装置的结构示意图； 图 14为本发明实施例中编码器的结构示意图。 具体实施方式
下面结合附图， 对本发明的最佳实施方案进行伴细描述。
请参阅图 1， 为本发明实施例中通道间延迟估计的方法的流程图； 所述方 法包括：
步骤 101 : 确定左右声道信号的互相关函数及累积互相关函数； 在本实施 例中该步骤为可选步骤。 其中，
ccf{d) r{n - d))
Figure imgf000006_0001
其中， d是延迟， 是常量， n是样点个数， 是变量； r为右声道的信号； 1 为左声道的信号； 当然，确定互相关函数也可以通过其他的公式得到，本实施例并不限于此， t匕
Figure imgf000007_0001
ccf(d) = 0) 而累积互相关函数， 是一级 MA函数， 比如
a _ccf{d) = a _ ccf(d) * + ccf(d) a≥0 在该步骤中， 《是加权系数， 是个变量， 对于确定互相关函数及其累积互 相关函数， 对于本领域技术人员来说， 已为公知技术， 在此不再赘述。
步骤 102: 从左右声道信号的互相关函数及累积互相关函数中分别获取信 号的声场信息；
可以从互相关函数中提取当前帧互相关函数的声场信息，以及从累积互相 关函数中提取当前帧互相关函数之前的累积互相关函数的声场信息；也可以从 互相关函数中提取短时互相关函数的声场信息，以及从累积互相关函数中提取 长时累积互相关函数的声场信息， 本实施例不作限制。 步骤 103: 根据分别获取的声场信息获得累积互相关函数调整信息， 采用 所述调整信息对所述累积互相关函数进行调整， 得到调整后的累积互相关函 数；
可以根据提取的不同声场信息确定累积互相关函数的加权系数，利用所述 加权系数对所述累积互相关函数进行调整，得到调整后的累积互相关函数； 也 可以在所述确定累积互相关函数的加权系数的基石出上，在乘以提取信号类型对 应的值来确定累积互相关函数的加权系数；也可以根据提取当前帧互相关函数 与累积互相关函数的不同声场信息，确定对应的声场类别； 判断所述对应的声 场类别是否相同， 并根据判断结果设置累积互相关函数的加权系数； 利用设置 的加权系数对所述累积互相关函数进行调整， 得到调整后的累积互相关函数。 步骤 104 : 确定所述调整后的累积互相关函数中最大值对应的时间为通道 间延迟。
所述方法还包括： 判断所述声场信息是否发生变化， 若是， 则执行步骤 103; 否则， 结束流程。 也就是说，在本实施例中估计声道间延时，先提取当前帧互相关函数及其 之前累积互相关函数的声场信息，根据提取的声场信息计算累积互相关函数的 加权系数，通过修改的加权系数来调整累积互相关函数，从而估计出声场变化 时左右声道信号间的延迟。 即本实施例提取声场变化的信息，根据声场信息的 变化对左右声道间的延时估计进行调整。具体包括：根据提取当前帧互相关函 数的声场信息同累积互相关函数的声场信息的变化对延迟估计的累积互相关 函数进行自适应加权调整，或根据短时互相关函数的声场信息同长时累积互相 关函数的声场信息变化对延迟估计相关函数的累积互相关函数进行自适应加 权调整，或者，根据提取的信号类型和提取的声场信息对延迟估计相关函数的 累积互相关函数进行自适应加权调整， 从而正确估计出左右声道信号间的延 迟， 并发送该延迟， 以便于接收端根据接收到的延迟正确对信号进行合成， 从 而提高合成立体声的声场的稳定性。
为了便于本领域技术人员的理解， 下面以具体的实施例来说明。
实施例一
请参阅图 2， 为本发明实施例 1中通道间延迟估计的方法的流程图， 所述 方法包括：
步骤 201 : 将左右声道的两路信号分别进行加窗处理， 并输出加窗处理的 信号， 该步骤为可选步骤；
步骤 202 : 求加窗处理后左右声道信号的互相关函数； 其具体的求互相关 函数的过程详见上述公式， 在此不再赘述。 步骤 203: 求所述累积互相关函数； 其具体的求累积互相关函数的过程详 见上述公式， 在此不再赘述。
步骤 204: 从所述互相关函数中提取当前帧互相关函数的声场信息； 具体 包括：
对第一部分延迟时刻当前帧互相关函数的和与第二部分延迟时刻当前帧 互相关函数的和进行操作，获得第一声场信息值； 对第一部分延迟时刻累积互 相关函数的和与第二部分延迟时刻累积互相关函数的和进行操作，获得第二声 场信息值。
其中， 所述第一部分延迟时刻当前帧互相关函数定义为延迟大于等于 0 的当前帧互相关函数；所述第二部分延迟时刻累积互相关函数定义为延迟小于 等于 0的累积互相关函数。
在本实施例中， 所述操作以除法和减法为例， 而所述第一声场信息值包括 下述的第一比值或第一差值，所述第二声场信息值包括下述的第二比值或第二 差值， 但并不限于此。
一种优选的提取声场信息的方式为：先确定延迟大于等于 0的当前帧互相 关函数的和，再确定延迟小于等于 0的当前帧互相关函数的和， 最后将求出延 迟大于等于 0的当前帧互相关函数的和与延迟小于等于 0的当前帧互相关函数 的和相除，得到的比值称为第一比值， 所述第一比值为提取当前帧互相关函数 的声场信息。
另一种提取声场信息的方式为：先确定延迟大于等于 0的当前帧互相关函 数的和，再确定延迟小于等于 0的当前帧互相关函数的和， 最后将求出延迟大 于等于 0的当前帧互相关函数的和与延迟小于等于 0的当前帧互相关函数的和 相减，得到的差值称为第一差值， 所述第一差值为提取当前帧互相关函数的声 场信息值。
当然， 本发明实施例并不限于此。 步骤 205 :将累积互相关函数延迟一帧或者多帧（本实施例不作限制）后， 提取该延迟后的累积互相关函数的声场信息， 比如， 如果当前帧为 N帧， 则累 积互相关函数即为过去 N-1帧累积互相关函数的声场信息； 一种提取的方式为： 先确定延迟大于等于 0的累积互相关函数的和，再确 定延迟小于等于 0的累积互相关函数的和，最后将求出延迟大于等于 0的累积 互相关函数的和与延迟小于等于 0的累积互相关函数的和相除，得到的比值称 为第二比值， 所述第二比值为提取累积互相关函数的声场信息。 另一种提取的方式为： 先确定延迟大于等于 0的累积互相关函数的和，再 确定延迟小于等于 0的累积互相关函数的和，最后将求出延迟大于等于 0的累 积互相关函数的和与延迟小于等于 0的累积互相关函数的和相减，得到的差值 称为第二差值， 所述第二差值为提取累积互相关函数的声场信息。
步骤 206 : 利用提取到变化的声场信息来计算累积互相关函数的加权系 数；
