WO2010013941A2 - An apparatus for processing an audio signal and method thereof - Google Patents

An apparatus for processing an audio signal and method thereof Download PDF

Info

Publication number
WO2010013941A2
WO2010013941A2 PCT/KR2009/004220 KR2009004220W WO2010013941A2 WO 2010013941 A2 WO2010013941 A2 WO 2010013941A2 KR 2009004220 W KR2009004220 W KR 2009004220W WO 2010013941 A2 WO2010013941 A2 WO 2010013941A2
Authority
WO
WIPO (PCT)
Prior art keywords
gain
signal
fast
slow
noise
Prior art date
Application number
PCT/KR2009/004220
Other languages
French (fr)
Other versions
WO2010013941A3 (en
Inventor
Hyen O Oh
Jong Ha Moon
Alexis Favrot
Christof Faller
Joon Il Lee
Myung Hoon Lee
Yang Won Jung
Original Assignee
Lg Electronics Inc.
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 Lg Electronics Inc. filed Critical Lg Electronics Inc.
Priority to CN2009801148963A priority Critical patent/CN102016994B/en
Publication of WO2010013941A2 publication Critical patent/WO2010013941A2/en
Publication of WO2010013941A3 publication Critical patent/WO2010013941A3/en

Links

Classifications

    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03GCONTROL OF AMPLIFICATION
    • H03G9/00Combinations of two or more types of control, e.g. gain control and tone control
    • H03G9/02Combinations of two or more types of control, e.g. gain control and tone control in untuned amplifiers
    • H03G9/025Combinations of two or more types of control, e.g. gain control and tone control in untuned amplifiers frequency-dependent volume compression or expansion, e.g. multiple-band systems
    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03GCONTROL OF AMPLIFICATION
    • H03G7/00Volume compression or expansion in amplifiers
    • H03G7/007Volume compression or expansion in amplifiers of digital or coded signals
    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03GCONTROL OF AMPLIFICATION
    • H03G9/00Combinations of two or more types of control, e.g. gain control and tone control
    • H03G9/005Combinations of two or more types of control, e.g. gain control and tone control of digital or coded signals

