AGC control device, AGC control method, and program
The AGC control device adapts to individual head shapes and content-specific frequency characteristics, optimizing audio reproduction by adjusting coefficients and reducing noise in sound image localization systems.
Patent Information
- Application Number
- JP2023080576
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-16
- Publication Date
- 2024-11-28
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing AGC control systems in sound image localization devices fail to account for individual head shape variations and content-specific frequency characteristics, leading to suboptimal audio reproduction.
An AGC control device and method that includes filter processing, maximum value detection, and adaptive smoothing to adjust AGC coefficients based on individual head measurements and content scenes, with noise detection and parameter adjustment to optimize audio output.
Enhances sound image localization by providing optimal AGC control tailored to individual head shapes and content characteristics, reducing noise and improving audio quality.
Smart Images

Figure 2024164882000001_ABST
Abstract
Description
[Technical Field]
[0001] The present disclosure relates to an AGC control device, an AGC control method, and a program. [Background technology]
[0002] One sound image localization technology is out-of-head localization, which uses headphones to localize a sound image outside the listener's head. The out-of-head localization technology in Patent Document 1 localizes a sound image outside the head by canceling the characteristics from the headphones to the ears and applying the characteristics of four lines from the stereo speakers to the ears.
[0003] In out-of-head localization playback, measurement signals (impulse sounds, etc.) emitted from two-channel (hereinafter referred to as ch) speakers are recorded by microphones (hereinafter referred to as mics) placed at the listener's (user's) ears. Then, a processing device creates a filter based on the picked-up signal obtained from the impulse response. This creates a filter that corresponds to the spatial acoustic transfer characteristics from the speaker to the microphone position in the ear canal. Out-of-head localization playback can be achieved by convolving the created filter with the two-channel audio signal.
[0004] Furthermore, in order to generate a filter to cancel the characteristics from the headphones to the ear, the characteristics from the headphones to the eardrum (also called the ear canal transfer function ECTF or ear canal transfer characteristics) are measured using a microphone placed in the listener's own ear.
[0005] Patent Document 2 discloses an automatic gain control (AGC) device. This AGC device divides an input audio signal using a band-splitting filter. The AGC device converts the divided signals into absolute values and multiplies the absolute values by weighting data. The multiplied values are added, and the gain is controlled based on the added value. [Prior art documents] [Patent documents]
[0006] [Patent Document 1] Japanese Patent Application Publication No. 2019-62430 [Patent Document 2] Japanese Patent Application Laid-Open No. 2002-162029 Summary of the Invention [Problem to be solved by the invention]
[0007] In such out-of-head localization processing devices, filtering is performed on input signals, such as by convolving various filters. The filtered signals are then output from headphones or the like. The frequency characteristics of these filters vary depending on the shape of an individual's head or pinna, and therefore the optimal AGC coefficients also vary. Furthermore, frequency characteristics vary depending on the content scene, so it is desirable to adjust the optimal AGC coefficients depending on the content scene and the individual's filters.
[0008] The present disclosure has been made in consideration of the above points, and provides an AGC control device, an AGC control method, and a program for performing appropriate AGC control. [Means for solving the problem]
[0009] The AGC control device of this embodiment includes a filter processing unit that generates a filtered signal by performing filtering on an input signal using a filter, a maximum value detection unit that divides the filtered signal over a predetermined period of content into frames and detects a maximum value for each frame, a coefficient calculation unit that calculates an AGC (Auto Gain Control) coefficient for each frame based on the maximum value, a smoothing unit that calculates a smoothing gain by smoothing the AGC coefficient for each frame using a smoothing parameter, a multiplication unit that generates a multiplied signal by multiplying the filtered signal by the smoothing gain, a noise detection unit that detects whether noise is present in the multiplied signal, and a smoothing parameter adjustment unit that changes the smoothing parameter when the noise is detected.
[0010] The AGC control method according to this embodiment includes the steps of: generating a filtered signal by performing filtering on an input signal using a filter; dividing the filtered signal over a predetermined period of content into frames and detecting a maximum value for each frame; calculating an AGC coefficient for each frame according to the maximum value; smoothing the AGC coefficient using a smoothing parameter to calculate a smoothing gain; generating a multiplied signal by multiplying the filtered signal by the smoothing gain; detecting whether noise is present in the multiplied signal; and, if the noise is detected, changing the smoothing parameter.
[0011] The program according to this embodiment is a program for causing a computer to execute an AGC control method, and the AGC control method includes the steps of: generating a filtered signal by performing filtering using a filter on an input signal during a predetermined period of content; dividing the filtered signal into frames and detecting a maximum value for each frame; calculating an AGC coefficient for each frame according to the maximum value; smoothing the AGC coefficient using a smoothing parameter to calculate a smoothing gain; generating a multiplied signal by multiplying the filtered signal by the smoothing gain; detecting whether noise is occurring in the multiplied signal; and changing the smoothing parameter if the noise is detected. [Effects of the Invention]
[0012] According to the present disclosure, it is possible to provide an AGC control device, an AGC control method, and a program that are capable of performing appropriate AGC control. [Brief explanation of the drawings]
[0013] [Figure 1] 1 is a block diagram showing an out-of-head localization processing device according to an embodiment of the present invention; [Figure 2] FIG. 1 is a block diagram illustrating a configuration of a processing system for performing AGC processing. [Figure 3] FIG. 10 is a block diagram schematically illustrating a configuration for adjusting a smoothing parameter. [Figure 4] FIG. 10 is a diagram schematically illustrating a waveform in which noise occurs between frames. [Figure 5] FIG. 10 is a diagram schematically illustrating a waveform in which noise occurs. [Figure 6] 1 is a flowchart illustrating an AGC control method. [Figure 7] 1 is a flowchart illustrating an AGC control method. DETAILED DESCRIPTION OF THE INVENTION
[0014] An overview of the sound image localization processing according to this embodiment will be described. The out-of-head localization processing according to this embodiment is performed using spatial acoustic transfer characteristics and ear canal transfer characteristics. The spatial acoustic transfer characteristics are the transfer characteristics from a sound source such as a speaker to the ear canal. The ear canal transfer characteristics are the transfer characteristics from a speaker unit of headphones or earphones to the eardrum. In this embodiment, the spatial acoustic transfer characteristics are measured when headphones or earphones are not worn, and the ear canal transfer characteristics are measured when headphones or earphones are worn, and the out-of-head localization processing is realized using these measurement data.