其中， 计算的方式有很多种， 本实施例以： 取所述第一比值与第二比值之 差的绝对值； 或者取第一差值与第二差值的绝对值为例，从而得到累积互相关 函数的加权系数， 但并不限于此。 步骤 207 : 根据所述累积互相关函数的加权系数来调整累积互相关函数； 其具体的调整的过程也就是用计算得到的加权系数作为调整加权系数来 计算累积互相关函数， 其具体的实现过程详见图 3和图 4。 步骤 208 : 搜索该累积互相关函数中最大值对应的时间， 所述时间为估计 的延迟； 具体的搜索方式对于本领域技术人员来说，已为公知技术，在此不再赘述。 步骤 209 : 判断所述延迟的变化与原来的延迟相比是否有效， 若有效， 执 行步骤 210; 否则， 执行步骤 211 ; 其判断的依据为： 将确定的延迟与原来的 延迟相比较， 如果确定的延迟满需条件， 则有效， 否则无效。 步骤 210: 输出所述延迟； 步骤 211: 输出原来的延迟。 在本实施例中，通过提取当前帧互相关函数的声场信息和当前帧互相关函 数之前的累积互相关函数的声场信息， 来判断左右声道的声场是否发生变化， 并根据声场的变化计算出不同的累积互相关函数的加权系数，根据所述加权系 数来调整累积互相关函数， 以便跟踪上声场的变化，从而估计出比较准确的延 迟。 还请参阅图 3， 为本发明实施例 1中利用声场信息调整累积互相关函数的 流程图， 在本实施例中， 当前帧互相关函数以 Cc/(")
Figure imgf000011_0001
对 于当前帧之前的累积互相关函数以 Cc/(")， ~ <η<Τ,Τ>0^^. 所述互相 关函数是包括归一化的互相关函数， 但并不限于此。 具体包括： 步骤 301: 求延迟大于等于零的当前帧互相关函数的^^延迟小 等于零 cur _ ratio = Ccf (ή) I 2 Ccf (n) 的当前巾贞互相关函数的和的比值（ cur— ratio ): ― "=。 r+i ； 在该步骤中， 可以将 cur_ratio限制在某一范围， 比如 <min，max>， 其中 min和 max的值可以根据经验来设定， 也可以将 min的值可以设置为 0， 而将 max 的值设置可以为无穷大， 本实施例不做限制， 其中， 设置 <min，max>的目 的是为了防止 cur-ratio过大或过小。 步骤 302: 求延迟大于等于零的累积互相关函数的和与延迟小于等于零的 累 积 互^_ 目 关 函 。数 的 和 的 比 值 ( prev-ratio ) ： prev _ ratio = ^ ac _ Ccf (n) I ^ ac _ Ccf (n)
"=° "=- r+1 ； prev—ratio 可以限制在<11^11， max>>^ 间， 其<1^11，111& >上述^11"_1&110的限定范围相同， 在此不再赞述。 步骤 303: 根据所述求出的 cur -ratio与 prev—ratio计算累积互相关函 数的加权系数； 一种方式为： 通过下述公式求出累积互相关函数的加权系数， /曰并不限于此. "―
Figure imgf000011_0002
lk+b 其中， a为累积互相关函数的加权系数， cur_ratio为延迟大于等于 0的 当前帧互相关函数的和与延迟小于等于 0 的当前帧互相关函数之后的比值； prev- rat io为延迟大于等于 0的累积互相关函数的和与延迟小于等于 0的累 积互相关函数的和的比值； k和 b为常数。 比如， 在实际应用中， 上述计算加 权系数中的一组参数为： min =0. 5， max =1. 5， k =—0. 2， b =1， 但并不限于 此。
步骤 304: 利用所述加权系数对累积互相关函数进行加权运算， 得到加权 后的累积互相关函数，也就是说，加权后的互相关函数能更好的跟踪上声场的 变化。
本实施例给出一种累积互相关函数的形式，即通道间延迟的累积互相关函 数是累积互相关函数乘以一个加权系数并和当前帧的互相关函数的和，但并不 仅限于所述的累积互相关函数， 具体为：
ac_ Ccfin) = ac_ Ccfin) * a + Ccfin) - T < n < T,T > 0 其中， a为加权系数。 还请参阅图 4， 为本发明实施例 1中另一利用声场信息调整累积互相关函 数的流程图， 具体包括： 步骤 401 : 求出延迟大于等于零的当前帧互相关函数的和与延迟小于等于 零的当前帧互相关函数的和的差， 该差值称为第一差值；
步骤 402: 求出延迟大于等于零的累积互相关函数的和与延迟小于等于零 的累积互相关函数的和的差， 该差值称为第二差值； 步骤 403: 求第一差值与第二差值之差的绝对值， 得到累积互相关函数的 力口权系数；
其中， 可以根据公式 " =1第一差值-第二差值 1 + ， 计算得到累积互相关函 数的加权系数； 当然， 计算所述加权系数的公式并不限于此， 也可以通过其他 的公式来计算。
步骤 404: 利用所述加权系数对累积互相关函数进行加权运算， 得到加权 后的累积互相关函数。 还请参阅图 5， 为本发明实施例 1中又一种利用声场信息调整累积互相关 函数的流程图，在该实施例中， 首先分别获取当前帧互相关函数中延迟大于等 于零和延迟小于等于零的和，根据两者的比值来确定判断声场的类别， 即判断 所述两者的比值对应的声场类别是否相同，并根据判断结果设置累积互相关函 数的加权系数； 利用设置的加权系数调整累积互相关函数。
其中，在该实施例中，当前帧互相关函数仍以 /(") - <n<T,T>0^^ 对于当前帧之前的累积互相关函数仍以" c- Cc (")， -T<n<T,T>0^^. 所述 互相关函数是包括归一化的互相关函数， 但并不限于此。 具体包括：
步骤 501: 求延迟大于等于零的当前帧互相关函数的和 延迟小 ^等于零 cur _ ratio = ^ Ccf (n) I ^ Ccf (n) 的当前巾贞互相关函数的和的比值（ cur— ratio ): "=。 ^ 步骤 502: 居所述比值 cur_ratio 确定当前帧对应的声场类别， 用 Cur_Flag来标记； 具体包括： 判断所述比值 cur_ratio是否大于预设的第一 阔值； 若是， 则设置比值 cur_ratio对应的声场类别的标志为 1; 否则， 继续 判断比值 cur_ratio 是否大于等于第二阔值， 若是， 则设置比值 cur_ratio 对应的声场类别的标志为 0; 否则， 设置比值 cur_rat io对应的声场类别的标 志为- 1， 其中， 所述第二阔值小于第一阔值； 其具体的实现过程详见图 6。 步骤 503: 求延迟大于等于零的累积互相关函数的和与延迟小于等于零的 累积互相关函数的和的比值（ prev-ratio ):
T-1 0
prev _ ratio = ^ ac _ Ccf n)l ^ ac _ Ccf (n)