Definitions

  • the present invention relates to an apparatus for processing an audio signal and method thereof.
  • the present invention is suitable for a wide scope of applications, it is particularly suitable for processing audio signals.
  • an audio signal is outputted via a loud speaker of a television or a portable device, a headset or the like.
  • an audio processor is able to perform noise canceling, normalizing, volume adjustment functions and the like to the audio signal.
  • the present invention is directed to an apparatus for processing an audio signal and method thereof that substantially obviate one or more of the problems due to limitations and disadvantages of the related art.
  • An object of the present invention is to provide an apparatus for processing an audio signal and method thereof, by adaptively normalizing a local dynamic range as well as a maximum dynamic range of an audio signal.
  • Another object of the present invention is to provide an apparatus for processing an audio signal and method thereof, by which an extent of noise canceling can be adjusted according to an extent of normalizing.
  • the present invention provides the following effects and/or advantages.
  • FIG. 1 is a block diagram of an audio signal processing apparatus according to an embodiment of the present invention
  • FIG. 2 is a detailed block diagram of a noise canceling unit shown in FIG. 1 ;
  • FIG. 3 is a detailed block diagram of a normalizing unit shown in FIG. 2;
  • FIG. 4 is a diagram for explaining a function of a fast gain generating part shown in FIG. 3;
  • FIG. 5 is a diagram for an example of a normalizing gain and attenuation limit for sample music;
  • FIG. 6 is a conceptional diagram for the relation between a normalizing gain and an attenuation limit value
  • FIG. 7 is a diagram for first and second examples of an audio signal processing apparatus according to another embodiment of the present invention.
  • FIG. 8 is a diagram for third and fourth examples of an audio signal processing apparatus according to another embodiment of the present invention.
  • FIG. 9 is a detailed block diagram of a volume controlling unit shown in FIG. 7 or FIG. 8;
  • FIG. 10 is a block diagram of an audio signal decoding apparatus to which an audio signal processing apparatus according to an embodiment of the present invention is applied;
  • FIG. 11 is a schematic diagram of a product in which an audio signal processing apparatus according to an embodiment of the present invention is implemented; and FIG. 12 is a diagram of products provided with an audio signal processing apparatus according to an embodiment of the present invention.
  • a method for processing an audio signal includes receiving a signal, by an audio processing apparatus; computing a long- term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.
  • the final gain is generated by multiplying the slow gain by the fast gain.
  • the fast gain is generated using at least one of a limiter threshold, a compressor threshold, and an expander threshold.
  • the slow gain and the fast gain are generated by being smoothed over time, and wherein the slow gain and the fast gain are within a limit range.
  • the slow gain, the fast gain and the final gain correspond to a full-band gain.
  • the slow gain and the fast gain are generated using a target level value included in at least one of user input information and device setting information.
  • the method further comprises generating a feedback information using the final gain.
  • the feedback information limit is associated with at least one of a minimum value and a maximum value of gain filter for noise canceling.
  • an apparatus for processing an audio signal includes a power estimating part computing a long-term power and a short-term power by estimating power of a signal; a slow gain generating part generating a slow gain based on the long-term power; a fast gain generating part generating a fast gain based on the short-term power; a final gain obtaining part obtaining a final gain by combining the slow gain and the fast gain; and, an output signal generating part modifying the signal using the final gain.
  • the final gain is generated by multiplying the slow gain by the fast gain.
  • the fast gain is generated using at least one of a limiter threshold, a compressor threshold, and an expander threshold.
  • the slow gain and the fast gain are generated by being smoothed over time, and wherein the slow gain and the fast gain are within a limit range.
  • the slow gain, the fast gain and the final gain correspond to a full-band gain.
  • the slow gain and the fast gain are generated using a target level value included in at least one of user input information and device setting information.
  • the apparatus further comprises a feedback information generating part generating feedback information using the final gain.
  • the feedback information is associated with at least one of a minimum value and a maximum value of gain filter for noise canceling.
  • a computer-readable medium having instructions stored thereon which, when executed by a processor, causes the processor to perform operations, includes receiving a signal, by an audio processing apparatus; computing a long-term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.
  • a method for processing an audio signal includes receiving, by an audio processing apparatus, a signal, and feedback information estimated based on a normalizing gain; generating a noise estimation based on the signal; computing a gain filter for noise canceling, based on the noise estimation and the signal; and, obtaining a restricted gain filter by applying the feedback information to the gain filter.
  • the method further comprises generating an output signal by applying the restricted gain filter to the signal.
  • the normalizing gain is generated based on the output signal and a target level value.
  • the gain filter is a time-frequency- variant value associated with a relative proportion of a noise signal in the signal.
  • the gain filter and the noise signal are in inverse proportion to each other.
  • the normalizing gain corresponds to a combination of a slow gain and a fast gain.
  • the slow gain corresponds to a gain for controlling a maximum dynamic range of the signal
  • the fast gain corresponds to a gain for controlling a local dynamic range of the signal.
  • an apparatus for processing an audio signal includes a noise detecting part generating a noise estimation based on a signal; and, a gain filter computing part computing a gain filter for noise canceling, based on the noise estimation and the signal, a gain filter computing part obtaining a restricted gain filter by applying feedback information to the gain filter, wherein the feedback information is received from normalizing part, wherein the feedback information is estimated based on a normalizing gain.
  • the apparatus further comprises a multifier generating an output signal by applying the restricted gain filter to the signal.
  • the normalizing gain is generated based on the output signal and a target level value.
  • the gain filter is a time-frequency-variant value associated with a relative proportion of a noise signal in the signal.
  • the gain filter and the noise signal are in inverse proportion to each other.
  • the normalizing gain corresponds to a combination of a slow gain and a fast gain.
  • the slow gain corresponds to a gain for controlling a maximum dynamic range of the signal
  • the fast gain corresponds to a gain for controlling a local dynamic range of the signal.
  • a computer-readable medium having instructions stored thereon which, when executed by a processor, causes the processor to perform operations, includes receiving, by an audio processing apparatus, a signal, and feedback information estimated based on a normalizing gain; generating a noise estimation based on the signal; computing a gain filter for noise canceling, based on the noise estimation and the signal; and, generating a restricted gain filter by applying the feedback information to the gain filter.
  • an audio signal in a broad sense, is conceptionally discriminated from a video signal and designates all kinds of signals that can be auditorily identified.
  • the audio signal means a signal having none or small quantity of speech characteristics.
  • Audio signal of the present, invention should be construed in a broad sense.
  • the audio signal of the present invention can be understood as a narrow-sense audio signal in case of being used by being discriminated from a speech signal.
  • FIG. 1 is a block diagram of an audio signal processing apparatus according to an embodiment of the present invention.
  • an audio signal processing apparatus includes a noise canceling unit 100 and a normalizing unit 200 and may further include a volume controlling unit 300.
  • the noise canceling unit 100 detects or estimates a noise included in an audio signal for an input audio signal y[n], determines a per- time per-frequency gain based on an relative proportion of the noise, and then applies the determined gain to the audio signal. In determining the gain, it is able to use feedback information outputted from the normalizing unit 200.
  • the noise canceling unit 100 will be described in detail with reference to FIG. 2 later.
  • the normalizing unit 200 is able to adaptively control a maximum dynamic range and local dynamic range of an input signal by performing normalizing on the input audio signal y[n] or a signal processed by the noise canceling unit, based on a target level value.
  • the normalizing unit 200 will be explained in detail with reference to FIGS. 3 to 6 later.
  • the target level value is an information inputted by a user or an device setting information and may correspond to a full-band gain (e.g., 1OdB irrespective of a frequency band).
  • the volume controlling unit 300 controls a volume for the normalized audio signal based on the target level value.
  • FIG. 2 is a detailed block diagram of a noise canceling unit shown in FIG. 1.
  • the noise canceling unit 100 can include a time-frequency analysis part 110, a gain filter computing part 120, a noise detecting part 130 and a time- frequency synthesis part 140.
  • Vu 1 I is an input signal
  • L J is a desired signal or a clean signal
  • v i n ⁇ is a noise signal
  • 'n' indicates a time index
  • the time-frequency analysis part 100 performs a time-frequency analysis on an input audio signal.
  • the time-frequency analysis can be performed by short-time Fourier transform (STFT) or can be performed by a filterbank or another time-frequency transform, by which the present invention is non-limited.
  • STFT short-time Fourier transform
  • Formula 1 can be represented as the following formula.
  • YIk 1 Tn] X[k, m] + V[k, m]
  • 'k' indicates a time frame number
  • 'm' indicates a frequency bin index
  • an estimated phase can be chosen to be equal to the noisy phase.
  • a speech signal x and a noise signal v can be modeled by an uncorrelated stationary process. This is represented as the following formula. If the modeling is performed by this formula, it is able to reconstruct a power spectrum
  • a gain filter can be expressed as follows. [Formula 4]
  • the ⁇ is a parameter to control amount of noise to be suppressed.
  • This parameter can be selected in association with a signal to noise ratio (SNR) and a frequency index m.
  • SNR signal to noise ratio
  • the noise detecting part 130 detects noise for an input signal and then generates a
  • This noise is such a signal as a white noise and the like.
  • This noise can be detected in various ways. In case of a noise signal, it may be a signal failing to have self-correlation of a signal and inter-channel correlation of a signal. Hence, when tonality and inter-channel correlation of an input signal are measured, if theses values are lost, it is able to determine that it is a noise. Yet, the present invention is non-limited by a method of detecting a noise.
  • the gain filter computing part 130 computes a gain filter ⁇ • I ' * ' for noise canceling based on the input signal and the noise estimated value and then generates a restricted gain filter G BL[k, m] by applying feedback information including attenuation limit value to the computed gain.
  • the gain filter may be equal to the expression of Formula 4.
  • the gain G B [k,m] indicates an extent of the noise estimated value
  • the gain filter is a time-frequency-variant value associated with a relative proportion of a noise signal in the signal.
  • the gain filter and the noise signal are in inverse proportion to each other. If a relative ratio of a noise estimated value is high, a gain becomes a low value. If a relative ratio of a noise estimated value is low, a gain becomes a
  • a power I ⁇ I ⁇ * > ⁇ J of an input signal is lowered by applying a low gain to a time-frequency band determined to include a large quantity of noise.
  • a power I * ⁇ ""> ⁇ J I of an input signal is raised by applying a high gain to a time-frequency band determined to include a small quantity of noise.
  • the gain filter defined in Formula 4 is smoothed over time and is restricted not to become smaller than L dB.
  • the attenuation include in the feedback information may correspond to maximum value or minimum value of gain filter, which is time-variant. This is represented as the following formula.
  • GB indicates a gain filter defined in Formula 4 and 'L' indicates an attenuation limit value [dB], G B L indicates a restricted gain filter.
  • the L is a lower limit of the gain G.
  • an attenuation limit value is a value that varies according to a frequency band.
  • the attenuation limit value can be determined based on the feedback information including the attenuation limit information received from a normalizing unit 200, which will be explained later.
  • this attenuation limit information is provided to use a feedback for a gain control in the normalizing step in canceling a noise. For instance, if a low gain is applied in normalizing, a power of an input signal is not excessively reduced in the noise canceling step.
  • the feedback information including attenuation limit information will be explained in detail together with the description of the normalizing unit 200 later.
  • a multiplier 140 generates an estimation of a clean signal (or a desired signal) resulting from removing a noise from an input signal, as represented in Formula 3, by applying the gain determined by the gain filter computing part 120 to the input signal.
  • the time-frequency synthesis part 150 generates a clean signal in time domain by performing a time-frequency synthesis on the clean signal.
  • FIG. 3 is a detailed block diagram of a normalizing unit shown in FIG. 2.
  • a normalizing unit 200 is a unit configured to normalize a clean signal generated by a noise canceling unit 100 from canceling a noise of an input signal.
  • An input signal of the normalizing unit 200 can be an output x(n) of the noise canceling unit and will be expressed as y[n] for convenience.
  • the normalizing unit 200 adaptively controls a dynamic range of an audio signal mainly in two ways. First of all, the normalizing unit 200 controls a maximum dynamic range of an input signal by performing slow adaptation. Secondly, the normalizing unit 200 controls a local dynamic range of an input signal by performing fast adaptation. Details will be described together with the respective elements.
  • a gain applied by the normalizing unit i.e., a normalizing gain can be defined as a full-band gain shown in the following formula.
  • Y[k,m] is a frequency transformed input signal
  • Z[k,m] is a frequency transformed output signal through a normalizing unit
  • G[k] is a normalizing gain
  • k is a time frame number
  • m is a frequency bin index.
  • G[k] is a time-variant value irrespective of a frequency band.
  • the frequency transform can include a short-time Fourier transform (STFT), by which the present invention is non-limited.
  • STFT short-time Fourier transform
  • a power estimating part 210 performs a frequency transform on an input signal and then computes an instantaneous power of the input signal by summing a power per frequency band m at a specific timing point k, as shown in the following formula.
  • the power estimating part 210 can receive a frequency transformed signal without performing the frequency transform by itself, as shown in the drawing.
  • Y[k,m] is a frequency transformed input signal
  • k is a time frame number
  • m is a frequency bin index
  • long-term power ⁇ L ⁇ J slow can be calculated using a single pole averaging filter having a small attack time constant and a large release time constant.
  • the long-term power may correspond to a maximum dynamic range estimator.
  • the * i ⁇ J s i ow is transferred to a
  • the slow gain generating part 220 generates a slow gain based on the long-term
  • the slow gain can be defined by the following formula. [Formula 8]
  • ⁇ NRM is a target level value
  • *-* slow I ⁇ J is a slow gain
  • the slow gain generating part 220 is provided to 2 j- simply apply the long-term power ⁇ * lAJslow to the target level value SNRM .
  • the target level value may be the information inputted by a user or device setting information. For instance, in case that a user inputs 10 dB as a target level value, a slow gain is provided to control a range of an input signal not to deviate from 1OdB mostly.
  • the fast gain generating part 240 generates a fast gain ⁇ riast W) dB based on the
  • the fast gain enables the normalizing unit to fast adapt to a dynamic range of a signal by acting like an expander, a compressor or a limiter.
  • ( ⁇ M ) ⁇ B is a level of an output signal
  • LT is a limiter threshold
  • CT is a limiter threshold
  • ET is an expander threshold
  • PC is a compression strength
  • ME is an expander strength
  • FIG. 4 is a diagram for explaining a function of a fast gain generating part shown in FIG. 3, in which a relation between levels of input and output signals represented as Formula 9 is shown.
  • a horizontal axis indicates a level of an input signal and a vertical axis indicates a level of an output signal.
  • the horizontal axis is divided into a region (a) in which a level of an input signal is greater than LT, a region (b) in which a level of an input signal is between LT and CT, and a region (c) in which a level of an input signal is smaller than CT.
  • LT in which a level of an input signal is greater than LT
  • a region (b) in which a level of an input signal is between LT and CT
  • a region (c) in which a level of an input signal is smaller than CT.
  • a level of an output signal is adjusted according to this slope. For instance, if a level of an input signal is -1OdB, a level of an output signal becomes -2OdB. If a level of an input signal is -2OdB, a level of an output signal can become -25dB. In the region (c) [corresponding to a third line in Formula 9], a level of an output signal becomes greater than that of an input signal according to an expander
  • a level of an input signal is -5OdB
  • a level of an output signal can become -45dB.
  • compressor and expander are performed in the regions (a), (b) and (c), respectively, fast normalizing is performed in a manner of increasing a small signal and decreasing a large signal.