[0015] The out-of-head localization processing according to this embodiment is executed by a user terminal such as a personal computer, a smartphone, or a tablet PC. The user terminal is an information processing device having processing means such as a processor, storage means such as a memory or a hard disk, display means such as an LCD monitor, and input means such as a touch panel, buttons, a keyboard, or a mouse. The user terminal may have a communication function for transmitting and receiving data. Furthermore, output means (output unit) having headphones or earphones is connected to the user terminal. The connection between the user terminal and the output means may be wired or wireless.
[0016] In the following explanation, a 2ch out-of-head localization processing device will be described as a sound field reproduction device using an AGC control device (also referred to as an AGC processing unit). Of course, the AGC control device can be applied to a multi-channel out-of-head localization processing device or other sound field reproduction devices. When the AGC control device is applied to a multi-channel out-of-head localization processing device, the out-of-head localization processing unit 101 (see FIG. 1) generates 2ch signals by convolving an out-of-head localization processing filter with each of the multi-channel input signals. Then, the AGC control device performs AGC processing on the 2ch signals output from the out-of-head localization processing unit.
[0017] (Extracranial stereotaxic processing device) 1 shows a block diagram of an out-of-head localization processing device 100, which is an example of a sound field reproduction device according to this embodiment. The out-of-head localization processing device 100 reproduces a sound field for a user U wearing headphones 43. To this end, the out-of-head localization processing device 100 performs sound image localization processing on Lch and Rch stereo input signals FL, FR. The Lch and Rch stereo input signals FL, FR are analog audio signals output from a DVD (Digital Versatile Disc) player, a BD (Blu-ray Disc: registered trademark) player, or the like, or digital audio data such as mp3 (MPEG Audio Layer-3).
[0018] Note that audio signals or digital audio data are collectively referred to as input signals. That is, the Lch and Rch stereo input signals FL and FR are input signals. The stereo input signals FL and FR are audio signals contained in various content such as movies and dramas. The content may be divided into multiple chapters. Each chapter is also called a scene. A silent period may be provided between chapters. For example, a silent period may occur when a scene changes.
[0019] In this embodiment, the out-of-head localization processing device 100 performs arithmetic processing for appropriately processing the filters. The arithmetic processing unit of the out-of-head localization processing device 100 is a personal computer (PC), a tablet terminal, a smartphone, or the like, and is equipped with a memory and a processor. The memory stores processing programs, various parameters, measurement data, and the like. The processor executes the processing programs stored in the memory. Each process is performed by the processor executing the processing programs. The processor may be, for example, a CPU (Central Processing Unit), FPGA (Field-Programmable Gate Array), DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), or GPU (Graphics Processing Unit).
[0020] The out-of-head localization processing device 100 is not limited to a single physical device, and some of the processing may be performed by different devices. For example, some of the processing may be performed by a smartphone or the like, and the remaining processing may be performed by a DSP (Digital Signal Processor) built into the headphones 43.
[0021] The out-of-head localization processing device 100 includes a spatial acoustic filter processing unit 10, an inverse filter unit 41, an inverse filter unit 42, AGC processing units 45L and 45R (collectively referred to as an AGC processing unit 45), and headphones 43. The spatial acoustic filter processing unit 10, the inverse filter unit 41, the inverse filter unit 42, and the AGC processing unit 45 can be specifically realized by a processor or the like. The spatial acoustic filter processing unit 10 and the inverse filter units 41 and 42 are collectively referred to as an out-of-head localization processing unit 101. In other words, the out-of-head localization processing unit 101 includes the spatial acoustic filter processing unit 10 and the inverse filter units 41 and 42.
[0022] The spatial acoustic filter processing unit 10 includes convolution calculation units 11-12, 21-22 that store spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs, and adders 24 and 25. The convolution calculation units 11-12, 21-22 perform convolution processing using the spatial acoustic transfer characteristics. Stereo input signals FL and FR from a DVD player or the like are input to the spatial acoustic filter processing unit 10. The spatial acoustic transfer characteristics are set in the spatial acoustic filter processing unit 10. The spatial acoustic filter processing unit 10 convolves the stereo input signals FL and FR of each channel with a filter of the spatial acoustic transfer characteristics (hereinafter also referred to as a spatial acoustic filter). The spatial acoustic transfer characteristics may be head-related transfer functions (HRTFs) measured using the head or pinnae of the subject, or head-related transfer functions of a dummy head or a third party.