~ "=o ~ "=- " 步骤 504: ^居所述比值 prev-ratio确定累积互相关函数对应的声场类 另 |J，用 prev-flag来标记；其确定的过程与步骤 503中的过程类似，具体包括： 判断所述比值 prev-ratio是否大于预设的第一阔值； 若是， 则设置比值 prev-ratio对应的声场类别的标志为 1; 否则继续判断比值 prev_ratio是否 大于等于第二阔值， 若是， 则设置比值 prev-ratio对应的声场类别的标志为 0; 否则， 设置比值 prev-ratio对应的声场类别的标志为- 1， 其中， 所述第二 阔值小于第一阔值。 其具体的实现过程详见图 5。
步骤 505: 判断所述比值 cur_ratio 对应的声场类别是否与比值 prev-ratio对应的声场类别相同， 若相同， 执行步骤 506和步骤 508; 否则执 行步骤 507和步骤 508; 步骤 506: 设置累积互相关函数的加权系数为 1; 步骤 507: 设置累积互相关函数的加权系数为小于 1， 一般设置为 0.85， 也可以是其他小于 1的值， 本实施例不作限制。
步骤 508: 根据设置的加权系数调整累积互相关函数。 即利用加权系数对累积互相关函数进行加权运算，以便于加权后的累积互 相关函数能更好的跟踪上声场的变化。本实施例提供一种累积互相关函数形式 但并不仅限于这一种累积互相关函数，即通道间延迟的累积互相关函数是累积 的互相关函数乘以一个加权系数并和当前帧的互相关函数的和，：
ac _ Ccf (n) = ac— Ccf (n) * rate + Ccf (n) -T < n <T,T > 0 其中， rate是加权系数的比值。 还请参阅图 6， 为本发明实施例 1中根据比值判断声场类别的一种应用实 例的流程图； 在该实施例中， 可以才 据该比值把声场分为 3类， 比值大于 1.2 (即第一阔值） 时， 声场类别的标志（Flag)设置 1; 比值大于 0.8同时小于 等于 1.2 (即第二阔值） 时， 声场类别的标志设置为 0; 比值小于 0.8时， 声 场类别的标志设置为 -1。 由此可知，可以根据声场的变化设置不同的加权系数 来调整累积互相关函数。 具体的判断过程包括：
步骤 601: 判断比值 ratio是否大于 1.2; 如果大于， 则执行步骤 602; 否则， 执行步骤 603; 步骤 602:设置比值 ratio对应的声场类别的标记为 1,比如 Cur_Flag=l; 或者 prev_flag=l; 步骤 603: 继续判断比值 rat io是否大于等于 0. 8， 若是， 执行步骤 604; 否则， 执行步骤 605; 步骤 604: 设置比值 rat io对应的声场类别的标记为 0; 步骤 605: 设置比值 rat io对应的声场类别的标记为 -1。 在本实施例中， 所述比值 rat io 可以是比值 cur— rat io, 也可以是比值 prev- rat io , 本实施例不作限制。 实施例二
请参阅图 7， 为本发明实施例 2中通道间延迟估计的方法的流程图， 本实 施例与实施例一的实现过程类似，其不同之处包括： 从互相关函数中提取短时 互相关函数的声场信息，以及从累积互相关函数中提取长时累积互相关函数的 声场信息， 然后根据所提取的不同声场信息计算累积互相关函数的加权系数。 在本实施例中， 所短时互相函数和长时累积互相关函数是一个相对的概念， 比 如，
a _ ccf\{d) = a— ccf\{d) * l + ccf (d) a _ ccf 2(d) = a _ ccf 2(d) * 2 + ccf (d) 如果《1大于《2， 那么" - /Κ 是长时累积互相关函数； "― /2 ( 是短时 累积互相关函数。 其具体的实现过程如图 6所示， 具体包括： 步骤 701 : 将左右声道的两路信号分别进行加窗处理， 并输出加窗处理的 信号， 该步骤为可选步骤；
步骤 702: 求加窗处理后左右声道信号的互相关函数； 步骤 703: 求所述累积互相关函数； 其中， 步骤 702和步骤 703的具体的实现过程详见实施例一，在此不再赞 述；
步骤 704: 从所述互相关函数中提取短时互相关函数的声场信息； 具体包 括：
对第三部分延迟时刻短时互相关函数的和与第四部分延迟时刻短时互相 关函数的和进行操作， 获得第三声场信息值；
对第三部分延迟时刻长时累积互相关函数的和与第四部分延迟时刻长时 累积互相关函数的和进行操作， 获得第四声场信息值。 其中，所述第三部分延迟时刻短时互相关函数定义为延迟大于等于 0的短 时互相关函数；所述第四部分延迟时刻长时累积互相关函数定义为延迟小于等 于 0的长时累积互相关函数。 在本实施例中， 所述操作以除法和减法为例， 而所述第三声场信息值包括 下述的第三比值或第三差值，所述第四声场信息值包括下述的第四比值或第四 差值， 但并不限于此。 一种优选提取声场信息的方式为：确定延迟大于等于 0的短时互相关函数 的和与延迟小于等于 0的短时互相关函数的和的比值， 为了方便描述，将其称 为第三比值； 所述第三比值为提取短时互相关函数的声场信息。 另一种提取声场信息的方式为：确定延迟大于等于 0的短时互相关函数的 和， 以及延迟小于等于 0的短时互相关函数的和，将所述延迟大于等于 0的短 时互相关函数的和与延迟小于等于 0的短时互相关函数的和相减，得到的差值 称为第三差值， 所述第三差值为提取短时互相关函数的声场信息。 步骤 705 : 将累积互相关函数延迟一帧或者多帧（本实施例以延迟一帧为 例）后，提取该延迟后的短时互相关函数之前累积的长时累积互相关函数的声 场信息；
一种提取的方式为：确定延迟大于等于 0的长时累积互相关函数的和与延 迟小于等于 0的长时累积互相关函数的和的比值。 为了方便描述，将其称为第 四比值； 所述第四比值为提取长时累积互相关函数的声场信息。 另一种提取的方式为： 确定延迟大于等于 0的长时累积互相关函数的和， 以及延迟小于等于 0的长时累积互相关函数的和；将延迟大于等于 0的长时累 积互相关函数的和与延迟小于等于 0的长时累积互相关函数的和相减，得到的 差值称为第四差值， 所述第四差值为提取长时累积互相关函数的声场信息。 步骤 706 : 利用提取到的变化的声场信息来计算累积互相关函数的加权系 数；
其中， 计算的方式有很多种， 本实施例以： 取所述第三比值与第四比值之 差的绝对值； 或者取第三差值与第四差值的绝对值为例，从而得到累积互相关 函数的加权系数， 但并不限于此。 步骤 707 : 根据所述累积互相关函数的加权系数来调整累积互相关函数； 具体过程包括： 其具体的实现过程详见图 8、 图 9和图 10。 步骤 708 : 搜索该累积互相关函数中最大值对应的时间， 所述时间为估计 的延迟；