  • a dynamic range of an output signal can range from negative infinity to a level (e.g., -2OdB) of an output signal corresponding to a case that an input signal is LT.
  • ⁇ IAJ fast M-B is a level of an input signal.
  • a first line in Formula 10 corresponds to a case that an input signal is greater than LT.
  • a second line indicates a case that an input signal is between CT and LT.
  • a third line indicates a case that an input signal is smaller than CT.
  • the fast gain represented as Formula 10 After the fast gain represented as Formula 10 has been obtained, the fast gain can be smoothed in time direction to prevent high variation artifacts like the slow gain.
  • the slow gain and the fast gain which are generated by the slow gain generating part 220 and the fast gain generating part 240 can be limited within a reasonable range by limiting parts 230 and 250, respectively.
  • the first limiting part 230 can generate a limited slow gain G S I OW ,L shown in the following formula.
  • a slow gain is limited within a range between — L s i ow and L s i ow -
  • G fast L
  • a fast gain is limited within a range between -Lf 3St and L fast .
  • the first and second limiting parts 230 and 250 are not mandatory and can be omitted if necessary.
  • a final gain obtaining part 260 obtains a final normalizing gain G[k] using a limited slow gain and a limited fast gain (occasionally, using a non-limited slow gain and a non-limited fast gain) according to the following formula. [Formula 12]
  • G[Ic] is a final gain
  • G s i Ow ,L[k] is a limited slow gain
  • Gf as t,L[k] is a limited fast gain.
  • a final gain can be generated from a combination of a slow gain and a fast gain.
  • the final gain can be generated from a multiplication of a slow gain and a fast gain.
  • a maximum dynamic range is controlled using a slow gain and a local dynamic range can be controlled using a fast gain.
  • An output signal generating part 270 modifies input signal using the final gain. For example, the output signal generating part generates an output signal by normalizing the input signal using a final gain.
  • a feedback information generating part 280 generates a feedback information including attenuation limit value ⁇ L[k]>dB using a final gain G[Ic]. This can be defined as the following formula. [Formula 13]
  • ⁇ L[k]>dB is an attenuation limit value and G[Ic] is a final gain for normalizing.
  • a normalizing gain applied to a TV audio content is smoothed over time to minimize an effect of artifacts attributed to quick variations of slow and fast gains. Yet, as a smoothing time constant needs to be sufficiently small, an overall adaptation of a normalizing unit follows the dynamic of an audio signal. Yet, as a result of non-instant adaptation, a background noise is modulated according to variations of gain. This is annoying especially when a volume of a signal is boosted by the normalizing unit. In other words, when a level of an input signal is too low, if a normalizer raises a level of the input signal as an expander, a level of noise of background is raised together to make the noise heard loudly. Hence, an audio quality is degraded. In order to reduce this effect, the aforesaid noise canceling unit 100 is able to limit the gain filter of noise canceling like Formula 5 using the attenuation limit value shown in the above formula.
  • an attenuation limit value becomes a negative final gain. If a final gain has a negative value, an attenuation becomes zero. Namely, if a normalizing gain raises a level of an input signal (i.e., if positive), an attenuation limit value becomes a negative final gain. And, a lower limit of a noise gain filter is lowered as many as the corresponding final gain, as shown in Formula 5. If a lower limit is further lowered, cancellation of noise can be further achieved.
  • FIG. 5 is a diagram for an example of a normalizing gain and attenuation limit for sample music.
  • a music clip y(n) of about 5 seconds and a normalizing gain G(n) corresponding to the music clip y(n) are shown. Since a level of an input signal is high at about 0.5, 2.6 or 4.7 seconds on a time axis, it can be observed that a normalizing gain in this interval is about -2dB ⁇ -3dB, which is low. Meanwhile, it can be observed that a normalizing gain in an interval having a relatively low level is about 0 - 3 dB, which is relatively high. Meanwhile, referring to (C) of FIG. 5, it can be observed that an attenuation limit value L(n) varies in a range between -6dB and 0 according to variations of a normalizing gain. In an interval having a relatively high normalizing gain
  • an attenuation limit value is relatively low.
  • FIG. 6 is a conceptional diagram for the relation between a normalizing gain and an attenuation limit value.
  • a range of a gain filter value G BL for noise canceling is between 10 L 2 ° and GB because a lower limit value is 10 L/2 °, as shown in Formula 5.
  • G B has a maximum value if noise is not detected at all. According to Formula 5, a maximum value of G B can become 1.
  • An interval having a large value of G B is a region having small noise, while an interval having a GB value close to a lower limit value is a region having many noises.
  • the former description relates to one embodiment of the present invention including a noise canceling unit and a normalizing unit.
  • another embodiment of the present invention including a volume controlling unit and other units (a noise detecting unit, etc.) is explained.
  • FIG. 7 is a diagram for first and second examples of an audio signal processing apparatus according to another embodiment of the present invention
  • FIG. 8 is a diagram for third and fourth examples of an audio signal processing apparatus according to another embodiment of the present invention.
  • an audio signal processing apparatus includes a noise detecting part 410 and a volume controlling unit 420.
  • the noise detecting part 410 analyses an input signal to detect an extent of noise included in the input signal per time and per frequency band and then outputs the corresponding information as noise information.
  • a method of detecting the noise may be equal to that of the noise detecting part 130 explained in the foregoing description with reference to FIG. 2. In this
  • the noise information can include an estimation I r v ? m ⁇ of a power spectrum of a noise signal, which is explained with reference to FIG. 2 or may correspond to a per- time per-frequency band gain filter (GBL)-
  • a lower limit value of the gain filter may be determined using the attenuation limit information received from the normalizing unit.
  • the volume controlling unit 420 controls a volume using the noise information received from the noise detecting part 410 and a target volume level received from a user input. For example, in controlling the volume, it is able to prevent a noise component from being heard by listener's ears abnormally and loudly. And, it is also able to prevent a level of a noise output from fluctuating. After an input signal has been analyzed, if a current input signal is a noise component, it is compared to a case of a normal tone component and another volume level scheme is then applicable. Details of the volume controlling unit 420 will be explained with reference to FIG. 9 later.
  • the first example corresponds to an example of detecting a noise and then using the corresponding result for a volume control.
  • the second example corresponds to an example of canceling a noise after completion of a volume control.
  • a volume controlling unit 520 controls a volume based on a target volume level for an input signal.
  • the volume controlled signal is sent to the noise canceling unit 510.
  • the noise canceling unit 510 detects a noise signal and then generates a clean signal by removing the noise signal from the volume controlled signal using the corresponding result.
  • the noise canceling unit 510 is ale to have the same element of the former noise canceling unit 100 described with reference to FIG. 1 and FIG. 2. In particular, it is able to determine a lower limit value of a gain filter for noise canceling using attenuation limit information received from a normalizing unit.
  • a noise detecting part 620 analyzes a noise component of an input signal, generates noise information based on this analysis, and then delivers the generated noise information to a volume controlling unit 620 and a noise reduction part 615.
  • the volume controlling unit 620 controls a volume of an input signal based on a target volume level and the noise information.
  • the noise reduction part 615 generates a clean signal by canceling or reducing the noise from a volume adjusted output based on the noise information.
  • the noise detecting part 610 and the noise reduction part 615 may perform the same function of the noise canceling unit 510 of the second example but may differ from the noise canceling unit 510 of the second example in a target of the noise detecting.
  • a target of noise detecting is an input signal itself in case of the third example.
  • a target of noise detecting is a volume control completed signal.
  • the noise reduction part 615 is able to use a gain filter limited by an attenuation limit like the case of the noise canceling unit described with reference to FIG. 2.
  • an audio signal processing apparatus according to another embodiment of the present invention is shown.
  • both noise detecting and noise reduction are performed before volume controlling.
  • a noise detecting part 710 detects a noise signal from an input signal.
  • a noise reduction part 710 removes the detected noise signal.
  • a volume controlling unit 720 controls a volume of the noise canceled signal based on a target volume level.
  • the noise detecting part 710 and the noise reduction part 715 can perform the same function of the former noise canceling unit 100 explained with reference to FIG. 1.
  • the volume controlling unit 720 can perform the same function of the former volume controlling unit 300 described with reference to FIG. 1.
  • FIG. 9 is a detailed block diagram of a volume controlling unit 420/520/620/720 shown in FIG. 7 or FIG. 8.
  • the volume controlling unit 420 is an element for controlling a volume of an input signal based on a target volume level. And, the volume controlling unit 420 is able to include a time-frequency analysis part 422, a gain processing part 424, a loudness estimating part 426 and a time-frequency synthesis part 428.
  • the time-frequency analysis part 422 generates a signal in a frequency domain by performing frequency transform.
  • the frequency transform can be performed by ERB (equivalent rectangular bandwidth) frequency scale, by which the present invention is non-limited.
  • the loudness estimating part 426 estimates a loudness of sound by performing a perceptual (psychoacoustic) loudness analysis on an input signal and generates a target gain value from the loudness.
  • the loudness is an auditory sense actually perceived by a man for each sound pressure level and has a concept (unit of phon) of representing a sound intensity of sound actually recognized by a human ear.
  • the loudness has the concept discriminated from a sound intensity (unit of dB) determined by a sound pressure level.
  • the loudness depends on a lasting time of sound, an occurring time of sound, a spectrum characteristic and the like as well as the sound intensity.
  • Perceptual loudness analysis is to analyze loudness of an input signal based on an equal loudness curve which is the result of a human test.
  • a target gain is obtained based on the loudness and a target volume level.
  • the target volume level is a full-band gain corresponding to a user inputted volume command or device setting information and may equal to the former target level value explained together with the former embodiment.
  • the target gain is a target value of volume in consideration of loudness and may include a value that varies per frequency band. Meanwhile, the target gain can be re-adjusted to raise a sound quality.
  • the gain processing part 424 applies the target gain to the frequency transformed input signal.
  • the time-frequency synthesis part 426 frequency-synthesizes the gain processed signal.
  • FIG. 10 is a block diagram of an audio signal decoding apparatus according to an embodiment of the present invention.
  • an audio signal decoding apparatus 800 can include a demultiplexer 810, an audio signal decoder 820, a band extension decoder 830, a band expansion decoding apparatus 840 and a multi-channel decoder 850. And, the audio signal decoding apparatus 860 further includes an audio signal processing apparatus 860 according to the present invention.
  • the audio signal processing apparatus 860 corresponds to one embodiment of the former audio signal processing apparatus described with reference to FIG. 1 or can correspond to another embodiment (first to fourth examples) of the former audio signal processing apparatus described with reference to FIG. 7 or FIG. 8.
  • the demultiplexer 810 extracts spectral data, band expansion information, spatial information and the like from an audio signal bitstream.
  • the audio signal decoder 820 decodes the spectral data by an audio coding scheme if the spectral data corresponding to a downmix signal has a large audio characteristic.
  • the audio coding scheme can follow AAC standard, HE-AAC standard or the like.
  • the audio signal decoder 820 reconstructs a frequency transformed residual by performing dequantization on the spectral data and scale factor carried on the bitstream. Subsequently, the audio signal decoder 820 generates a residual (inverse transformed) by performing inverse frequency transform on the frequency transformed residual.
  • the speech signal decoder 830 decodes the downmix signal by a speech coding scheme if the spectral data has a large speech characteristic.
  • the speech coding scheme can follow the AMR-WB (adaptive multi-rate wide-band) standard, by which the present invention is non-limited.
  • the band expansion decoder 840 decodes a band extension information bitstream and then generates an audio signal (or, spectral data) of another band (e.g., high frequency band) from a portion or all of the audio signal (or, spectral data) using this information.
  • an audio signal or, spectral data
  • another band e.g., high frequency band
  • the multi-channel decoder 850 If the decoded audio signal is a downmix, the multi-channel decoder 850 generates an output channel signal of a multi-channel signal (stereo signal included) using the spatial information.
  • the audio signal processing apparatus 860 is the former audio signal processing apparatus according to an embodiment of the present invention described with reference to FIG. 1, FIG. 7 or FIG. 8. And, the audio signal processing apparatus 860 performs such processing as noise canceling, normalizing, volume controlling and the like on the output channel signal from the multi-channel decoder.
  • the audio signal processing apparatus is available for various products to use. Theses products can be grouped into a stand alone group and a portable group. A TV, a monitor, a settop box and the like belong to the stand alone group. And, a PMP, a mobile phone, a navigation system and the like belong to the portable group.
  • FIG. 11 is a block diagram of a product in which an audio signal processing apparatus according to an embodiment of the present invention is implemented
  • FIG. 12 is a diagram for relations between products in which an audio signal processing apparatus according to an embodiment of the present invention is implemented.
  • a wire/wireless communication unit 910 receives a bitstream via wire/wireless communication system.
  • the wire/wireless communication unit 910 can include at least one of a wire communication unit 910A, an infrared communication unit 910B, a Bluetooth unit 910C and a wireless LAN communication unit
  • a user authenticating unit 920 receives an input of user information and then performs user authentication.
  • the user authenticating unit 920 can include at least one of a fingerprint recognizing unit 920A, an iris recognizing unit 920B, a face recognizing unit 920C and a speech recognizing unit 920D.
  • 920D receive fingerprint information, iris information, face contour information and speech information and then convert them into user informations, respectively. Whether each of the user informations matches pre-registered user data is determined to perform the user authentication.
  • An input unit 930 is an input device enabling a user to input various kinds of commands and can include at least one of a keypad unit 930A, a touchpad unit 930B and a remote controller unit 930C, by which the present invention is non-limited.
  • a signal coding unit 940 performs encoding or decoding on an audio signal and/or a video signal, which is received via the wire/wireless communication unit 910, and then output an audio signal in time domain.
  • the signal coding unit 940 includes an audio signal processing apparatus 945.
  • the audio signal processing apparatus 945 corresponds to one of the former audio signal processing apparatuses according to one embodiment and other embodiments (first to fourth examples) of the present invention. And, the audio signal processing apparatus 945 performs such processing as noise canceling, normalizing, volume controlling and the like on an audio signal before being outputted via the output unit.
  • the audio signal processing apparatus and the signal coding unit including the audio signal processing apparatus can be implemented by at least one or more processors.
  • a control unit 950 receives input signals from input devices and controls all processes of the signal decoding unit 940 and an output unit 960.
  • the output unit 960 is an element configured to output an output signal generated by the signal decoding unit 940 and the like and can include a speaker unit 960A and a display unit 960B. If the output signal is an audio signal, it is outputted to a speaker. If the output signal is a video signal, it is outputted via a display.
  • FIG. 12 shows the relation between the terminal corresponding to the product shown in FIG. 11 and a server.
  • a first terminal 900.1 and a second terminal 900.2 can exchange data or bitstreams bi-directionally with each other via the wire/wireless communications units.
  • a server 1000 and a first terminal 900.1 can perform wire/wireless communication with each other.
  • An audio signal processing method can be implemented into a computer-executable program and can be stored in a computer- readable recording medium.
  • multimedia data having a data structure of the present invention can be stored in the computer-readable recording medium.
  • the computer- readable media include all kinds of recording devices in which data readable by a computer system are stored.
  • the computer-readable media include ROM, RAM, CD-ROM, magnetic tapes, floppy discs, optical data storage devices, and the like for example and also include carrier- wave type implementations (e.g., transmission via Internet).
  • a bitstream generated by the above encoding method can be stored in the computer-readable recording medium or can be transmitted via wire/wireless communication network.
  • the present invention is applicable to processing and outputting an audio signal.