[0023] A set of four spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs is defined as a spatial acoustic transfer function. Data used for convolution in convolution calculation units 11, 12, 21, and 22 becomes a spatial acoustic filter. A spatial acoustic filter is generated by cutting out the spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs with a predetermined filter length.
[0024] The spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs are obtained in advance by impulse response measurement or the like. For example, a user U wears a microphone on each of his or her left and right ears. Left and right speakers placed in front of the user U output impulse sounds for impulse response measurement. Then, a measurement signal such as the impulse sound output from the speakers is picked up by the microphone. The spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs are obtained based on the signal picked up by the microphone. The spatial acoustic transfer characteristic Hls between the left speaker and the left microphone, the spatial acoustic transfer characteristic Hlo between the left speaker and the right microphone, the spatial acoustic transfer characteristic Hro between the right speaker and the left microphone, and the spatial acoustic transfer characteristic Hrs between the right speaker and the right microphone are measured.
[0025] The convolution calculation unit 11 then convolves the Lch stereo input signal FL with a spatial acoustic filter that corresponds to the spatial acoustic transfer characteristic Hls. The convolution calculation unit 11 outputs the convolution calculation data to the adder 24. The convolution calculation unit 21 convolves the Rch stereo input signal FR with a spatial acoustic filter that corresponds to the spatial acoustic transfer characteristic Hro. The convolution calculation unit 21 outputs the convolution calculation data to the adder 24. The adder 24 adds the two convolution calculation data and outputs the result to the inverse filter unit 41.
[0026] The convolution calculation unit 12 convolves the left channel stereo input signal FL with a spatial acoustic filter corresponding to the spatial acoustic transfer characteristic Hlo. The convolution calculation unit 12 outputs the convolution calculation data to the adder 25. The convolution calculation unit 22 convolves the right channel stereo input signal FR with a spatial acoustic filter corresponding to the spatial acoustic transfer characteristic Hrs. The convolution calculation unit 22 outputs the convolution calculation data to the adder 25. The adder 25 adds the two convolution calculation data and outputs the result to the inverse filter unit 42.
[0027] Inverse filters Linv and Rinv that cancel headphone characteristics (characteristics between the headphone playback unit and microphone) are set in the inverse filter units 41 and 42. The inverse filter unit 41 stores the inverse filter Linv, and the inverse filter unit 42 stores the inverse filter Rinv. The inverse filters Linv and Rinv are then convolved with the playback signal (convolution signal) that has been processed by the spatial acoustic filter processing unit 10. The inverse filter unit 41 convolves the inverse filter Linv of the left-channel headphone characteristics with the left-channel signal from the adder 24. Similarly, the inverse filter unit 42 convolves the right-channel signal from the adder 25 with the inverse filter Rinv of the right-channel headphone characteristics. When headphones 43 are worn, the inverse filters Linv and Rinv cancel the characteristics from the headphone unit to the microphone. The microphone may be placed anywhere between the entrance of the ear canal and the eardrum.
[0028] Hereinafter, the signals subjected to out-of-head localization processing by the out-of-head localization processing unit 101 are referred to as filtered signals XL and XR. The filtered signals XL and XR are signals obtained by convolving a filter with the input signal. The filtered signal XL is the Lch signal. The filtered signal XR is the Rch signal.
[0029] The inverse filter unit 41 outputs the filtered signal XL of the left channel to the AGC processing unit 45L. The AGC processing unit 45L generates an output signal YL by performing AGC processing on the filtered signal XL. Specifically, the AGC processing unit 45L multiplies each sample of the filtered signal XL by a smoothing gain. The AGC processing unit 45L outputs the output signal YL to the left unit 43L of the headphones 43.
[0030] The inverse filter unit 42 outputs the filtered Rch signal XR to the AGC processing unit 45R. The AGC processing unit 45R generates an output signal YR by performing AGC processing on the filtered signal XR. Specifically, the AGC processing unit 45R multiplies each sample of the filtered signal XR by a smoothing gain. The AGC processing unit 45R outputs the output signal YR to the right unit 43R of the headphones 43. Note that, for convenience, FIG. 1 shows a configuration in which AGC processing units 45L and 45R are provided for the Lch and Rch, respectively, but in reality, as shown in FIG. 2, a single AGC processing unit 45 performs AGC processing on the Lch and Rch signals.
[0031] The user U is wearing headphones 43. The headphones 43 output an Lch output signal YL and an Rch output signal YR (hereinafter, the Lch output signal YL and the Rch output signal YR are also collectively referred to as stereo output signals) to the user U. That is, the headphones 43 reproduce content by outputting the stereo output signals YL and YR. This makes it possible to reproduce a sound image localized outside the head of the user U.
[0032] In this way, the out-head localization processing device 100 performs out-head localization processing using spatial acoustic filters corresponding to the spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs, and inverse filters Linv and Rinv of headphone characteristics. In the following description, the spatial acoustic filters corresponding to the spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs, and the inverse filters Linv and Rinv of headphone characteristics are collectively referred to as out-head localization processing filters. In the case of a 2-channel stereo playback signal, the out-head localization filters are composed of four spatial acoustic filters and two inverse filters. The out-head localization processing device 100 then performs convolution operation processing on the stereo playback signal using a total of six out-head localization filters, thereby executing out-head localization processing. It is preferable that the out-head localization filters be based on personal measurements of the user U. For example, the out-head localization filters are set based on sound signals picked up by microphones attached to the user U's ears.