步骤 709 : 判断所述延迟的变化与原来的延 目比较是否有效， 若有效， 则执行步骤 710: 否则， 返回步骤 711。 其判断的依据为： 将确定的延迟与原 来的延 目比较， 如果确定的延迟满足条件， 则有效， 否则无效。 步骤 710: 输出所述延迟； 步骤 711 : 输出原来的延迟。 在本实施例中，通过提取短时互相关函数的声场信息和短时互相关函数之 前累积的长时累积互相关函数声场信息， 来判断左右声道的声场是否发生变 化， 并根据发生变化的声场信息计算出不同的累积互相关函数的加权系数，根 据所述加权系数来调整累积互相关函数， 以便跟踪上声场的变化，从而估计出 比较准确的延迟。 还请参阅图 8， 为本发明实施例 2中利用声场信息调整累积互相关函数的 流程图，在本实施例中，首先分别获取短时互相关函数中延迟大于等于零的和， 与延迟小于等于零的和，根据两者的比值来确定判断声场的类别， 即判断所述 两者的比值对应的声场类别是否相同，并根据判断结果设置累积互相关函数的 加权系数； 利用设置的加权系数调整累积互相关函数。
其中，在该实施例中，短时互相关函数以 "—cc/2 ( -T<d<T,T>o
Figure imgf000018_0001
对于累积互相关函数中的长时累积互相关函数以 a— ccfl(d)， ~T<d<T,T>0^ 例； 具体包括：
步骤 801: 求延迟大于等于零的短时互相关函数的和与延迟小于等于零的 短时互相关函数的和的比值（cur_ratio)， 为了便于描述， 该比值称为第三比 值， 具体的公 ^ 。
cur _ ratio - _ acf2(d) I _ acf\{d) 步骤 802: 居所述比值 cur_ratio 确定当前帧对应的声场类别， 用 Cur—Flag来标记； 具体过程包括： 判断所述第三比值是否大于预设的第一阔值； 若是， 则设置第三比值对应 的声场类别的标志为 1;否则继续判断第三比值是否大于等于第二阔值，若是， 则设置第三比值对应的声场类别的标志为 0; 否则， 设置第三比值对应的声场 类别的标志为 -1， 其中， 所述第二阔值小于第一阔值； 步骤 803: 求延迟大于等于零的长时累积互相关函数的和与延迟小于等于 零的长时累积互相关函数的和的比值（pr^Y_ratio): 为了便于描述， 该比值 prev ratio = V a ccf\(d)l V a ccf2(d)
称为第三比值， 具体的公式： _ — ^ - 步骤 804: ^居所述比值 prev-ratio确定累积互相关函数对应的声场类 另 |J，用 prev-flag来标记；其确定的过程与步骤 703中的过程类似，具体包括： 判断所述第四比值是否大于预设的第一阔值； 若是， 则设置第四比值对应 的声场类别的标志为 1;否则继续判断第四比值是否大于等于第二阔值，若是， 则设置第四比值对应的声场类别的标志为 0; 否则， 设置第四比值对应的声场 类别的标志为 -1， 其中， 所述第四阔值小于第三阔值。 步骤 805: 判断所述比值 cur_ratio 对应的声场类别是否与比值 prev-ratio对应的声场类别相同， 若相同， 执行步骤 806和步骤 808; 否则执 行步骤 807和步骤 808; 步骤 806: 设置累积互相关函数的加权系数为 1; 步骤 807: 设置累积互相关函数的加权系数为小于 1， 一般设置为 0.85， 也可以是其他小于 1的值， 本实施例不作限制。 步骤 808: 根据设置的加权系数调整累积互相关函数。 本实施例中， 首先对左右声道的两路信号分别进行加窗处理， 并求所述两 路信号间的互相关函数；从互相关函数中提取短时互相关函数的声场信息及过 去 N- 1帧累积长时累积互相关函数的声场信息，利用提取到不同的声场信息来 调整相关累积互相关函数的加权系数；从累积互相关函数搜索出最大的累积互 相关函数的值， 并求该累积互相关函数中最大值的对应的时间， 所述时间为估 计的延迟； 然后判断所述延时是否为有效的延迟， 若是， 则输出该延迟输出， 以便于接收端根据接收到的延迟正确对信号进行合成，从而提高合成立体声的 声场的稳定性。 还请参阅图 9, 为本发明实施例 2中另一种利用声场信息调整累积互相关 函数的流程图， 其中， 在该实施例中， 短时互相关函数以 " -cc/2(° -Τ^ΤΉ为例， 对于累积互相关函数中的长时累积互相关函数以 a_ccf\{d) ^ _Γ<ί/<Γ,Γ>0为例； 具体包括： 步骤 901: 求延迟大于等于零的短时互相关函数的和与延迟小于等于零的 短时互相关函数的和的比值（cur_ratio)， 为了便于描述， 该比值称为第三比 值， 具体的公 。
cur _ ratio - acf2(d) I acf\{d) 步骤 902: 求延迟大于等于零的长时累积互相关函数的和与延迟小于等于 零的长时累积互相关函数的和的比值（prev-ratio): 为了便于描述， 该比值 Γ-1 0
prev ratio = V a ccf\(d)l V a ccf2(d)
称为第三比值， 具体的公式： _ — ^ - 步骤 903: 根据所述求出的 cur -ratio与 prev_ ratio计算累积互相关函 数的加权系数； 具体可以通过下述公式算累积互相关函数的加权系数，但并不 限于计该公式， 其公式为：
Figure imgf000020_0001
lk+b a为累积互相关函数的加权系数， k和 b为常数。 比如， 在实际应用中， 上述计算加权系数中的一组参数为： min =0.5， max =1.5， k = -0.2， b =1， 但并不限于此。 步骤 904: 利用所述加权系数对累积互相关函数进行加权运算， 得到加权 后的累积互相关函数。 其累积互相关函数的形式详见上述。 还请参阅图 10， 为本发明实施例 1 中又一利用声场信息调整累积互相关 函数的流程图， 具体包括： 步骤 1001: 求出延迟大于等于零的短时互相关函数的和与延迟小于等于 零的短时互相关函数的和的差， 该差值称为第三差值； 步骤 1002: 求出延迟大于等于零的长时累积互相关函数的和与延迟小于 等于零的长时累积互相关函数的和的差， 该差值称为第四差值；
步骤 1003: 求第三差值与第四差值之差的绝对值， 得到累积互相关函数 的加权系数； 其中， 可以根据公式 "=1第三差值―第四差值 1 +6， 计算得到累积互相关函 数的加权系数； 当然， 计算所述加权系数的公式并不限于此， 也可以通过其他 的公式来计算。 步骤 1004: 利用所述加权系数对累积互相关函数进行加权运算， 得到加 权后的累积互相关函数。 实施例三 还请参阅图 11， 为本发明实施例中 3中通道间延迟估计的方法的流程图， 所述方法包括：
步骤 111 : 将左右声道的两路信号分别进行加窗处理， 并输出加窗处理的 信号， 该步骤为可选步骤；
步骤 112: 求加窗处理后左右声道信号的互相关函数； 步骤 113: 求所述累积互相关函数；
其中，步骤 112和步骤 113的具体实现过程详见实施例一，在此不再赞述； 步骤 114: 从所述互相关函数中提取信号类型， 以及当前帧或短时互相关 函数的声场信息； 其中，对于从所述互相关函数中提取当前帧或短时互相关函数的声场信息 的具体实现过程详见上述， 在此不再赘述。