Abstract

An apparatus for processing an audio signal and method thereof are disclosed, by which a local dynamic range of an audio signal can be adaptively normalized as well as a maximum dynamic range of the audio signal. The present invention includes receiving a signal, by an audio processing apparatus; computing a long-term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.

Description

[DESCRIPTION]
AN APPARATUS FOR PROCESSING AN AUDIO SIGNALAND METHOD
THEREOF
TECHNICAL FIELD
The present invention relates to an apparatus for processing an audio signal and method thereof. Although the present invention is suitable for a wide scope of applications, it is particularly suitable for processing audio signals.
BACKGROUND ART
Generally, an audio signal is outputted via a loud speaker of a television or a portable device, a headset or the like. Before the audio signal is outputted via a speaker or the like, an audio processor is able to perform noise canceling, normalizing, volume adjustment functions and the like to the audio signal.
DISCLOSURE OF THE INVENTION
TECHNICAL PROBLEM
However, in the related art, such processing as noise canceling, normalizing and the like is performed individually and independently. Moreover, when an audio content varies widely (e.g., zapping between TV channels or radio channels (several songs) is performed), normalizing may not be performed according to a characteristic o the audio content.
Moreover, since noise varies rapidly when volume goes up or down, a listener may hear how a normalizer of a background noise adjusts the volume. TECHNICAL SOLUTION
Accordingly, the present invention is directed to an apparatus for processing an audio signal and method thereof that substantially obviate one or more of the problems due to limitations and disadvantages of the related art. An object of the present invention is to provide an apparatus for processing an audio signal and method thereof, by adaptively normalizing a local dynamic range as well as a maximum dynamic range of an audio signal.
Another object of the present invention is to provide an apparatus for processing an audio signal and method thereof, by which an extent of noise canceling can be adjusted according to an extent of normalizing.
ADVANTAGEOUS EFFECTS
Accordingly, the present invention provides the following effects and/or advantages. First of all, as normalizing is adaptively performed on a local dynamic range as well as a maximum dynamic range, the present invention can minimize a damage caused to a sound quality as well as automatically prevents a volume from being excessively raised or lowered.
Secondly, as the extent of noise cancelling is adjusted according to the extent of normalizing, it is able to prevent level of noise from rising together when level of signal is automatically raised by normalizing.
DESCRIPTION OF DRAWINGS
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention.
In the drawings:
FIG. 1 is a block diagram of an audio signal processing apparatus according to an embodiment of the present invention;
FIG. 2 is a detailed block diagram of a noise canceling unit shown in FIG. 1 ;
FIG. 3 is a detailed block diagram of a normalizing unit shown in FIG. 2;
FIG. 4 is a diagram for explaining a function of a fast gain generating part shown in FIG. 3; FIG. 5 is a diagram for an example of a normalizing gain and attenuation limit for sample music;
FIG. 6 is a conceptional diagram for the relation between a normalizing gain and an attenuation limit value;
FIG. 7 is a diagram for first and second examples of an audio signal processing apparatus according to another embodiment of the present invention;
FIG. 8 is a diagram for third and fourth examples of an audio signal processing apparatus according to another embodiment of the present invention;
FIG. 9 is a detailed block diagram of a volume controlling unit shown in FIG. 7 or FIG. 8; FIG. 10 is a block diagram of an audio signal decoding apparatus to which an audio signal processing apparatus according to an embodiment of the present invention is applied;
FIG. 11 is a schematic diagram of a product in which an audio signal processing apparatus according to an embodiment of the present invention is implemented; and FIG. 12 is a diagram of products provided with an audio signal processing apparatus according to an embodiment of the present invention.
BEST MODE
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
To achieve these and other advantages and in accordance with the purpose of the present invention, as embodied and broadly described, a method for processing an audio signal includes receiving a signal, by an audio processing apparatus; computing a long- term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.
Preferably, the final gain is generated by multiplying the slow gain by the fast gain.
Preferably, the fast gain is generated using at least one of a limiter threshold, a compressor threshold, and an expander threshold.
Preferably, the slow gain and the fast gain are generated by being smoothed over time, and wherein the slow gain and the fast gain are within a limit range.
Preferably, the slow gain, the fast gain and the final gain correspond to a full-band gain.
Preferably, the slow gain and the fast gain are generated using a target level value included in at least one of user input information and device setting information. Preferably, the method further comprises generating a feedback information using the final gain. In this case, the feedback information limit is associated with at least one of a minimum value and a maximum value of gain filter for noise canceling.
To further achieve these and other advantages and in accordance with the purpose of the present invention, an apparatus for processing an audio signal includes a power estimating part computing a long-term power and a short-term power by estimating power of a signal; a slow gain generating part generating a slow gain based on the long-term power; a fast gain generating part generating a fast gain based on the short-term power; a final gain obtaining part obtaining a final gain by combining the slow gain and the fast gain; and, an output signal generating part modifying the signal using the final gain. Preferably, the final gain is generated by multiplying the slow gain by the fast gain.
Preferably, the fast gain is generated using at least one of a limiter threshold, a compressor threshold, and an expander threshold.
Preferably, the slow gain and the fast gain are generated by being smoothed over time, and wherein the slow gain and the fast gain are within a limit range. Preferably, the slow gain, the fast gain and the final gain correspond to a full-band gain.
Preferably, the slow gain and the fast gain are generated using a target level value included in at least one of user input information and device setting information.
Preferably, the apparatus further comprises a feedback information generating part generating feedback information using the final gain. In this case, the feedback information is associated with at least one of a minimum value and a maximum value of gain filter for noise canceling.
To further achieve these and other advantages and in accordance with the purpose of the present invention, a computer-readable medium having instructions stored thereon, which, when executed by a processor, causes the processor to perform operations, includes receiving a signal, by an audio processing apparatus; computing a long-term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.
To further achieve these and other advantages and in accordance with the purpose of the present invention, a method for processing an audio signal includes receiving, by an audio processing apparatus, a signal, and feedback information estimated based on a normalizing gain; generating a noise estimation based on the signal; computing a gain filter for noise canceling, based on the noise estimation and the signal; and, obtaining a restricted gain filter by applying the feedback information to the gain filter.
Preferably, the method further comprises generating an output signal by applying the restricted gain filter to the signal. In this case, the normalizing gain is generated based on the output signal and a target level value. Preferably, the gain filter is a time-frequency- variant value associated with a relative proportion of a noise signal in the signal.
Preferably, the gain filter and the noise signal are in inverse proportion to each other.
Preferably, the normalizing gain corresponds to a combination of a slow gain and a fast gain. In this case, the slow gain corresponds to a gain for controlling a maximum dynamic range of the signal, and, the fast gain corresponds to a gain for controlling a local dynamic range of the signal.
To further achieve these and other advantages and in accordance with the purpose of the present invention, an apparatus for processing an audio signal includes a noise detecting part generating a noise estimation based on a signal; and, a gain filter computing part computing a gain filter for noise canceling, based on the noise estimation and the signal, a gain filter computing part obtaining a restricted gain filter by applying feedback information to the gain filter, wherein the feedback information is received from normalizing part, wherein the feedback information is estimated based on a normalizing gain.
Preferably, the apparatus further comprises a multifier generating an output signal by applying the restricted gain filter to the signal. In this case, the normalizing gain is generated based on the output signal and a target level value.
Preferably, the gain filter is a time-frequency-variant value associated with a relative proportion of a noise signal in the signal.
Preferably, the gain filter and the noise signal are in inverse proportion to each other.
Preferably, the normalizing gain corresponds to a combination of a slow gain and a fast gain. In this case, the slow gain corresponds to a gain for controlling a maximum dynamic range of the signal, and, the fast gain corresponds to a gain for controlling a local dynamic range of the signal.
To further achieve these and other advantages and in accordance with the purpose of the present invention, a computer-readable medium having instructions stored thereon, which, when executed by a processor, causes the processor to perform operations, includes receiving, by an audio processing apparatus, a signal, and feedback information estimated based on a normalizing gain; generating a noise estimation based on the signal; computing a gain filter for noise canceling, based on the noise estimation and the signal; and, generating a restricted gain filter by applying the feedback information to the gain filter.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are intended to provide further explanation of the invention as claimed.
MODE FOR INVENTION
Reference will now be made in detail to the preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings. First of all, terminologies or words used in this specification and claims are not construed as limited to the general or dictionary meanings and should be construed as the meanings and concepts matching the technical idea of the present invention based on the principle that an inventor is able to appropriately define the concepts of the terminologies to describe the inventor's invention in best way. The embodiment disclosed in this disclosure and configurations shown in the accompanying drawings are just one preferred embodiment and do not represent all technical idea of the present invention. Therefore, it is understood that the present invention covers the modifications and variations of this invention provided they come within the scope of the appended claims and their equivalents at the timing point of filing this application.
The following terminologies in the present invention can be construed based on the following criteria and other terminologies failing to be explained can be construed according to the following purposes. First of all, it is understood that the concept 'coding' in the present invention includes both encoding and decoding. Secondly, 'information' in this disclosure is the terminology that generally includes values, parameters, coefficients, elements and the like and its meaning can be construed as different occasionally, by which the present invention is non-limited.
In this disclosure, in a broad sense, an audio signal is conceptionally discriminated from a video signal and designates all kinds of signals that can be auditorily identified. In a narrow sense, the audio signal means a signal having none or small quantity of speech characteristics. Audio signal of the present, invention should be construed in a broad sense. And, the audio signal of the present invention can be understood as a narrow-sense audio signal in case of being used by being discriminated from a speech signal.
FIG. 1 is a block diagram of an audio signal processing apparatus according to an embodiment of the present invention.
Referring to FIG. 1, an audio signal processing apparatus includes a noise canceling unit 100 and a normalizing unit 200 and may further include a volume controlling unit 300.
The noise canceling unit 100 detects or estimates a noise included in an audio signal for an input audio signal y[n], determines a per- time per-frequency gain based on an relative proportion of the noise, and then applies the determined gain to the audio signal. In determining the gain, it is able to use feedback information outputted from the normalizing unit 200. The noise canceling unit 100 will be described in detail with reference to FIG. 2 later. The normalizing unit 200 is able to adaptively control a maximum dynamic range and local dynamic range of an input signal by performing normalizing on the input audio signal y[n] or a signal processed by the noise canceling unit, based on a target level value. The normalizing unit 200 will be explained in detail with reference to FIGS. 3 to 6 later. In this case, the target level value is an information inputted by a user or an device setting information and may correspond to a full-band gain (e.g., 1OdB irrespective of a frequency band).
And, the volume controlling unit 300 controls a volume for the normalized audio signal based on the target level value.
FIG. 2 is a detailed block diagram of a noise canceling unit shown in FIG. 1. Referring to FIG. 2, the noise canceling unit 100 can include a time-frequency analysis part 110, a gain filter computing part 120, a noise detecting part 130 and a time- frequency synthesis part 140.
First of all, an input audio signal y[n] is assumed as the following formula.
[Formula 1]
Figure imgf000011_0001
In Formula I5 Vu 1I is an input signal, L J is a desired signal or a clean signal,
v in\ is a noise signal, and 'n' indicates a time index.
The time-frequency analysis part 100 performs a time-frequency analysis on an input audio signal. Meanwhile, the time-frequency analysis can be performed by short-time Fourier transform (STFT) or can be performed by a filterbank or another time-frequency transform, by which the present invention is non-limited. In case that the time-frequency analysis is performed, Formula 1 can be represented as the following formula.
[Formula 2]
YIk1 Tn] = X[k, m] + V[k, m] In this case, 'k' indicates a time frame number and 'm' indicates a frequency bin index.
As human perception is relatively insensitive to phase corruption, an estimated phase can be chosen to be equal to the noisy phase. In this case, an estimation of \Λ [k, m\\ js suffϊcient to recover a desired speech signal or a desired audio signal. Within a single processing frame, a speech signal x and a noise signal v can be modeled by an uncorrelated stationary process. This is represented as the following formula. If the modeling is performed by this formula, it is able to reconstruct a power spectrum
I i- '' ?/2J I of a desired signal or a clean signal by subtracting an estimation I V [Zc. m\ of a power spectrum of a noise signal from I IA 5 771J I . According to the analogy with Wiener filtering, a spectral magnitude of a clean signal can be reconstructed by the following formula. [Formula 3]
Figure imgf000012_0001
In the above formula, a gain filter can be expressed as follows. [Formula 4]
Figure imgf000012_0002
In case of under or over estimation, the β is a parameter to control amount of noise to be suppressed. This parameter can be selected in association with a signal to noise ratio (SNR) and a frequency index m.