[0033] In this way, the spatial acoustic filter and the headphone characteristic inverse filters Linv and Rinv are filters for audio signals. These filters are convoluted with the playback signals (stereo input signals FL and FR), so that the out-of-head localization processing device 100 performs out-of-head localization processing.
[0034] The AGC processing unit 45 performs AGC processing on the filtered signals XL and XR. The AGC processing unit 45 multiplies each sample of the filtered signals XL and XR by an AGC coefficient α. The AGC coefficient α varies depending on the sample. In other words, the AGC coefficient α changes over time. The filtered signals that have undergone AGC processing become the output signals YL and YR that are output from the headphones 43.
[0035] (System Configuration) Next, the configuration of a processing system for performing AGC processing will be described with reference to Fig. 2. Fig. 2 is a block diagram showing the overall configuration of a processing system 200 that performs AGC processing. The processing system 200 includes an AGC control unit 210, an AGC processing unit 45, a smoothing parameter storage unit 241, and an AGC parameter storage unit 242. The AGC processing unit 45 shown in Fig. 2 corresponds to the AGC processing unit 45L and AGC processing unit 45R in Fig. 1. In other words, the AGC processing unit 45L and AGC processing unit 45R in Fig. 1 function as one AGC processing unit 45.
[0036] As described above, the AGC processing unit 45 performs online processing while the user is viewing content. The AGC control unit 210 performs processing before the user views content. The AGC control unit 210 can perform offline processing in advance. The AGC control unit 210 is a processing device equipped with a processor, memory, etc., and performs processing described below. The processing device of the AGC control unit 210 and the out-of-head localization processing device 100 may be physically the same device or different devices.
[0037] The AGC control unit 210 sets parameters for calculating AGC coefficients. Specifically, the AGC control unit 210 calculates AGC coefficients for each frame and smoothing parameters for smoothing the AGC coefficients. The AGC parameter storage unit 242 stores AGC parameters for performing AGC processing. The smoothing parameter storage unit 241 stores parameters for smoothing. For example, the smoothing parameter is the cutoff frequency of an IIR filter.
[0038] The AGC control unit 210 includes a filter generation unit 211, a filter storage unit 212, and an analysis unit 213. The filter generation unit 211 generates filters used for out-of-head localization processing. When stereo input signals FL and FR are used, the filter generation unit 211 generates six filters as described above. The filter generation unit 211 generates spatial acoustic filters according to spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs, and inverse filters Linv and Rinv of headphone characteristics.
[0039] The filter generation unit 211 can generate an out-of-head localization filter based on an impulse response measurement performed while the user is wearing a microphone. The filter generation unit 211 generates a spatial acoustic filter by, for example, measuring an impulse response using a microphone and a speaker. The filter generation unit 211 generates inverse filters Linv, Rinv by measuring an impulse response using a microphone and headphones.
[0040] The filter generation unit 211 may generate an out-of-head localization filter suitable for the user from a database, etc. For example, the filter generation unit 211 may measure the ear canal transfer characteristics of the user and generate an out-of-head localization filter by matching the ear canal transfer characteristics with the ear canal transfer characteristics. Since the filter generation unit 211 can use known processing to generate a filter, detailed description thereof will be omitted.
[0041] The filter storage unit 212 stores the filter coefficients of the filters generated by the filter generation unit 211. The filter storage unit 212 stores spatial acoustic filters according to the spatial acoustic transfer characteristics Hls, Hlo, Hro, and Hrs, and inverse filters Linv and Rinv of the headphone characteristics. The filters stored in the filter storage unit 212 are the same data as the out-of-head localization filters shown in Fig. 1. In other words, the analysis unit 213 performs the analysis process described below using the same filters as when performing out-of-head localization listening.
[0042] The analysis unit 213 analyzes the input signals FL and FR to calculate smoothing parameters for setting AGC coefficients. Specifically, the analysis unit 213 performs out-of-head localization processing on the input signals FL and FR using filters stored in the filter storage unit 212, thereby generating filtered signals. Here, the analysis unit 213 performs out-of-head localization processing using the six out-of-head localization filters as described above. Then, the analysis unit 213 analyzes the filtered signals to generate smoothing parameters and AGC coefficients. The processing of the analysis unit 213 will be described later.
[0043] The smoothing parameter storage unit 241 stores the smoothing parameters in a memory etc. The AGC parameter storage unit 242 stores the AGC parameters in a memory etc. The AGC parameters are, for example, a threshold value vth, a half life, an attack time, and a release time.
[0044] Next, the configuration and processing of the analysis unit 213 will be described with reference to Fig. 3. Fig. 3 is a block diagram for explaining the configuration and processing of the analysis unit 213. The analysis unit 213 includes an input unit 401, an out-of-head localization processing unit 101, a maximum value detection unit 402, a coefficient calculation unit 403, a smoothing unit 404, a multiplication unit 405, a storage unit 406, a noise detection unit 408, and an analysis parameter storage unit 412.
[0045] Stereo input signals FL and FR are input to the input unit 401. As in FIG. 1 , the out-of-head localization processing unit 101 performs out-of-head localization processing on the stereo input signals FL and FR to generate filtered signals XL and XR. The out-of-head localization processing unit 101 performs out-of-head localization processing using six out-of-head localization filters. The out-of-head localization processing unit 101 outputs the filtered signals XL and XR to the maximum value detection unit 402 and the multiplication unit 405.