从所述互相关函数中提取信号类型的过程包括：可以从所述互相关函数中 采集信号类型， 具体的采集过程对于本领域技术人员来说， 已为公知技术， 在 此不再赘述。
步骤 115: 将累积互相关函数延迟一帧或者多帧（本实施例以延迟一帧为 例）后，提取该延迟后的当前帧互相关函数之前的累积互相关函数或累积互相 关函数中长时累积互相关函数的声场信息， 具体的实现过程， 详见上述， 在此 不再赘述。
步骤 116: 利用提取到的变化的声场信息来计算累积互相关函数的加权系 数；
其中， 计算的方式有很多种， 比如， 取所述第一比值与第二比值之差的绝 对值，再将取得的绝对值乘以信号类型对应的值； 或者取第一差值与第二差值 的绝对值， 再将取得的绝对值乘以信号类型对应的值； 当然， 也可以是其他的 计算方式， 本实施例不作限定。
步骤 117: 根据所述累积互相关函数的加权系数来调整累积互相关函数； 步骤 118: 搜索该累积互相关函数中最大值对应的时间， 所述时间为估计 的延迟；
步骤 119: 判断所述延迟的变化原来的延迟相比是否有效， 若有效， 执行 步骤 120: 否则执行步骤 121 ; 步骤 120: 输出所述延迟； 步骤 121 : 输出原来的延迟。 在本实施例中，通过提取信号类型， 以及当前帧或短时互相关函数的声场 信息，以及当前帧互相关函数之前的累积互相关函数或长时累积互相关函数的 声场信息， 来判断左右声道的声场是否发生变化， 并根据发生变化的声场信息 计算出不同的累积互相关函数的加权系数，根据所述加权系数来调整累积互相 关函数， 以便跟踪上声场的变化， 从而估计出比较准确的延迟。
段立体声信号延迟估计的对比示意图，从该图中对应的波形可以看出，本发明 实施例改进后估计的延迟快于现有技术， 从而更加正确的跟踪上延迟的变化。 基于上述方法的实现过程，本发明实施例还提供有一种通道间延迟估计的 装置， 其结构示意图详见图 13， 包括： 提取单元 131、 调整单元 132和延迟估 计单元 133， 其中， 所述提取单元 131， 用于从左右声道信号的互相关函数及 累积互相关函数中分别获取信号的声场信息； 所述调整单元 132， 用于根据提 取单元分别获取的所述声场信息获得累积互相关函数调整信息，采用所述调整 信息对所述累积互相关函数进行调整，得到调整后的累积互相关函数； 所述延 迟估计单元 133， 用于确定调整单元调整后的所述累积互相关函数中最大值对 应的时间为通道间延迟。 所述提取单元具体包括： 第一提取单元和第二提取单元， 所述第一提取单 元， 用于从左右声道信号的互相关函数中提取声场信息； 所述第二提取单元， 用于从第一提取单元提取的互相关函数之前的累积互相关函数中提取声场信 息。 所述第一提取单元包括： 第一计算单元和第一确定单元， 所述第一计算单 元，用于计算第一部分延迟时刻互相关函数的和与第二部分延迟时刻互相关函 数的和； 所述第一确定单元， 用于对第一计算单元计算第一部分延迟时刻互相 关函数的和与第二部分延迟时刻互相关函数的和进行操作，获得第一声场信息 值；
所述第二提取单元包括： 第二计算单元和第二确定单元， 所述第二计算单 元，用于计算第一部分延迟时刻累积互相关函数的和与第二部分延迟时刻累积 互相关函数的和； 所述第二确定单元， 用于对第二计算单元计算第一部分延迟 时刻累积互相关函数的和与第二部分延迟时刻累积互相关函数的和进行操作， 获得第二声场信息值。 所述调整单元包括： 第一系数计算单元与第一调整单元， 其中， 所述第一 系数计算单元，用于根据所述第一声场信息值与第二声场信息值计算累积互相 关函数的加权系数； 所述第一调整单元， 用于利用所述第一系数计算单元计算 的加权系数对所述累积互相关函数进行调整， 得到调整后累积互相关函数。 所述调整单元包括： 第一声场类别确定单元、 第一判断单元、 第一设置单 元与第二调整单元， 其中， 所述第一声场类别确定单元， 用于根据所述第一和 第二确定单元确定的第一声场信息值与第二声场信息值确定对应的声场类别； 所述第一判断单元，用于判断所述第一声场信息值与第二声场信息值对应的声 场类别是否相同， 并发送判断结果； 所述第一设置单元， 用于根据接收到第一 判断单元发送的判断结果设置累积互相关函数不同的加权系数； 第二调整单 元，用于利用第一设置单元设置的不同加权系数对所述累积互相关函数进行调 整， 得到调整后的累积互相关函数。 其中， 所述第一提取单元从左右声道信号的互相关函数包括： 当前帧互相 关函数，所述第二提取单元从第一提取单元提取的互相关函数之前的累积互相 关函数包括： 当前帧互相关函数之前的累积互相关函数； 所述第一提取单元从左右声道信号的互相关函数包括： 短时互相关函数； 所述第二提取单元从第一提取单元提取的互相关函数之前的累积互相关函数 包括： 长时互相关函数。 在所述互相关函数为短时互相关函数时， 所述第一提取单元还包括： 第三提取单元， 用于从左右声道信号的互相关 函数中提取信号类型； 所述调整单元还包括：
第二系数计算单元，用于根据所述信号类型对应的值对所述第一系数计算 单元算累积互相关函数的加权系数再次进行加权计算，得到计算累积互相关函 数的加权系数； 第三调整单元，用于利用所述第二系数计算单元计算的加权系数对所述累 积互相关函数进行调整， 得到调整后累积互相关函数。 其中， 所述装置还可以包括： 判断单元， 用于判断所述声场信息是否发生 变化， 并将发生变化的判断结果发送给调整单元。
其中， 所述通道间延迟估计的装置可以集成在编码器中，也可以集成在用 于通信的多用户定位设备中，还可以集成在多声源位置判定设备中，本发明实 施例不作限制。 所述通道间延迟估计的装置中各个单元的功能的作用的实现过程详见上 述方法中对应的实现过程， 在此不再赘述。 为了便于本领域技术人员的理解，下面互相关函数分别以当前帧互相关函 数和短时互相关函数；累积互相关函分别以当前帧互相关函数之前的累积互相 关函数及长时互相关函数来说明， 但并不限于此。 一种以当前帧互相关函数为例的实施例中： 所述提取单元具体包括： 当前帧提取单元和累积提取单元， 其中， 所述当 前帧提取单元，用于从左右声道信号的互相关函数中提取当前帧互相关函数的 声场信息； 所述累积提取单元， 用于从当前帧互相关函数之前的累积互相关函 数提取声场信息。 其中， 所述当前帧提取单元包括： 第一计算单元和第一确定单元， 所述第 一计算单元，用于计算第一部分延迟时刻当前帧互相关函数的和与第二部分延 迟时刻当前帧互相关函数的和，比如计算延迟大于等于 0的当前帧互相关函数 的和与延迟小于等于 0的当前帧互相关函数的和； 所述第一确定单元， 用于对 第一计算单元计算第一部分延迟时刻当前帧互相关函数的和与第二部分延迟 时刻当前帧互相关函数的和进行操作，获得第一声场信息值， 比如用于确定延 迟大于等于 0的当前帧互相关函数的和与延迟小于等于 0的当前帧互相关函数 的和的比， 称为第一比值； 或者， 用于确定延迟大于等于 0的当前帧互相关函 数的和与延迟小于等于 0的当前帧互相关函数的和的差， 称为第一差值。 