The noise detecting part 130 detects noise for an input signal and then generates a
noise estimation! L ■ " J I . This noise is such a signal as a white noise and the like. This noise can be detected in various ways. In case of a noise signal, it may be a signal failing to have self-correlation of a signal and inter-channel correlation of a signal. Hence, when tonality and inter-channel correlation of an input signal are measured, if theses values are lost, it is able to determine that it is a noise. Yet, the present invention is non-limited by a method of detecting a noise.
The gain filter computing part 130 computes a gain filter τ • I ' * ' for noise canceling based on the input signal and the noise estimated value and then generates a restricted gain filter GBL[k, m] by applying feedback information including attenuation limit value to the computed gain. In this case, the gain filter may be equal to the expression of Formula 4. The gain GB[k,m] indicates an extent of the noise estimated value
I K [ft, TTiJ I included per time-frequency band, compared to \ * IΛ> '' ^H . In other words, the gain filter is a time-frequency-variant value associated with a relative proportion of a noise signal in the signal. The gain filter and the noise signal are in inverse proportion to each other. If a relative ratio of a noise estimated value is high, a gain becomes a low value. If a relative ratio of a noise estimated value is low, a gain becomes a
high value. Namely, a power I ^ IΛ*> ^J of an input signal is lowered by applying a low gain to a time-frequency band determined to include a large quantity of noise. On the
contrary, a power I * \""> ^J I of an input signal is raised by applying a high gain to a time-frequency band determined to include a small quantity of noise.
Meanwhile, in order to prevent musical-like noise artifacts (i.e., in order to prevent a problem that noise is excessively canceled if noise is similar to music), the gain filter defined in Formula 4 is smoothed over time and is restricted not to become smaller than L dB. The attenuation include in the feedback information may correspond to maximum value or minimum value of gain filter, which is time-variant. This is represented as the following formula.
[Formula 5]
GBLΨ-, TΠ] = max j 1020 , GB \
In Formula 5, GB indicates a gain filter defined in Formula 4 and 'L' indicates an attenuation limit value [dB], GBL indicates a restricted gain filter.
In particular, since a gain G is impossible to be smaller than a value L (exactly, 10L/2°), the L is a lower limit of the gain G. By applying the attenuation limit, it is able to prevent a power of an input signal from being excessively lowered as a too-low gain is applied.
Meanwhile, an attenuation limit value (attenuation limit information) is a value that varies according to a frequency band. And, the attenuation limit value can be determined based on the feedback information including the attenuation limit information received from a normalizing unit 200, which will be explained later. In case that an input signal is normalized after the noise canceling, this attenuation limit information is provided to use a feedback for a gain control in the normalizing step in canceling a noise. For instance, if a low gain is applied in normalizing, a power of an input signal is not excessively reduced in the noise canceling step. The feedback information including attenuation limit information will be explained in detail together with the description of the normalizing unit 200 later.
A multiplier 140 generates an estimation of a clean signal (or a desired signal) resulting from removing a noise from an input signal, as represented in Formula 3, by applying the gain determined by the gain filter computing part 120 to the input signal. The time-frequency synthesis part 150 generates a clean signal in time domain by performing a time-frequency synthesis on the clean signal.
FIG. 3 is a detailed block diagram of a normalizing unit shown in FIG. 2.
Referring to FIG. 3, a normalizing unit 200 is a unit configured to normalize a clean signal generated by a noise canceling unit 100 from canceling a noise of an input signal. An input signal of the normalizing unit 200 can be an output x(n) of the noise canceling unit and will be expressed as y[n] for convenience.
The normalizing unit 200 adaptively controls a dynamic range of an audio signal mainly in two ways. First of all, the normalizing unit 200 controls a maximum dynamic range of an input signal by performing slow adaptation. Secondly, the normalizing unit 200 controls a local dynamic range of an input signal by performing fast adaptation. Details will be described together with the respective elements.
Meanwhile, a gain applied by the normalizing unit, i.e., a normalizing gain can be defined as a full-band gain shown in the following formula.
[Formula 6] Z[k, m] = G[k]Y[kM
In Formula 6, Y[k,m] is a frequency transformed input signal, Z[k,m] is a frequency transformed output signal through a normalizing unit, G[k] is a normalizing gain, k is a time frame number, and m is a frequency bin index.
In particular, G[k] is a time-variant value irrespective of a frequency band. Meanwhile, the frequency transform can include a short-time Fourier transform (STFT), by which the present invention is non-limited.
First of all, a power estimating part 210 performs a frequency transform on an input signal and then computes an instantaneous power of the input signal by summing a power per frequency band m at a specific timing point k, as shown in the following formula. Of course, the power estimating part 210 can receive a frequency transformed signal without performing the frequency transform by itself, as shown in the drawing.
[Formula 7]
Y[kf = Y \Y[k,m]\ m
In Formula 7, Y[k,m] is a frequency transformed input signal, k is a time frame number, and m is a frequency bin index.
Since it is attempted to find a full band gain G[k] to be applied by the normalizing unit, the instantaneous power is calculated by the above formula. By the above formula, a
long-term power ^ L^J slow can be calculated using a single pole averaging filter having a small attack time constant and a large release time constant. The long-term power may correspond to a maximum dynamic range estimator. On the other hand, a short-term power
* IAJ fast can be calculated using a small release time constant. The short-term power
may correspond to a local dynamic range estimator. The * i^Jsiow is transferred to a
slow gain generating part 220 and the * IAJ fast is transferred to a fast gain generating part 240.
The slow gain generating part 220 generates a slow gain based on the long-term
power Y L^J slow received from the power estimating part 210. The slow gain can be defined by the following formula. [Formula 8]
Figure imgf000016_0001
In Formula 8, ζNRM is a target level value, *-* slow IΛJ is a slow gain, and
" IAJ .slow is a long-term power.
Referring to the above formula, the slow gain generating part 220 is provided to 2 j- simply apply the long-term power * lAJslow to the target level value SNRM .
Meanwhile, the target level value may be the information inputted by a user or device setting information. For instance, in case that a user inputs 10 dB as a target level value, a slow gain is provided to control a range of an input signal not to deviate from 1OdB mostly.
Meanwhile, the slow gain GsiOw[k] by Formula 8 can be smoothed over time to avoid high variation artifacts. As a result of the smoothing, it is able to center a signal into the target level value. The fast gain generating part 240 generates a fast gain \^riast W) dB based on the
short-term power Y [®\ fast. The fast gain enables the normalizing unit to fast adapt to a dynamic range of a signal by acting like an expander, a compressor or a limiter. The fast
gain is provided to generate a level (zl^l )dB of an output signal. [Formula 9]
Figure imgf000017_0001
CT if LT < (ΨΪMJjdB
CT - CT) if CT ≤ (Y[k]ljdB < LT
ET
Figure imgf000017_0002
~ ET) if (Y[k]ljdB < CT,
In Formula 9, (^M )^B is a level of an output signal, LT is a limiter threshold, CT
is a compressor threshold, ET is an expander threshold,
Figure imgf000017_0003
is a level of an input
signal, PC is a compression strength, and ME is an expander strength.
FIG. 4 is a diagram for explaining a function of a fast gain generating part shown in FIG. 3, in which a relation between levels of input and output signals represented as Formula 9 is shown.
Referring to FIG. 4, a horizontal axis indicates a level of an input signal and a vertical axis indicates a level of an output signal. The horizontal axis is divided into a region (a) in which a level of an input signal is greater than LT, a region (b) in which a level of an input signal is between LT and CT, and a region (c) in which a level of an input signal is smaller than CT. In this case, even if an input signal is equal to or greater than LT, it is limited to LT in the region (a) [corresponding to a first line in Formula 9]. In the region (b) [corresponding to a second line in Formula 9], a slope varies according to a compression strength βo. And, a level of an output signal is adjusted according to this slope. For instance, if a level of an input signal is -1OdB, a level of an output signal becomes -2OdB. If a level of an input signal is -2OdB, a level of an output signal can become -25dB. In the region (c) [corresponding to a third line in Formula 9], a level of an output signal becomes greater than that of an input signal according to an expander
strength HE. For instance, if a level of an input signal is -5OdB, a level of an output signal can become -45dB. As functions of a limiter, compressor and expander are performed in the regions (a), (b) and (c), respectively, fast normalizing is performed in a manner of increasing a small signal and decreasing a large signal. For instance, a dynamic range of an output signal can range from negative infinity to a level (e.g., -2OdB) of an output signal corresponding to a case that an input signal is LT.
Thus, in order to obtain an output signal shown in Formula 9 and FIG. 4, it is able to apply a fast gain defined in the following formula to an input signal. In this case, the calculation of the fast gain can be performed by a simple line equation in a logarithmic domain. [Formula 10]
Figure imgf000018_0001
o
In Formula 10, \ IAJ fast M-B is a level of an input signal. CS is a slope factor of compression and CS = 1 — tan/% _ g§ js a sjOpe factor of an expander and ES = l - tan /fø. Like Formula 9, a first line in Formula 10 corresponds to a case that an input signal is greater than LT. A second line indicates a case that an input signal is between CT and LT. And, a third line indicates a case that an input signal is smaller than CT.
After the fast gain represented as Formula 10 has been obtained, the fast gain can be smoothed in time direction to prevent high variation artifacts like the slow gain.
Thus, the slow gain and the fast gain, which are generated by the slow gain generating part 220 and the fast gain generating part 240 can be limited within a reasonable range by limiting parts 230 and 250, respectively. The first limiting part 230 can generate a limited slow gain GSIOW,L shown in the following formula.
[Formula 11]
(Gάow,L[k})dB = max J min | {Gsiow [fc])dj5 , Lsiow | , -Lsiow |
In Formula 11,
Figure imgf000019_0001
W )dB is a fast gain, and Lsiow is a limit range of a slow gain.
Namely, a slow gain is limited within a range between — Lsiow and Lsiow- Likewise, the second limiting part 240 can generate a limited fast gain Gfast,L according to the following formula. [Formula 11] {GtatΛiφ])dB = max J min |{Grast[A;]}dB, Lfast,| ; -Lfast;|
In Formula 11, (^fast.L W)dS is a limited fast gain, (Φast Ψ])dB is a fast gain, and Lfast is a limit range of a fast gain.
Namely, a fast gain is limited within a range between -Lf3St and Lfast.
The first and second limiting parts 230 and 250 are not mandatory and can be omitted if necessary.
Meanwhile, a final gain obtaining part 260 obtains a final normalizing gain G[k] using a limited slow gain and a limited fast gain (occasionally, using a non-limited slow gain and a non-limited fast gain) according to the following formula. [Formula 12]
G[k] = GdwώklGtoΛtΦ]
In Formula 12, G[Ic] is a final gain, GsiOw,L[k] is a limited slow gain, and Gfast,L[k] is a limited fast gain. Namely, a final gain can be generated from a combination of a slow gain and a fast gain. In particular, the final gain can be generated from a multiplication of a slow gain and a fast gain. A maximum dynamic range is controlled using a slow gain and a local dynamic range can be controlled using a fast gain.
An output signal generating part 270 modifies input signal using the final gain. For example, the output signal generating part generates an output signal by normalizing the input signal using a final gain.
Meanwhile, a feedback information generating part 280 generates a feedback information including attenuation limit value <L[k]>dB using a final gain G[Ic]. This can be defined as the following formula. [Formula 13]
(L[k])dB = mm { - (G[k])dB, 0}
In Formula 13, <L[k]>dB is an attenuation limit value and G[Ic] is a final gain for normalizing.
A normalizing gain applied to a TV audio content is smoothed over time to minimize an effect of artifacts attributed to quick variations of slow and fast gains. Yet, as a smoothing time constant needs to be sufficiently small, an overall adaptation of a normalizing unit follows the dynamic of an audio signal. Yet, as a result of non-instant adaptation, a background noise is modulated according to variations of gain. This is annoying especially when a volume of a signal is boosted by the normalizing unit. In other words, when a level of an input signal is too low, if a normalizer raises a level of the input signal as an expander, a level of noise of background is raised together to make the noise heard loudly. Hence, an audio quality is degraded. In order to reduce this effect, the aforesaid noise canceling unit 100 is able to limit the gain filter of noise canceling like Formula 5 using the attenuation limit value shown in the above formula.
Looking into Formula 13, if a final gain has a positive value, an attenuation limit value becomes a negative final gain. If a final gain has a negative value, an attenuation becomes zero. Namely, if a normalizing gain raises a level of an input signal (i.e., if positive), an attenuation limit value becomes a negative final gain. And, a lower limit of a noise gain filter is lowered as many as the corresponding final gain, as shown in Formula 5. If a lower limit is further lowered, cancellation of noise can be further achieved.
FIG. 5 is a diagram for an example of a normalizing gain and attenuation limit for sample music.
Referring to (A) and (B) of FIG. 5, a music clip y(n) of about 5 seconds and a normalizing gain G(n) corresponding to the music clip y(n) are shown. Since a level of an input signal is high at about 0.5, 2.6 or 4.7 seconds on a time axis, it can be observed that a normalizing gain in this interval is about -2dB ~ -3dB, which is low. Meanwhile, it can be observed that a normalizing gain in an interval having a relatively low level is about 0 - 3 dB, which is relatively high. Meanwhile, referring to (C) of FIG. 5, it can be observed that an attenuation limit value L(n) varies in a range between -6dB and 0 according to variations of a normalizing gain. In an interval having a relatively high normalizing gain
(e.g., interval between 0 ~ 0.5 seconds, interval between 1.7 ~ 2.6 seconds, etc.), it can be observed that an attenuation limit value is relatively low.
FIG. 6 is a conceptional diagram for the relation between a normalizing gain and an attenuation limit value. Referring to FIG. 6, first of all, since a range of a gain filter value GBL for noise canceling is between 10L 2° and GB because a lower limit value is 10L/2°, as shown in Formula 5. GB has a maximum value if noise is not detected at all. According to Formula 5, a maximum value of GB can become 1. An interval having a large value of GB is a region having small noise, while an interval having a GB value close to a lower limit value is a region having many noises. In this case, the lower limit value (10L/2°) is not stationary, as mentioned in the foregoing description, but varies from \0'G^/2° to 1 (L=O) according to a normalizing gain G[Ic]. If a normalizing gain is negative (i.e., if a level of an input signal needs to be lowered), a lower limit of GB is raised to have noise canceled less. On the contrary, if a normalizing gain is positive (i.e., if a level of an input signal needs to be raised), a lower limit of GB is lowered to have noise canceled more. In other words, in case that a level of an input signal is considerably raised in a normalizing process, noise can be canceled more by further lowering a lower limit value of a gain applied in a noise canceling process. On the contrary, in case that a level of an input signal is lowered in a normalizing process, it is able to prevent noise from excessively canceled by raising a lower limit value of a gain in a noise canceling process.