[0046] The maximum value detection unit 402 divides the filtered signals XL and XR into frames and detects the maximum value for each frame. The maximum value detection unit 402 divides the filtered signals XL and XR for one scene into multiple frames. The maximum value detection unit 402 divides the filtered signals XL and XR for a predetermined time into frames using a predetermined frame length and period. One frame contains multiple samples. For example, one frame contains 512 samples. The maximum value detection unit 402 then detects the maximum value for each frame. The maximum value detection unit 402 detects the maximum value vMaxLR for the number of frames.
[0047] Here, the maximum value detection unit 402 detects one maximum value vMaxLR from the filtered signals XL and XR of the Lch and Rch. That is, it detects one maximum value vMaxLR from the two filtered signals XL and XR in one frame.
[0048] For example, in one frame, the maximum value of the filtered left channel signal XL is taken as vMaxL, and the maximum value of the filtered right channel signal XR is taken as vMaxR. If the maximum value vMaxL of the left channel is equal to or greater than the maximum value vMaxR of the right channel, then the maximum value vMaxL of the left channel becomes the maximum value vMaxLR of the frame. If the maximum value vMaxL of the left channel is smaller than the maximum value vMaxR of the right channel, then the maximum value vMaxR of the left channel becomes the maximum value vMaxLR of the frame. In this way, the maximum value vMaxLR of each frame is calculated from the filtered left and right channel signals XL and XR.
[0049] The coefficient calculation unit 403 calculates an AGC coefficient based on the maximum value vMaxLR. The coefficient calculation unit 403 calculates the AGC coefficient by referring to the AGC parameters stored in the AGC parameter storage unit 242. The AGC coefficient is a value set for each frame. Here, the AGC coefficient set for each frame is also referred to as a frame coefficient. The AGC parameter storage unit 242 stores AGC parameters such as the threshold value vth, half life, attack time, and release time. These AGC parameters are stored in advance in the AGC parameter storage unit 242.
[0050] The threshold value vth is a value used to determine whether or not to compress the filtered signal using AGC processing. In other words, when the filtered signals XL and XR exceed the threshold value vth, the amplitude is compressed. The half-life indicates the time it takes for the maximum value in the i-th frame to decay to half. The attack time is a time constant that indicates the time it takes for compression to begin after the signal exceeds the threshold value vth. The release time is a time constant that indicates the time it takes for compression to end after the signal falls below the threshold value vth.
[0051] For example, the coefficient calculation unit 403 calculates the AGC coefficient α for each frame using the following equation (1): In the following equation (1), x=vMaxLR.
number
[0052] When the maximum value is equal to or less than the threshold value vth, the AGC coefficient α is 1. When the maximum value is greater than the threshold value vth, it is calculated from the AGC parameters. The AGC coefficient α is a frame coefficient set for each frame.
[0053] The smoothing unit 404 smoothes the AGC coefficient α for each frame to obtain a smoothed gain α s Calculate the smoothing gain α s is calculated for each sample included in the frame. Therefore, the smoothing gain α s can be a different value for each sample.
[0054] The smoothing unit 404 smoothes the AGC coefficient α using an IIR (Infinite Impulse Response) filter. The smoothing unit 404 reads out analysis parameters stored in the analysis parameter storage unit 412. The analysis parameters are parameters for adjusting the smoothing process in the smoothing unit 404.
[0055] The smoothing unit 404 smoothes the AGC coefficient α by referring to the analysis parameters. As a result, the smoothing unit 404 smooths the smoothing gain α for each sample included in the frame. s For example, if one frame contains 512 samples, 512 smoothing gains are calculated for each of the filtered signals of Lch and Rch in one frame.
[0056] Here, the IIR filter is an LPF (low-pass filter) having a cutoff frequency. Analysis parameter storage unit 412 stores a plurality of cutoff frequencies. The cutoff frequencies are, for example, 5 kHz, 4 kHz, 3 kHz, 2 kHz, 1 kHz, 800 Hz, 500 Hz, 300 Hz, 200 Hz, 100 Hz, and 50 Hz. Smoothing unit 404 reads one cutoff frequency from the plurality of cutoff frequencies stored in analysis parameter storage unit 412.
[0057] For example, the smoothing unit 404 reads out 1 kHz as the cutoff frequency. The smoothing unit 404 smoothes the AGC coefficient α using an IIR filter with a cutoff frequency of 1 kHz. The smoothing unit 404 smooths the smoothing gain α by inputting the AGC coefficient set for each frame to the IIR filter. s Calculate.
[0058] For example, assume that the AGC coefficient α(i) of the i-th frame is set to 0.7, and the AGC coefficient α(i+1) of the (i+1)-th frame is set to 0.6. Assume that one frame contains 512 samples. In this case, the smoothing unit 404 applies a smoothing gain α s The smoothing unit 404 performs smoothing processing using an IIR filter. This prevents a large difference in smoothing gain between the last sample of the i-th frame and the first sample of the (i+1)-th frame. For example, the smoothing gain value for the last sample of the i-th frame is set to 0.649, and the smoothing gain value for the first sample of the (i+1)-th frame is set to 0.650. The smoothing unit 404 can set the smoothing gain so that it shows a smooth curve between frames.
[0059] In this way, the smoothing unit 404 calculates the smoothing gain α s The smoothing unit 404 calculates the smoothing gain α s Multiply the filtered signals XL and XR by the smoothing gain α sis common to the filtered signals XL and XR. As mentioned above, the smoothing gain α s is different for each sample. And the smoothing gain α s is set.