所述累积提取单元包括： 第二计算单元和第二确定单元， 其中， 所述第二 计算单元，用于计算第一部分延迟时刻累积互相关函数的和与第二部分延迟时 刻累积互相关函数的和，比如用于计算延迟大于等于 0的累积互相关函数的和 与延迟小于等于 0的累积互相关函数的和； 所述第二确定单元， 用于对第二计 算单元计算第一部分延迟时刻累积互相关函数的和与第二部分延迟时刻累积 互相关函数的和进行操作，获得第二声场信息值， 比如用于确定延迟大于等于 0的累积互相关函数的和与延迟小于等于 0的累积互相关函数的和的比， 称为 第二比值； 或者， 用于确定延迟大于等于 0的累积互相关函数的和与延迟小于 等于 0的累积互相关函数的和的差， 称为第二差值。 所述调整单元包括： 系数计算单元与第一调整单元， 和 /或， 第一声场类 别确定单元、 第一判断单元、 第一设置单元与第二调整单元， 其中 所述系数计算单元，用于根据所述第一声场信息值与第二声场信息值计算 累积互相关函数的加权系数，比如根据所述第一和第二确定单元确定的第一比 值与第二比值计算累积互相关函数的加权系数； 所述第一调整单元，用于利用所述系数计算单元计算的加权系数对所述累 积互相关函数进行调整， 得到调整后累积互相关函数； 所述第一声场类别确定单元，用于根据所述第一和第二确定单元确定的第 一声场信息值与第二声场信息值确定对应的声场类别，比如用于根据所述第一 和第二确定单元确定的第一比值与第二比值确定对应的声场类别； 所述第一判断单元，用于判断所述第一声场信息值与第二声场信息值对应 的声场类别是否相同， 并发送判断结果， 比如用于判断所述第一比值与第二比 值对应的声场类别是否相同， 并发送判断结果； 所述第一设置单元，用于根据接收到第一判断单元发送的判断结果设置累 积互相关函数不同的加权系数； 第二调整单元，用于利用第一设置单元设置的不同加权系数对所述累积互 相关函数进行调整， 得到调整后的累积互相关函数。 另一以短时互相关函数为例的实施例： 所述提取单元具体包括：
的声场信息； 长时累积提取单元，用于从确定单元确定的累积互相关函数中提取短时互 相关函数之前的长时累积互相关函数的声场信息。 所述短时提取单元包括： 第三计算单元和第三确定单元， 所述第三计算单 元，用于计算第三部分延迟时刻短时互相关函数的和与第四部分延迟时刻短时 互相关函数的和，比如用于计算延迟大于等于 0的短时互相关函数的和与延迟 小于等于 0的短时互相关函数的和； 所述第三确定单元， 用于对第三部分延迟 时刻短时互相关函数的和与第四部分延迟时刻短时互相关函数的和进行操作， 获得第三声场信息值，比如用于确定延迟大于等于 0的短时互相关函数的和与 延迟小于等于 0的短时互相关函数的和的比，称为第三比值； 或者用于确定延 迟大于等于 0的短时互相关函数的和与延迟小于等于 0的短时互相关函数的和 的差， 称为第三差值； 所述长时累积提取单元包括： 第四计算单元和第四确定单元， 所述第四计 算单元，用于计算第三部分延迟时刻长时累积互相关函数的和与第四部分延迟 时刻长时累积互相关函数的和，比如用于计算延迟大于等于 0的长时累积互相 关函数的和与延迟小于等于 0 的长时累积互相关函数的和； 所述第四确定单 元，用于对第三部分延迟时刻长时累积互相关函数的和与第四部分延迟时刻长 时累积互相关函数的和进行操作，获得第四声场信息值， 比如用于确定延迟大 于等于 0的长时累积互相关函数的和与延迟小于等于 0的长时累积互相关函数 的和的比，称为第四比值； 或者用于确定延迟大于等于 0的长时累积互相关函 数的和与延迟小于等于 0的长时累积互相关函数的和的差， 称为第四差值。 所述调整单元包括： 第二声场类别确定单元、 第二判断单元、 第二设置单 元和第三调整单元， 其中， 所述第二声场类别确定单元， 用于根据所述第三和 第四确定单元确定的第三声场信息值与第四声场信息值确定对应的声场类别， 比如用于才 据所述第三和第四确定单元确定的第三比值与第四比值确定对应 的声场类别； 所述第二判断单元， 用于判断所述第三声场信息值与第四声场信 息值对应的声场类别是否相同， 并发送判断结果， 比如用于判断所述第三比值 与第四比值对应的声场类别是否相同， 并发送判断结果； 所述第二设置单元， 用于根据接收到第二判断单元发送的判断结果设置累积互相关函数不同的加 权系数； 所述第三调整单元， 用于利用第二设置单元设置的不同加权系数对所 述累积互相关函数进行调整， 得到调整后的累积互相关函数。 本发明实施例还提供一种编码器 14， 其结构示意图详见图 14， 包括： 通 道间延迟估计的装置 141和编码装置 142， 所述道间延迟估计的设备 141 , 用 于确定左右声道信号间的互相关函数及累积互相关函数，从所述互相关函数及 累积互相关函数中分别获取信号的声场信息，根据所述分别获取的声场信息获 得累积互相关函数调整信息，采用所述调整信息对所述累积互相关函数进行调 整，得到调整后的累积互相关函数，确定所述调整后的累积互相关函数中最大 值对应的时间为通道间延迟， 并将所述通道间延迟输出至编码装置； 所述编码 装置， 用于对接收到的通道间延迟进行编码， 并发送编码后的通道间延迟。 其中，在该编码器中， 所述通道间延迟估计的装置中各个单元的功能的作 用的实现过程详见上述方法中对应的实现过程， 在此不再赘述。 由上述实施例可知，在左右声道信号的声场发生变化时，通过从确定左右 声道信号间的互相关函数及累积互相关函数中提取声场变化的信息，根据声场 变化的信息可以正确估计出左右声道信号间的延迟，以便于对端根据接收到的 延迟正确对信号进行合成， 从而提高合成立体声的声场的稳定性。 通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到本发明 可借助软件加必需的通用硬件平台的方式来实现， 当然也可以通过硬件，但很 多情况下前者是更佳的实施方式。基于这样的理解，本发明的技术方案本质上 或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机 软件产品可以存储在存储介质中， 如 R0M/RAM、 磁碟、 光盘等， 包括若干指令 用以使得一台计算机设备（可以是个人计算机， 服务器， 或者网络设备等）执
以上所述仅是本发明的优选实施方式，应当指出，对于本技术领域的普通 技术人员来说， 在不脱离本发明原理的前提下， 还可以作出若干改进和润饰， 这些改进和润饰也应视为本发明的保护范围。