The former description relates to one embodiment of the present invention including a noise canceling unit and a normalizing unit. In the following description, another embodiment of the present invention including a volume controlling unit and other units (a noise detecting unit, etc.) is explained.
FIG. 7 is a diagram for first and second examples of an audio signal processing apparatus according to another embodiment of the present invention, and FIG. 8 is a diagram for third and fourth examples of an audio signal processing apparatus according to another embodiment of the present invention. Referring to a first example shown in (A) of Fig. 7, an audio signal processing apparatus includes a noise detecting part 410 and a volume controlling unit 420. The noise detecting part 410 analyses an input signal to detect an extent of noise included in the input signal per time and per frequency band and then outputs the corresponding information as noise information. A method of detecting the noise may be equal to that of the noise detecting part 130 explained in the foregoing description with reference to FIG. 2. In this
case, the noise information can include an estimation I rv? m\ of a power spectrum of a noise signal, which is explained with reference to FIG. 2 or may correspond to a per- time per-frequency band gain filter (GBL)- In this case, as mentioned in the foregoing description with reference to FIG. 2, a lower limit value of the gain filter may be determined using the attenuation limit information received from the normalizing unit.
Meanwhile, the volume controlling unit 420 controls a volume using the noise information received from the noise detecting part 410 and a target volume level received from a user input. For example, in controlling the volume, it is able to prevent a noise component from being heard by listener's ears abnormally and loudly. And, it is also able to prevent a level of a noise output from fluctuating. After an input signal has been analyzed, if a current input signal is a noise component, it is compared to a case of a normal tone component and another volume level scheme is then applicable. Details of the volume controlling unit 420 will be explained with reference to FIG. 9 later.
Thus, the first example corresponds to an example of detecting a noise and then using the corresponding result for a volume control. And, the second example corresponds to an example of canceling a noise after completion of a volume control. Referring to (B) of Fig. 7, a volume controlling unit 520 controls a volume based on a target volume level for an input signal. The volume controlled signal is sent to the noise canceling unit 510. The noise canceling unit 510 detects a noise signal and then generates a clean signal by removing the noise signal from the volume controlled signal using the corresponding result.
Meanwhile, the noise canceling unit 510 is ale to have the same element of the former noise canceling unit 100 described with reference to FIG. 1 and FIG. 2. In particular, it is able to determine a lower limit value of a gain filter for noise canceling using attenuation limit information received from a normalizing unit.
Referring to (A) of FIG. 8, a third example of an audio signal processing apparatus according to another embodiment of the present invention is shown. Like the former noise detecting part 410 of the first example, a noise detecting part 620 analyzes a noise component of an input signal, generates noise information based on this analysis, and then delivers the generated noise information to a volume controlling unit 620 and a noise reduction part 615. Like the former volume controlling unit 420 of the first example, the volume controlling unit 620 controls a volume of an input signal based on a target volume level and the noise information. The noise reduction part 615 generates a clean signal by canceling or reducing the noise from a volume adjusted output based on the noise information. Meanwhile, the noise detecting part 610 and the noise reduction part 615 may perform the same function of the noise canceling unit 510 of the second example but may differ from the noise canceling unit 510 of the second example in a target of the noise detecting. In particular, a target of noise detecting is an input signal itself in case of the third example. In case of the second example, a target of noise detecting is a volume control completed signal. In case of the third example, the noise reduction part 615 is able to use a gain filter limited by an attenuation limit like the case of the noise canceling unit described with reference to FIG. 2.
Referring to (B) of FIG. 8, an audio signal processing apparatus according to another embodiment of the present invention is shown. In case of a fourth example, both noise detecting and noise reduction are performed before volume controlling. A noise detecting part 710 detects a noise signal from an input signal. A noise reduction part 710 removes the detected noise signal. A volume controlling unit 720 controls a volume of the noise canceled signal based on a target volume level. In this case, the noise detecting part 710 and the noise reduction part 715 can perform the same function of the former noise canceling unit 100 explained with reference to FIG. 1. And, the volume controlling unit 720 can perform the same function of the former volume controlling unit 300 described with reference to FIG. 1.
FIG. 9 is a detailed block diagram of a volume controlling unit 420/520/620/720 shown in FIG. 7 or FIG. 8. The volume controlling unit 420 is an element for controlling a volume of an input signal based on a target volume level. And, the volume controlling unit 420 is able to include a time-frequency analysis part 422, a gain processing part 424, a loudness estimating part 426 and a time-frequency synthesis part 428. In case that an input signal is a signal in a time domain, the time-frequency analysis part 422 generates a signal in a frequency domain by performing frequency transform. In this case, the frequency transform can be performed by ERB (equivalent rectangular bandwidth) frequency scale, by which the present invention is non-limited.
The loudness estimating part 426 estimates a loudness of sound by performing a perceptual (psychoacoustic) loudness analysis on an input signal and generates a target gain value from the loudness. In this case, the loudness is an auditory sense actually perceived by a man for each sound pressure level and has a concept (unit of phon) of representing a sound intensity of sound actually recognized by a human ear. In particular, the loudness has the concept discriminated from a sound intensity (unit of dB) determined by a sound pressure level. And, the loudness depends on a lasting time of sound, an occurring time of sound, a spectrum characteristic and the like as well as the sound intensity. Perceptual loudness analysis is to analyze loudness of an input signal based on an equal loudness curve which is the result of a human test. Thus, if loudness per frequency band is obtained, a target gain is obtained based on the loudness and a target volume level. In this case, the target volume level is a full-band gain corresponding to a user inputted volume command or device setting information and may equal to the former target level value explained together with the former embodiment. The target gain is a target value of volume in consideration of loudness and may include a value that varies per frequency band. Meanwhile, the target gain can be re-adjusted to raise a sound quality.
The gain processing part 424 applies the target gain to the frequency transformed input signal. The time-frequency synthesis part 426 frequency-synthesizes the gain processed signal.
FIG. 10 is a block diagram of an audio signal decoding apparatus according to an embodiment of the present invention.
Referring to FIG. 10, an audio signal decoding apparatus 800 can include a demultiplexer 810, an audio signal decoder 820, a band extension decoder 830, a band expansion decoding apparatus 840 and a multi-channel decoder 850. And, the audio signal decoding apparatus 860 further includes an audio signal processing apparatus 860 according to the present invention. The audio signal processing apparatus 860 corresponds to one embodiment of the former audio signal processing apparatus described with reference to FIG. 1 or can correspond to another embodiment (first to fourth examples) of the former audio signal processing apparatus described with reference to FIG. 7 or FIG. 8.
The demultiplexer 810 extracts spectral data, band expansion information, spatial information and the like from an audio signal bitstream.
The audio signal decoder 820 decodes the spectral data by an audio coding scheme if the spectral data corresponding to a downmix signal has a large audio characteristic. In this case, as mentioned in the foregoing description, the audio coding scheme can follow AAC standard, HE-AAC standard or the like. Meanwhile, the audio signal decoder 820 reconstructs a frequency transformed residual by performing dequantization on the spectral data and scale factor carried on the bitstream. Subsequently, the audio signal decoder 820 generates a residual (inverse transformed) by performing inverse frequency transform on the frequency transformed residual.
The speech signal decoder 830 decodes the downmix signal by a speech coding scheme if the spectral data has a large speech characteristic. In this case, as mentioned in the foregoing description, the speech coding scheme can follow the AMR-WB (adaptive multi-rate wide-band) standard, by which the present invention is non-limited.
The band expansion decoder 840 decodes a band extension information bitstream and then generates an audio signal (or, spectral data) of another band (e.g., high frequency band) from a portion or all of the audio signal (or, spectral data) using this information.
If the decoded audio signal is a downmix, the multi-channel decoder 850 generates an output channel signal of a multi-channel signal (stereo signal included) using the spatial information.
As mentioned in the foregoing description, the audio signal processing apparatus 860 is the former audio signal processing apparatus according to an embodiment of the present invention described with reference to FIG. 1, FIG. 7 or FIG. 8. And, the audio signal processing apparatus 860 performs such processing as noise canceling, normalizing, volume controlling and the like on the output channel signal from the multi-channel decoder.
The audio signal processing apparatus according to the present invention is available for various products to use. Theses products can be grouped into a stand alone group and a portable group. A TV, a monitor, a settop box and the like belong to the stand alone group. And, a PMP, a mobile phone, a navigation system and the like belong to the portable group.
FIG. 11 is a block diagram of a product in which an audio signal processing apparatus according to an embodiment of the present invention is implemented, and FIG. 12 is a diagram for relations between products in which an audio signal processing apparatus according to an embodiment of the present invention is implemented.
Referring to FIG. 11, a wire/wireless communication unit 910 receives a bitstream via wire/wireless communication system. In particular, the wire/wireless communication unit 910 can include at least one of a wire communication unit 910A, an infrared communication unit 910B, a Bluetooth unit 910C and a wireless LAN communication unit
910D.
A user authenticating unit 920 receives an input of user information and then performs user authentication. The user authenticating unit 920 can include at least one of a fingerprint recognizing unit 920A, an iris recognizing unit 920B, a face recognizing unit 920C and a speech recognizing unit 920D. The fingerprint recognizing unit 920A, the iris recognizing unit 920B, the face recognizing unit 920C and the speech recognizing unit
920D receive fingerprint information, iris information, face contour information and speech information and then convert them into user informations, respectively. Whether each of the user informations matches pre-registered user data is determined to perform the user authentication.
An input unit 930 is an input device enabling a user to input various kinds of commands and can include at least one of a keypad unit 930A, a touchpad unit 930B and a remote controller unit 930C, by which the present invention is non-limited. A signal coding unit 940 performs encoding or decoding on an audio signal and/or a video signal, which is received via the wire/wireless communication unit 910, and then output an audio signal in time domain. The signal coding unit 940 includes an audio signal processing apparatus 945.
As mentioned in the foregoing description, the audio signal processing apparatus 945 corresponds to one of the former audio signal processing apparatuses according to one embodiment and other embodiments (first to fourth examples) of the present invention. And, the audio signal processing apparatus 945 performs such processing as noise canceling, normalizing, volume controlling and the like on an audio signal before being outputted via the output unit. Thus, the audio signal processing apparatus and the signal coding unit including the audio signal processing apparatus can be implemented by at least one or more processors. A control unit 950 receives input signals from input devices and controls all processes of the signal decoding unit 940 and an output unit 960. In particular, the output unit 960 is an element configured to output an output signal generated by the signal decoding unit 940 and the like and can include a speaker unit 960A and a display unit 960B. If the output signal is an audio signal, it is outputted to a speaker. If the output signal is a video signal, it is outputted via a display.
FIG. 12 shows the relation between the terminal corresponding to the product shown in FIG. 11 and a server.
Referring to (A) of FIG. 12, it can be observed that a first terminal 900.1 and a second terminal 900.2 can exchange data or bitstreams bi-directionally with each other via the wire/wireless communications units.
Referring to FIG. (B) of FIG. 12, it can be observed that a server 1000 and a first terminal 900.1 can perform wire/wireless communication with each other.
An audio signal processing method according to the present invention can be implemented into a computer-executable program and can be stored in a computer- readable recording medium. And, multimedia data having a data structure of the present invention can be stored in the computer-readable recording medium. The computer- readable media include all kinds of recording devices in which data readable by a computer system are stored. The computer-readable media include ROM, RAM, CD-ROM, magnetic tapes, floppy discs, optical data storage devices, and the like for example and also include carrier- wave type implementations (e.g., transmission via Internet). And, a bitstream generated by the above encoding method can be stored in the computer-readable recording medium or can be transmitted via wire/wireless communication network.
INDUSTRIAL APPLICABILITY Accordingly, the present invention is applicable to processing and outputting an audio signal.
While the present invention has been described and illustrated herein with reference to the preferred embodiments thereof, it will be apparent to those skilled in the art that various modifications and variations can be made therein without departing from the spirit and scope of the invention. Thus, it is intended that the present invention covers the modifications and variations of this invention that come within the scope of the appended claims and their equivalents.