[0060] The storage unit 406 stores the multiplied values obtained by multiplying each sample by the smoothing gain in a memory or the like. The storage unit 406 stores the multiplied value data for one scene as a multiplied signal. The storage unit 406 stores the multiplied signals for Lch and Rch in a memory or the like.
[0061] The noise detection unit 408 detects whether noise is present in the multiplied signal for one scene. The noise detection unit 408 detects noise using various methods. For example, as shown in FIG. 4, the noise detection unit 408 determines that noise is present when the difference in signal level between frames is equal to or greater than a predetermined value. In FIG. 4, the level difference between the i-th frame and the (i+1)-th frame is large. In other words, the signal level difference between the last sample of the i-th frame and the first sample of the (i+1)-th frame is large. If the signal level difference exceeds the predetermined value, the noise detection unit 408 detects the presence of noise. The noise detection unit 408 compares the amplitude difference between two consecutive samples with a predetermined value. If the amplitude difference exceeds the predetermined value, the noise detection unit 408 detects the presence of noise.
[0062] Alternatively, the noise detection unit 408 can detect noise from the difference in amplitude between the filtered signals XL and XR. For example, the noise detection unit 408 calculates the difference between the original signal and the multiplied signal. When noise occurs, the difference becomes large.
[0063] The original signal is the filtered signal before being multiplied by the smoothing gain. The difference value between the filtered signal and the multiplied signal is denoted as diff. For example, the difference value of the first sample is diff(1), and the difference value of the second sample is diff(2). The noise detection unit 408 calculates the amount of change (slope) of the difference value diff. For example, the noise detection unit 408 calculates the amount of change diff' from two consecutive difference values diff. For example, diff'(1) = diff(2) - diff(1). Then, as shown in FIG. 5, the noise detection unit 408 detects the occurrence of noise at the point where diff' suddenly increases.
[0064] The noise detection unit 408 outputs a detection signal indicating that noise has been detected to the smoothing parameter adjustment unit 413. When noise occurs, the smoothing parameter adjustment unit 413 adjusts the smoothing parameter. As described above, the smoothing parameter is the cutoff frequency. Therefore, the smoothing parameter adjustment unit 413 changes the cutoff frequency. For example, initially, the cutoff frequency is 1 kHz. Therefore, the noise detection unit 408 changes the cutoff frequency to 2 kHz.
[0065] Then, the analysis unit 213 analyzes the input signal again using the changed cutoff frequency. Specifically, the smoothing unit 404 generates a smoothing gain by performing smoothing processing using the changed cutoff frequency. Then, the multiplication unit 405 generates a multiplied signal by multiplying the filter-processed signal by the smoothing gain obtained by the smoothing processing. The noise detection unit determines whether noise is occurring in the multiplied signal. The smoothing parameter adjustment unit 413 changes the smoothing parameter until noise no longer occurs. In other words, the smoothing parameter adjustment unit 413 adjusts the smoothing parameter so that noise no longer occurs.
[0066] Furthermore, it is preferable that the smoothing parameter adjustment unit 413 alternately changes the cutoff frequency from the initial cutoff frequency to the higher frequency side and the lower frequency side. For example, assume that the initial cutoff frequency is 1 kHz. If noise occurs at a cutoff frequency of 1 kHz, the smoothing parameter adjustment unit 413 changes the cutoff frequency to 2 kHz, which is higher than 1 kHz. If noise still occurs at a cutoff frequency of 2 kHz, the smoothing parameter adjustment unit 413 changes the cutoff frequency to 800 Hz, which is lower than 1 kHz. If noise still occurs at a cutoff frequency of 800 Hz, the smoothing parameter adjustment unit 413 changes the cutoff frequency to 3 kHz, which is higher than 1 kHz. If noise still occurs at a cutoff frequency of 3 kHz, the smoothing parameter adjustment unit 413 changes the cutoff frequency to 500 Hz, which is lower than 1 kHz.
[0067] The smoothing parameter adjustment unit 413 alternately changes the cutoff frequency to a value higher than the initial value and a value lower than the initial value until noise is no longer detected. In this way, an appropriate smoothing parameter can be quickly found. The smoothing parameter adjustment unit 413 then stores the cutoff frequency at which noise is not detected in a memory or the like as a smoothing parameter. The smoothing parameter is determined for each scene. In other words, if the content has multiple scenes (chapters), the smoothing parameter adjustment unit 413 sets a smoothing parameter for each scene at which noise is not detected. The smoothing parameter for the nth chapter is denoted as smoothing parameter [n], where n is an integer greater than or equal to 1.
[0068] The smoothing parameter adjustment unit 413 outputs the adjusted smoothing parameters to the smoothing parameter storage unit 241 in FIG. 2. The smoothing parameter storage unit 241 stores the adjusted smoothing parameters in a memory or the like. The smoothing parameter storage unit 241 stores the smoothing parameters for each chapter. As described above, the smoothing parameter is the cutoff frequency of the LPF. Therefore, a different cutoff frequency is set for each chapter in the smoothing parameter storage unit 241.
[0069] 2, the following describes the configuration and processing of the AGC processing unit 45. The AGC processing unit 45 includes a signal input unit 451, a maximum value detection unit 452, a coefficient calculation unit 453, a smoothing unit 454, a multiplication unit 455, and a signal output unit 457.