Claims

权 利 要 求
1、 一种通道间延迟估计的方法， 其特征在于， 包括：
从左右声道信号的互相关函数及累积互相关函数中分别获取信号的声场 信息；
根据所述分别获取的声场信息获得累积互相关函数调整信息；
采用所述调整信息对所述累积互相关函数进行调整，得到调整后的累积互 目关函数；
确定所述调整后的累积互相关函数中最大值对应的时间为通道间延迟。
2、 根据权利要求 1所述的方法， 其特征在于， 所述从左右声道信号的互 相关函数及累积互相关函数中分别获取信号的声场信息包括：
从左右声道信号的互相关函数中提取声场信息，以及提取互相关函数之前 的累积互相关函数的声场信息。
3、 根据权利要求 2所述的方法， 其特征在于， 所述从左右声道信号的互 相关函数中提取声场信息，以及从提取互相关函数之前的累积互相关函数的声 场信息具体包括：
对第一部分延迟时刻互相关函数的和与第二部分延迟时刻互相关函数的 和进行操作， 获得第一声场信息值；
对第一部分延迟时刻累积互相关函数的和与第二部分延迟时刻累积互相 关函数的和进行操作， 获得第二声场信息值。
4、 根据权利要求 3所述的方法， 其特征在于， 所述根据所述分别获取的 声场信息获得累积互相关函数调整信息包括：
根据所述第一声场信息值与第二声场信息值计算累积互相关函数的加权 系数；
所述采用所述调整信息对所述累积互相关函数进行调整，得到调整后的累 积互相关函数包括：
利用所述加权系数对所述累积互相关函数进行调整，得到调整后的累积互 目关函数。
5、 根据权利要求 3所述的方法， 其特征在于， 所述根据所述分别获取的 声场信息获得累积互相关函数调整信息包括：
确定所述第一声场信息值与第二声场信息值对应的声场类别；
判断所述第一声场信息值与第二声场信息值对应的声场类别是否相同，并 根据判断结果设置累积互相关函数的加权系数；
所述采用所述调整信息对所述累积互相关函数进行调整得到调整后累积 互相关函数包括：
利用设置的加权系数对所述累积互相关函数进行调整，得到调整后的累积 互 目关函数。
6、 根据权利要求 5所述的方法， 其特征在于，
所述确定第一声场信息值对应的声场类别包括：判断所述第一声场信息值 是否大于预设的第一阔值； 若是， 则设置第一声场信息值对应的声场类别的标 志为 1 ; 否则继续判断第一声场信息值是否大于等于第二阔值， 若是， 则设置 第一声场信息值对应的声场类别的标志为 0; 否则， 设置第一声场信息值对应 的声场类别的标志为 -1， 其中， 所述第二阔值小于第一阔值；
所述确定第二声场信息值对应的声场类别包括：判断所述第二声场信息值 是否大于预设的第一阔值； 若是， 则设置第二声场信息值对应的声场类别的标 志为 1 ; 否则继续判断第二声场信息值是否大于等于第二阔值， 若是， 则设置 第二声场信息值对应的声场类别的标志为 0; 否则， 设置第二声场信息值对应 的声场类别的标志为 -1， 其中， 所述第二阔值小于第一阔值。
7、 根据权利要求 1至 6任一项所述的方法， 其特征在于，
所述互相关函数为： 当前帧互相关函数， 所述互相关函数之前的累积互相 关函数为： 当前帧互相关函数之前的累积互相关函数；
所述互相关函数为： 短时互相关函数， 所述互相关函数之前的累积互相关 函数累积互相关函为： 长时互相关函数。
8、 根据权利要求 7所述的方法， 其特征在于， 所述互相关函数为短时互 相关函数时， 所述方法还包括： 从左右声道信号的互相关函数中提取信号类型；
根据所述分别获取的声场信息和信号类型获得累积互相关函数调整信息。
9、 根据权利要求 8所述的方法， 其特征在于， 所述根据所述分别获取的 声场信息和信号类型获得累积互相关函数调整信息包括：
根据所述确定的第一声场信息值、第二声场信息值及信号类型对应的值计 算累积互相关函数的加权系数；
所述采用所述调整信息对所述累积互相关函数进行调整得到调整后累积 互相关函数包括：
利用所述加权系数对所述累积互相关函数进行调整，得到调整后的累积互 相关函数。
10、 根据权利要求 1至 6任一项所述的方法， 其特征在于， 还包括： 判断所述声场信息是否发生变化，若是， 则执行根据所述声场信息调整累 积互相关函数的步骤。
11、 一种通道间延迟估计的装置， 其特征在于， 包括：
提取单元，用于从左右声道信号的所述互相关函数及累积互相关函数中分 别获取信号的声场信息；
调整单元，用于根据提取单元分别获取的所述声场信息获得累积互相关函 数调整信息， 采用所述调整信息对所述累积互相关函数进行调整，得到调整后 的累积互相关函数；
延迟估计单元，用于确定调整单元调整后的所述累积互相关函数中最大值 对应的时间为通道间延迟。
12、根据权利要求 11所述的装置， 其特征在于， 所述提取单元具体包括： 第一提取单元， 用于从左右声道信号的互相关函数中提取声场信息； 第二提取单元，用于从第一提取单元提取的互相关函数之前的累积互相关 函数中提取声场信息。
13、 根据权利要求 12所述的装置， 其特征在于，
所述第一提取单元包括： 第一计算单元，用于计算第一部分延迟时刻互相关函数的和与第二部分延 迟时刻互相关函数的和；
第一确定单元，用于对第一计算单元计算第一部分延迟时刻互相关函数的 和与第二部分延迟时刻互相关函数的和进行操作， 获得第一声场信息值； 所述第二提取单元包括：
第二计算单元，用于计算第一部分延迟时刻累积互相关函数的和与第二部 分延迟时刻累积互相关函数的和；
第二确定单元，用于对第二计算单元计算第一部分延迟时刻累积互相关函 数的和与第二部分延迟时刻累积互相关函数的和进行操作，获得第二声场信息 值。
14、 根据权利要求 13所述的装置， 其特征在于， 所述调整单元包括： 第 一系数计算单元与第一调整单元， 其中
所述第一系数计算单元，用于根据所述第一声场信息值与第二声场信息值 计算累积互相关函数的加权系数；
所述第一调整单元，用于利用所述第一系数计算单元计算的加权系数对所 述累积互相关函数进行调整， 得到调整后累积互相关函数。
15、 根据权利要求 13所述的装置， 其特征在于， 所述调整单元包括： 第 一声场类别确定单元、 第一判断单元、 第一设置单元与第二调整单元， 其中， 所述第一声场类别确定单元，用于根据所述第一和第二确定单元确定的第 一声场信息值与第二声场信息值确定对应的声场类别；
所述第一判断单元，用于判断所述第一声场信息值与第二声场信息值对应 的声场类别是否相同， 并发送判断结果；
所述第一设置单元，用于根据接收到第一判断单元发送的判断结果设置累 积互相关函数不同的加权系数；
第二调整单元，用于利用第一设置单元设置的不同加权系数对所述累积互 相关函数进行调整， 得到调整后的累积互相关函数。
16、 根据权利要求 11至 15任一项所述的装置， 其特征在于， 所述第一提取单元从左右声道信号的互相关函数包括：当前帧互相关函数 和短时互相关函数；
所述第二提取单元从第一提取单元提取的互相关函数之前的累积互相关 函数包括： 当前帧互相关函数之前的累积互相关函数和长时互相关函数。
17、 根据权利要求 14所述的装置， 其特征在于， 在所述互相关函数为短 时互相关函数时，
所述第一提取单元还包括： 第三提取单元， 用于从左右声道信号的互相关 函数中提取信号类型；
所述调整单元还包括：
第二系数计算单元，用于根据所述信号类型对应的值对所述第一系数计算 单元算累积互相关函数的加权系数再次进行加权计算，得到计算累积互相关函 数的加权系数；
第三调整单元，用于利用所述第二系数计算单元计算的加权系数对所述累 积互相关函数进行调整， 得到调整后累积互相关函数。
18、 根据权利要求 11至 15任一项所述的装置， 其特征在于， 还包括： 判断单元， 用于判断所述声场信息是否发生变化， 并将发生变化的判断结 果发送给调整单元。
19、根据权利要求 11至 15任一项所述的装置， 其特征在于， 所述通道间 延迟估计的装置集成在编码器中， 集成在用于通信的多用户定位设备中， 集成 在多声源位置判定设备中。
20、一种编码器，其特征在于， 包括：通道间延迟估计的装置和编码装置， 其中
所述通道间延迟估计的装置，用于从左右声道信号的互相关函数及累积互 相关函数中分别获取信号的声场信息，根据所述分别获取的声场信息获得累积 互相关函数调整信息， 采用所述调整信息对所述累积互相关函数进行调整，得 到调整后的累积互相关函数，确定所述调整后的累积互相关函数中最大值对应 的时间为通道间延迟， 并将所述通道间延迟输出至编码装置； 所述编码装置， 用于对接收到的通道间延迟进行编码， 并发送编码后的通 道间延迟。
PCT/CN2010/071314 2009-03-25 2010-03-25 通道间延迟估计的方法及其装置和编码器 Ceased WO2010108445A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP10755445.3A EP2413598B1 (en) 2009-03-25 2010-03-25 Method for estimating inter-channel delay and apparatus and encoder thereof
US13/245,303 US8417473B2 (en) 2009-03-25 2011-09-26 Method for estimating inter-channel delay and apparatus and encoder thereof