Claims

rCLAIMSl
1. A method for processing an audio signal, comprising: receiving a signal, by an audio processing apparatus; computing a long-term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.
2. The method of claim 1, wherein the final gain is generated by multiplying the slow gain by the fast gain.
3. The method of claim 1, wherein the fast gain is generated using at least one of a limiter threshold, a compressor threshold, and an expander threshold.
4. The method of claim 1, wherein the slow gain and the fast gain are generated by being smoothed over time, and wherein the slow gain and the fast gain are within a limit range.
5. The method of claim 1, wherein the slow gain, the fast gain and the final gain correspond to a full-band gain.
6. The method of claim 1, wherein the slow gain and the fast gain are generated using a target level value included in at least one of user input information and device setting information.
7. The method of claim 1, further comprising: generating a feedback information using the final gain, wherein the feedback information limit is associated with at least one of a minimum value and a maximum value of gain filter for noise canceling.
8. An apparatus for processing an audio signal, comprising: a power estimating part computing a long-term power and a short- term power by estimating power of a signal; a slow gain generating part generating a slow gain based on the long- term power; a fast gain generating part generating a fast gain based on the short- term power; a final gain obtaining part obtaining a final gain by combining the slow gain and the fast gain; and, an output signal generating part modifying the signal using the final gain.
9. The apparatus of claim 8, wherein the final gain is generated by multiplying the slow gain by the fast gain.
10. The apparatus of claim 8, wherein the fast gain is generated using at least one of a limiter threshold, a compressor threshold, and an expander threshold.
11. The apparatus of claim 8, wherein the slow gain and the fast gain are generated by being smoothed over time, and wherein the slow gain and the fast gain are within a limit range.
12. The apparatus of claim 8, wherein the slow gain, the fast gain and the final gain correspond to a full-band gain.
13. The apparatus of claim 8, wherein the slow gain and the fast gain are generated using a target level value included in at least one of user input information and device setting information.
14. The apparatus of claim 8, further comprising a feedback information generating part generating feedback information using the final gain, wherein the feedback information is associated with at least one of a minimum value and a maximum value of gain filter for noise canceling.
15. A computer-readable medium having instructions stored thereon, which, when executed by a processor, causes the processor to perform operations, comprising: receiving a signal, by an audio processing apparatus; computing a long-term power and a short-term power by estimating power of the signal; generating a slow gain based on the long-term power; generating a fast gain based on the short-term power; obtaining a final gain by combining the slow gain and the fast gain; and, modifying the signal using the final gain.
PCT/KR2009/004220 2008-07-29 2009-07-29 An apparatus for processing an audio signal and method thereof WO2010013941A2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2009801148963A CN102016994B (en) 2008-07-29 2009-07-29 An apparatus for processing an audio signal and method thereof