[0070] The signal input unit 451 receives filtered signals XL and XR that have been subjected to out-of-head localization processing (see FIG. 1). The maximum value detection unit 452 detects the maximum value vMaxLR of the filtered signals XL and XR for each frame. The coefficient calculation unit 453 calculates an AGC coefficient α for each frame based on the maximum value vMaxLR.
[0071] The processing in the maximum value detection unit 452 and the coefficient calculation unit 453 is similar to that in the maximum value detection unit 402 and the coefficient calculation unit 403, and therefore a description thereof will be omitted. Furthermore, the coefficient calculation unit 453 acquires AGC parameters such as the threshold value vth, half life, attack time, and release time from the AGC parameter storage unit 242. These AGC parameters have the same values as the AGC parameters used in the coefficient calculation unit 403. Therefore, the AGC coefficient α calculated in the coefficient calculation unit 453 is the same value as the AGC coefficient α calculated in the coefficient calculation unit 403.
[0072] The smoothing unit 454 performs a smoothing process on the AGC coefficient for each frame to obtain a smoothing gain α s Here, the smoothing unit 454 performs the same smoothing process as the smoothing unit 404. The smoothing unit 454 smooths the AGC coefficient α using an IIR filter. The IIR filter is an LPF having a cutoff frequency.
[0073] Furthermore, the smoothing unit 454 reads out the cutoff frequency as a smoothing parameter from the smoothing parameter storage unit 241. Therefore, the smoothing unit 454 smoothes the AGC coefficients using the cutoff frequency that has been adjusted so that noise is not detected. Therefore, the smoothing unit 454 smooths the AGC coefficients using the smoothing gain α s The smoothing gain α sis a value set for each sample. The smoothing unit 454 uses the smoothing gain α s is output to the multiplication unit 455.
[0074] The multiplication unit 455 calculates the smoothing gain α s The multiplication unit 455 multiplies the filtered signal by the filtered signal to generate a multiplied signal. The processing by the multiplication unit 455 is similar to the processing by the multiplication unit 405. The multiplication unit 455 outputs the multiplied signal to the signal output unit 457. The signal output unit 457 outputs the multiplied signals as output signals YL and YR to the headphones 43 (see FIG. 1). As a result, the left unit 43L of the headphones 43 reproduces the Lch output signal YL, and the right unit 43R reproduces the Rch output signal YR.
[0075] Therefore, the user U can listen to a stereo output signal that has been appropriately AGC-processed. When performing out-of-head localization playback processing, the out-of-head localization processing device 100 performs out-of-head localization processing on the input signal of content that is input in real time. Furthermore, the AGC processing unit 45 can perform optimal AGC processing for each scene using preset AGC parameters and predetermined smoothing parameters. The smoothing parameters are set according to the content and the out-of-head localization filter. Therefore, the AGC processing unit 45 can perform AGC processing that is suitable for the content and the out-of-head localization filter. Since the optimal smoothing gain can be set according to the content and the user to prevent noise from occurring, the AGC processing unit 45 can perform AGC processing. This can improve the out-of-head localization listening effect.
[0076] 6 and 7 are flowcharts showing the AGC control method. First, input signals FL and FR are input to the input unit 401 (S101). The out-of-head localization processing unit 101 performs out-of-head localization processing on the input signals FL and FR using an out-of-head localization filter (S102). As a result, filtered signals XL and XR are generated.
[0077] The maximum value detection unit 402 divides the filtered signals XL and XR into frames (S103). The maximum value detection unit 402 detects the maximum value vMaxLR of each frame (S104). The coefficient calculation unit 403 sets AGC parameters (S105). Here, the coefficient calculation unit 403 reads AGC parameters such as the threshold value vth from the AGC parameter storage unit 242.
[0078] The coefficient calculation unit 403 calculates the AGC coefficient α from the maximum value vMaxLR using the AGC parameters (S106). The coefficient calculation unit 403 calculates the AGC coefficient α for each frame. The smoothing unit 404 acquires analysis parameters (S107). The smoothing unit 404 reads out analysis parameters such as the cutoff frequency from the analysis parameter storage unit 412. In the first process, the coefficient calculation unit 403 acquires an initial value of the cutoff frequency (for example, 1 kHz) as the analysis parameter.
[0079] The smoothing unit 404 calculates the smoothing gain α from the AGC coefficient α using the analysis parameters. s Here, the smoothing unit 404 smoothes the AGC coefficients using an IIR filter having a cutoff frequency. The smoothing unit 404 calculates a smoothing gain α s The smoothing gain α can be calculated. s is common to both the left and right channels.
[0080] The multiplication unit 405 multiplies the filter-processed signal by the smoothing gain (S109). That is, for each sample, the smoothing gain α s and the filtered signals XL and XR are multiplied. This generates a time-series multiplied signal. The storage unit 406 stores the multiplied signal (S110).
[0081] The noise detection unit 408 detects whether noise has occurred in the multiplication signal (S111). The noise detection unit 408 can detect noise using various noise detection processes. Then, the noise detection unit 408 determines whether noise has occurred (S112). If noise has occurred (YES in S112), the smoothing parameter adjustment unit 413 changes the smoothing parameter (S113). Then, the process returns to step S104, and the analysis unit 213 repeats the process.