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200910129492.3 2009-03-25
CN2009101294923A CN101848412B (zh) 2009-03-25 2009-03-25 通道间延迟估计的方法及其装置和编码器

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/245,303 Continuation US8417473B2 (en) 2009-03-25 2011-09-26 Method for estimating inter-channel delay and apparatus and encoder thereof

Publications (1)

Publication Number Publication Date
WO2010108445A1 true WO2010108445A1 (zh) 2010-09-30

Family

ID=42772838

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2010/071314 Ceased WO2010108445A1 (zh) 2009-03-25 2010-03-25 通道间延迟估计的方法及其装置和编码器

Country Status (4)

Country Link
US (1) US8417473B2 (zh)
EP (1) EP2413598B1 (zh)
CN (1) CN101848412B (zh)
WO (1) WO2010108445A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110301962A1 (en) * 2009-02-13 2011-12-08 Wu Wenhai Stereo encoding method and apparatus

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3048818B1 (en) * 2015-01-20 2018-10-10 Yamaha Corporation Audio signal processing apparatus
TWI714046B (zh) * 2018-04-05 2020-12-21 弗勞恩霍夫爾協會 用於估計聲道間時間差的裝置、方法或計算機程式
CN113948098B (zh) 2020-07-17 2025-06-10 华为技术有限公司 一种立体声音频信号时延估计方法及装置
CN112383855A (zh) * 2020-11-04 2021-02-19 北京安声浩朗科技有限公司 蓝牙耳机充电盒、录音方法及计算机可读存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050179777A1 (en) * 2004-02-18 2005-08-18 Ferguson Kevin M. Relative channel delay measurement
CN1860526A (zh) * 2003-09-29 2006-11-08 皇家飞利浦电子股份有限公司 音频信号编码
WO2008016097A1 (en) * 2006-08-04 2008-02-07 Panasonic Corporation Stereo audio encoding device, stereo audio decoding device, and method thereof

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4890065A (en) * 1987-03-26 1989-12-26 Howe Technologies Corporation Relative time delay correction system utilizing window of zero correction
US5434948A (en) 1989-06-15 1995-07-18 British Telecommunications Public Limited Company Polyphonic coding
US6973184B1 (en) 2000-07-11 2005-12-06 Cisco Technology, Inc. System and method for stereo conferencing over low-bandwidth links
US7006587B1 (en) * 2001-11-20 2006-02-28 Cisco Technolgy, Inc. Preamble aided synchronization
JP5017121B2 (ja) 2004-11-30 2012-09-05 アギア システムズ インコーポレーテッド 外部的に供給されるダウンミックスとの空間オーディオのパラメトリック・コーディングの同期化
JP5025485B2 (ja) 2005-10-31 2012-09-12 パナソニック株式会社 ステレオ符号化装置およびステレオ信号予測方法
KR101329308B1 (ko) 2006-11-22 2013-11-13 삼성전자주식회사 오디오 신호의 저주파 성분 보강 방법 및 그 장치, 오디오신호의 기본 주파수 계산 방법 및 그 장치
CN101681625B (zh) * 2007-06-08 2012-11-07 杜比实验室特许公司 用于从两个输入的音频信号获得两个环绕声音频通道的方法和设备
JP4455614B2 (ja) * 2007-06-13 2010-04-21 株式会社東芝 音響信号処理方法及び装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1860526A (zh) * 2003-09-29 2006-11-08 皇家飞利浦电子股份有限公司 音频信号编码
US20050179777A1 (en) * 2004-02-18 2005-08-18 Ferguson Kevin M. Relative channel delay measurement
WO2008016097A1 (en) * 2006-08-04 2008-02-07 Panasonic Corporation Stereo audio encoding device, stereo audio decoding device, and method thereof

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110301962A1 (en) * 2009-02-13 2011-12-08 Wu Wenhai Stereo encoding method and apparatus
US8489406B2 (en) * 2009-02-13 2013-07-16 Huawei Technologies Co., Ltd. Stereo encoding method and apparatus

Also Published As

Publication number Publication date
EP2413598A4 (en) 2012-02-08
CN101848412A (zh) 2010-09-29
US20120016632A1 (en) 2012-01-19
EP2413598A1 (en) 2012-02-01
CN101848412B (zh) 2012-03-21
EP2413598B1 (en) 2014-11-12
US8417473B2 (en) 2013-04-09

Similar Documents

Publication Publication Date Title
US12334084B2 (en) Multi-channel signal encoding method and encoder
US10311881B2 (en) Determining the inter-channel time difference of a multi-channel audio signal
JP6841900B2 (ja) マルチチャネル信号を符号化する方法及びエンコーダ
US11380337B2 (en) Method and apparatus for increasing stability of an inter-channel time difference parameter
CN106373589B (zh) 一种基于迭代结构的双耳混合语音分离方法
WO2011006369A1 (zh) 一种改进的离散余弦变换域音频丢帧补偿器和补偿方法
CN102292769B (zh) 一种立体声编码方法和装置
WO2022012629A1 (zh) 一种立体声音频信号时延估计方法及装置
CN110114828B (zh) 使用比率作为分离特征来分解音频信号的装置和方法
CN110114827A (zh) 使用可变阈值来分解音频信号的装置和方法
CN118280375A (zh) 用于多声道音频编码的方法和设备
KR20190085988A (ko) 상관해제 필터들의 적응적 제어를 위한 방법 및 장치
CN101848412B (zh) 通道间延迟估计的方法及其装置和编码器
JP5468020B2 (ja) 音響信号復号装置及びバランス調整方法
WO2010127489A1 (zh) 检测信号延迟的方法、检测装置及编码器
CN105981412A (zh) 用于基于至少第一对空间脉冲响应估计总体混合时间的装置和方法以及对应的计算机程序
US10063907B1 (en) Differential audio-video synchronization
TW201921338A (zh) 時間偏移估計
CN103403801A (zh) 参数多通道编码器和解码器
WO2024160859A1 (en) Refined inter-channel time difference (itd) selection for multi-source stereo signals
WO2024202972A1 (ja) チャネル間時間差推定装置、及び、チャネル間時間差推定方法
WO2017193550A1 (zh) 多声道信号的编码方法和编码器
WO2017193549A1 (zh) 多声道信号的编码方法和编码器
HK40002235A (zh) 多声道信号的编码方法和编码器

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10755445

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2010755445

Country of ref document: EP