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
US8426508P 2008-07-29 2008-07-29
US61/084,265 2008-07-29
US15838809P 2009-03-08 2009-03-08
US61/158,388 2009-03-08
US16445909P 2009-03-29 2009-03-29
US61/164,459 2009-03-29

Publications (2)

Publication Number Publication Date
WO2010013941A2 true WO2010013941A2 (en) 2010-02-04
WO2010013941A3 WO2010013941A3 (en) 2010-06-24

Family

ID=41119934

Family Applications (2)

Application Number Title Priority Date Filing Date
PCT/KR2009/004218 WO2010013939A2 (en) 2008-07-29 2009-07-29 An apparatus for processing an audio signal and method thereof
PCT/KR2009/004220 WO2010013941A2 (en) 2008-07-29 2009-07-29 An apparatus for processing an audio signal and method thereof

Family Applications Before (1)

Application Number Title Priority Date Filing Date
PCT/KR2009/004218 WO2010013939A2 (en) 2008-07-29 2009-07-29 An apparatus for processing an audio signal and method thereof

Country Status (5)

Country Link
US (2) US8275154B2 (en)
EP (2) EP2149985B1 (en)
CN (2) CN102016994B (en)
TW (2) TWI422147B (en)
WO (2) WO2010013939A2 (en)

Families Citing this family (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7742746B2 (en) * 2007-04-30 2010-06-22 Qualcomm Incorporated Automatic volume and dynamic range adjustment for mobile audio devices
US9020158B2 (en) * 2008-11-20 2015-04-28 Harman International Industries, Incorporated Quiet zone control system
US8135140B2 (en) 2008-11-20 2012-03-13 Harman International Industries, Incorporated System for active noise control with audio signal compensation
US8718289B2 (en) 2009-01-12 2014-05-06 Harman International Industries, Incorporated System for active noise control with parallel adaptive filter configuration
US8189799B2 (en) * 2009-04-09 2012-05-29 Harman International Industries, Incorporated System for active noise control based on audio system output
US8199924B2 (en) * 2009-04-17 2012-06-12 Harman International Industries, Incorporated System for active noise control with an infinite impulse response filter
US8077873B2 (en) * 2009-05-14 2011-12-13 Harman International Industries, Incorporated System for active noise control with adaptive speaker selection
JP4982617B1 (en) * 2011-06-24 2012-07-25 株式会社東芝 Acoustic control device, acoustic correction device, and acoustic correction method
TWI466014B (en) * 2011-08-19 2014-12-21 Acer Inc Method and control device for establishing connection and status relationship between at least two devices
JP5849677B2 (en) * 2011-12-15 2016-02-03 船井電機株式会社 Audio output device
KR101945816B1 (en) 2012-06-08 2019-02-11 삼성전자주식회사 Device and method for adjusting volume in terminal
US8873782B2 (en) 2012-12-20 2014-10-28 Starkey Laboratories, Inc. Separate inner and outer hair cell loss compensation
US20160065160A1 (en) * 2013-03-21 2016-03-03 Intellectual Discovery Co., Ltd. Terminal device and audio signal output method thereof
JP6323089B2 (en) * 2014-03-14 2018-05-16 ヤマハ株式会社 Level adjusting method and level adjusting device
FR3020711B1 (en) * 2014-05-02 2016-05-13 3Db COMPRESSION METHOD AND AUDIO DYNAMIC COMPRESSOR
CN110808723A (en) * 2014-05-26 2020-02-18 杜比实验室特许公司 Audio signal loudness control
EP3133736A1 (en) * 2015-08-17 2017-02-22 Nxp B.V. Gain control apparatus
CN105322904B (en) * 2015-09-08 2017-03-22 湖南国科微电子股份有限公司 Method for controlling level of digital audio signal
US10733971B2 (en) * 2016-06-13 2020-08-04 Sony Corporation Sound processing device, sound processing method, and computer program
WO2018130287A1 (en) * 2017-01-12 2018-07-19 Sonova Ag Hearing device with acoustic shock control and method for acoustic shock control in a hearing device
WO2018201374A1 (en) * 2017-05-04 2018-11-08 Harman International Industries, Incorporated Method and device for adjusting audio signal, and audio system
CN109854230B (en) * 2017-11-30 2022-05-10 中国石油天然气股份有限公司 Well testing method and device
CN110631691B (en) * 2019-09-09 2021-06-11 国网湖南省电力有限公司 Power equipment noise separation effect verification method and system, separation equipment and medium
CN114125658B (en) * 2020-08-25 2023-12-19 上海艾为电子技术股份有限公司 Dynamic range control circuit, audio processing chip and audio processing method thereof
CN114915275A (en) * 2021-02-08 2022-08-16 祖玛视频通讯公司 Two-stage digital automatic gain control

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6122384A (en) * 1997-09-02 2000-09-19 Qualcomm Inc. Noise suppression system and method
US20050228651A1 (en) * 2004-03-31 2005-10-13 Microsoft Corporation. Robust real-time speech codec
EP1729287A1 (en) * 1999-01-07 2006-12-06 Tellabs Operations, Inc. Method and apparatus for adaptively suppressing noise
US20070124139A1 (en) * 2000-10-25 2007-05-31 Broadcom Corporation Method and apparatus for one-stage and two-stage noise feedback coding of speech and audio signals

Family Cites Families (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US96762A (en) * 1869-11-16 Improved horseshoe-machine
US4490843A (en) * 1982-06-14 1984-12-25 Bose Corporation Dynamic equalizing
DE3689035T2 (en) * 1985-07-01 1994-01-20 Motorola Inc NOISE REDUCTION SYSTEM.
US4739514A (en) 1986-12-22 1988-04-19 Bose Corporation Automatic dynamic equalizing
US5481615A (en) 1993-04-01 1996-01-02 Noise Cancellation Technologies, Inc. Audio reproduction system
US5768473A (en) * 1995-01-30 1998-06-16 Noise Cancellation Technologies, Inc. Adaptive speech filter
US5832097A (en) 1995-09-19 1998-11-03 Gennum Corporation Multi-channel synchronous companding system
US6965676B1 (en) * 1999-10-19 2005-11-15 Texas Instruments Incorporated Volume-responsive loudness compensation circuits, systems, and methods
WO2001050459A1 (en) 1999-12-31 2001-07-12 Octiv, Inc. Techniques for improving audio clarity and intelligibility at reduced bit rates over a digital network
JP3964092B2 (en) * 2000-02-17 2007-08-22 アルパイン株式会社 Audio adaptive equalizer and filter coefficient determination method
US7016509B1 (en) 2000-09-08 2006-03-21 Harman International Industries, Inc. System and method for varying low audio frequencies inversely with audio signal level
EP1835487B1 (en) 2003-05-28 2013-07-10 Dolby Laboratories Licensing Corporation Method, apparatus and computer program for calculating and adjusting the perceived loudness of an audio signal
CN1795490A (en) * 2003-05-28 2006-06-28 杜比实验室特许公司 Method, apparatus and computer program for calculating and adjusting the perceived loudness of an audio signal
US20050090295A1 (en) * 2003-10-14 2005-04-28 Gennum Corporation Communication headset with signal processing capability
US20060106472A1 (en) * 2004-11-16 2006-05-18 Romesburg Eric D Method and apparatus for normalizing sound recording loudness
WO2006106479A2 (en) 2005-04-08 2006-10-12 Nxp B.V. A method of and a device for processing audio data, a program element and a computer-readable medium
WO2006111370A1 (en) 2005-04-19 2006-10-26 Epfl (Ecole Polytechnique Federale De Lausanne) A method and device for removing echo in a multi-channel audio signal
US8275153B2 (en) * 2007-04-16 2012-09-25 Evertz Microsystems Ltd. System and method for generating an audio gain control signal
KR101469739B1 (en) 2007-06-21 2014-12-05 코닌클리케 필립스 엔.브이. A device for and a method of processing audio signals

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6122384A (en) * 1997-09-02 2000-09-19 Qualcomm Inc. Noise suppression system and method
EP1729287A1 (en) * 1999-01-07 2006-12-06 Tellabs Operations, Inc. Method and apparatus for adaptively suppressing noise
US20070124139A1 (en) * 2000-10-25 2007-05-31 Broadcom Corporation Method and apparatus for one-stage and two-stage noise feedback coding of speech and audio signals
US20050228651A1 (en) * 2004-03-31 2005-10-13 Microsoft Corporation. Robust real-time speech codec

Also Published As

Publication number Publication date
TW201106618A (en) 2011-02-16
EP2149986B1 (en) 2017-10-25
EP2149985A1 (en) 2010-02-03
TWI422147B (en) 2014-01-01
WO2010013939A2 (en) 2010-02-04
TW201030736A (en) 2010-08-16
WO2010013941A3 (en) 2010-06-24
CN102016995A (en) 2011-04-13
EP2149986A1 (en) 2010-02-03
CN102016994B (en) 2013-07-17
EP2149985B1 (en) 2013-04-03
CN102016995B (en) 2013-07-10
US20100085117A1 (en) 2010-04-08
WO2010013939A3 (en) 2010-06-24
US20100061566A1 (en) 2010-03-11
US8275154B2 (en) 2012-09-25
US8275150B2 (en) 2012-09-25
TWI397058B (en) 2013-05-21
CN102016994A (en) 2011-04-13

Similar Documents

Publication Publication Date Title
EP2149985B1 (en) An apparatus for processing an audio signal and method thereof
JP6633239B2 (en) Loudness adjustment for downmixed audio content
US9361901B2 (en) Integrated speech intelligibility enhancement system and acoustic echo canceller
US8219389B2 (en) System for improving speech intelligibility through high frequency compression
US10871940B2 (en) Systems and methods for sound enhancement in audio systems
US20130030800A1 (en) Adaptive voice intelligibility processor
EP3107097A1 (en) Improved speech intelligilibility
EP2828853B1 (en) Method and system for bias corrected speech level determination
CN109841223B (en) Audio signal processing method, intelligent terminal and storage medium
GB2536727B (en) A speech processing device
Brouckxon et al. Time and frequency dependent amplification for speech intelligibility enhancement in noisy environments

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 200980114896.3

Country of ref document: CN

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

Ref document number: 09803146

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09803146

Country of ref document: EP

Kind code of ref document: A2