[0082] If no noise is occurring (NO in S112), the smoothing parameter adjustment unit 413 determines a smoothing parameter [n] for the n-th chapter (S114). The analysis unit 213 determines whether or not analysis of the final chapter is complete (S115). If analysis of the final chapter is not complete (NO in S115), the process proceeds to processing of the next chapter (S116). Then, the process returns to step S103, and the analysis unit 213 repeats the process. That is, the analysis unit 213 determines a smoothing parameter for the next chapter. If analysis of the final chapter is complete (YES in S115), the analysis unit 213 ends the process.
[0083] In this way, the analysis unit 213 can determine an appropriate smoothing parameter for each chapter, and the out-of-head localization processing device 100 can therefore play back each chapter appropriately.
[0084] The smoothing process in the smoothing unit 404 is not limited to the process using an IIR filter. It may also be the process using an FIR filter, a moving average filter, or the like, as long as it is an LPF with a cutoff frequency. Furthermore, before noise analysis, a low-pass filter or a band-limiting filter may be used to limit the band in which noise analysis is performed. The method of calculating the AGC coefficient is not limited to the feedforward type or square root type, and a feedback type or series expansion type may also be used.
[0085] In the above description, an example has been described in which an out-of-head localization filter is convolved with the input signals FL and FR, but the filter convolved with the input signals is not limited to an out-of-head localization filter. In other words, a signal convolved with a filter other than an out-of-head localization filter may also be used as the filtered signal. The above method can be used for audio signal processing using various filters, such as stereo expansion, sound image localization, or noise cancellation. When playing back content, AGC processing can be performed appropriately based on a preset filter.
[0086] Some or all of the above processes may be executed by a computer program. The above-described program can be stored and provided to a computer using various types of non-transitory computer-readable media. Non-transitory computer-readable media include various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic storage media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical storage media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). The program may also be provided to a computer by various types of transitory computer-readable media. Examples of transitory computer-readable media include electrical signals, optical signals, and electromagnetic waves. The transitory computer-readable media can provide the program to a computer via a wired communication path such as an electric wire or optical fiber, or via a wireless communication path.
[0087] The invention made by the inventor has been specifically described above based on an embodiment, but it goes without saying that the present invention is not limited to the above embodiment and can be modified in various ways without departing from the gist of the invention. [Explanation of symbols]
[0088] U User 1 Person to be measured 2L Left microphone 2R Right microphone 5L Left speaker 5R Right speaker 9L left ear 9R Right ear 10 Spatial audio filter processing section 11 Convolution operation unit 12 Convolution operation unit 21 Convolution operation unit 22 Convolution operation unit 24 Adder 25 Adder 41 Inverse filter section 42 Inverse filter section 43 Headphones 45 AGC processing unit 100 Extracranial stereotaxic processing device 101 Out-of-head localization processing unit 210 AGC control section 211 Filter Generation Unit 212 Filter storage section 213 Analysis Department 241 Smoothing parameter storage 242 AGC parameter storage section 401 Input section 402 Maximum value detector 403 Coefficient calculation unit 404 Smoothing section 405 Multiplication section 406 Storage section 408 Noise detection unit 412 Analysis parameter storage section 413 Smoothing parameter adjustment unit 452 Maximum value detector 453 Coefficient Calculation Unit 454 Smoothing section 455 Multiplication Unit
Claims
1. a filtering unit that performs filtering on an input signal using a filter to generate a filtered signal; a maximum value detection unit that divides the filtered signal during a predetermined period of the content into frames and detects a maximum value for each frame; a coefficient calculation unit that calculates an AGC (Auto Gain Control) coefficient for each frame based on the maximum value; a smoothing unit that calculates a smoothing gain by smoothing the AGC coefficient for each frame using a smoothing parameter; a multiplication unit that generates a multiplied signal by multiplying the filtered signal by the smoothing gain; a noise detection unit that detects whether or not noise is generated in the multiplication signal; and a smoothing parameter adjustment unit that changes the smoothing parameter when the noise is detected.
2. the smoothing unit performs a smoothing process on the AGC coefficient using a low-pass filter having a cutoff frequency; 2. The AGC control device of claim 1, wherein the smoothing parameter is the cutoff frequency of the low pass filter.
3. A plurality of cutoff frequencies are set as the smoothing parameters, 3. The AGC control device according to claim 2, wherein the smoothing parameter adjustment unit alternately changes the cutoff frequency to a value higher and a value lower than an initial value until the noise is no longer detected.
4. filtering an input signal using a filter to generate a filtered signal; splitting the filtered signal into frames during a predetermined period of the content and detecting a maximum value for each frame; calculating an AGC coefficient for each frame according to the maximum value; calculating a smoothing gain by smoothing the AGC coefficient using a smoothing parameter; multiplying the filtered signal by the smoothing gain to generate a multiplied signal; detecting whether noise is present in the multiplied signal; and changing the smoothing parameter if the noise is detected.
5. A program for causing a computer to execute an AGC control method, The AGC control method generating a filtered signal by filtering an input signal during a predetermined period of the content using a filter; splitting the filtered signal into frames and detecting a maximum value for each frame; calculating an AGC coefficient for each frame according to the maximum value; calculating a smoothing gain by smoothing the AGC coefficient using a smoothing parameter; multiplying the filtered signal by the smoothing gain to generate a multiplied signal; detecting whether noise is present in the multiplied signal; and modifying the smoothing parameter if the noise is detected.
Citation Information
Patent Citations
Pre-heating timer control device
JP2002162029A
Processing device, processing method, and program
JP2019062430A
