WO2013015887A1 - Apparatus and method for automatic gain control - Google Patents
Apparatus and method for automatic gain control Download PDFInfo
- Publication number
- WO2013015887A1 WO2013015887A1 PCT/US2012/041024 US2012041024W WO2013015887A1 WO 2013015887 A1 WO2013015887 A1 WO 2013015887A1 US 2012041024 W US2012041024 W US 2012041024W WO 2013015887 A1 WO2013015887 A1 WO 2013015887A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- speech signal
- speech
- signal
- output
- dynamic adjustment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03G—CONTROL OF AMPLIFICATION
- H03G3/00—Gain control in amplifiers or frequency changers
- H03G3/20—Automatic control
- H03G3/30—Automatic control in amplifiers having semiconductor devices
- H03G3/3089—Control of digital or coded signals
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03G—CONTROL OF AMPLIFICATION
- H03G3/00—Gain control in amplifiers or frequency changers
- H03G3/20—Automatic control
- H03G3/30—Automatic control in amplifiers having semiconductor devices
- H03G3/3005—Automatic control in amplifiers having semiconductor devices in amplifiers suitable for low-frequencies, e.g. audio amplifiers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M1/00—Substation equipment, e.g. for use by subscribers
- H04M1/60—Substation equipment, e.g. for use by subscribers including speech amplifiers
- H04M1/6008—Substation equipment, e.g. for use by subscribers including speech amplifiers in the transmitter circuit
Definitions
- the invention relates generally to providing voice signals to users, and, more specifically to providing automatic gain control for these signals.
- AGC Automatic gain control
- Still another problem associated with previous AGC approaches is the tendency to over or under amplify the input signal under certain conditions, creating a non- desirable output. More specifically and when normalizing audio signals, unintentional over- amplification can produce a noticeably unpleasant (e.g., noisy) output. When used in vehicle -based hands-free applications where an AGC algorithm is intended to normalize speech, over amplified signals can be unintentionally produced by noise and other sounds common to the environment in the input signal that are misrecognised as speech. [0005]
- the problems described above result in the limited use of AGC algorithms in automobile and other noisy applications. And when used, user dissatisfaction with these approaches has been high.
- FIG. 1 comprises a block diagram of an apparatus that provides automatic gain control according to various embodiments of the present invention
- FIG. 2 comprises a flowchart of an approach for providing automatic gain control according to various embodiments of the present invention
- FIG. 3 comprises a flowchart of an approach for providing automatic gain control according to various embodiments of the present invention
- FIG. 4 comprises a flowchart showing an approach for providing automatic gain control according to various embodiments of the present invention
- FIG. 5 comprises a graph showing the provision of gain control according to various embodiments of the present invention.
- AGC AGC algorithms that are particularly applicable and effective in noisy environments (e.g., vehicular applications) are provided. Some of these approaches normalize the level speech signals instantaneously without adding any or substantially any distortion.
- AGC approaches tend to increase the noise power level much more than speech signals which is not desired. Consequently, many of the present approaches employ a robust noise adaptive mechanism that controls the amplification of noise and, as a result, the noise power level is not increased more than the speech power level.
- the approaches described herein are also applicable to a wide range of applications such as hands-free communications in vehicles. Many of these approaches are based upon calculations involving the background noise estimate, slow and fast envelope tracking of speech signals, speech pause detection, and speech peak power estimate, among other factors.
- AGC Automatic Control
- a speech signal is received at an input. At least one electrical value associated with the received speech signal is tracked. A dynamic adjustment of the speech signal is determined. The dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis of the at least one electrical value. The dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal presented at an output. The dynamic adjustment value is applied to the speech signal and the adjusted speech signal is presented at the output.
- a power level is measured that is associated with an electrical envelope.
- a speech pause associated with an electrical envelope is detected.
- the tracking may include tracking at least one difference between a power level associated with the speech signal and a desired power level.
- the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level.
- Other tracking approaches may also be used.
- the analysis may include a variety of different calculations.
- the analysis includes a computation of a maximum gain based upon a desired output divided the input and applying the dynamic adjustment value includes limiting the gain of the adjusted speech signal at the output to substantially the maximum computed gain.
- the speech signal is transmitted to an audio application in a vehicle.
- the audio application may be a speaker associated with any application such as hands-free applications in vehicles.
- a speech signal is received. Based upon at least one characteristic of the speech signal and at least one characteristic of the desired output signal, a maximum gain so as to limit over amplification of the received speech signal is determined. The determined maximum gain is applied to the signal so as to limit over- amplification of the input signal.
- the maximum gain is the gain that would be effective to saturate the speech signal.
- saturation it is meant .
- the maximum gain is the maximum desired output divided by a maximum magnitude of the received speech signal.
- the interface includes an input and an output and the input is configured to receive a speech signal.
- the controller is coupled to the interface and configured to track at least one electrical value associated with the received speech signal.
- the controller is further configured to determine a dynamic adjustment of the speech signal and the dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over- amplification of the speech signal based at least in part upon an analysis by the controller of the at least one electrical value.
- the dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal to be presented at the output.
- the controller is configured to apply the dynamic adjustment value to the speech signal and present the adjusted speech signal at the output of the interface.
- the controller is configured to track a power level measured associated with an electrical envelope. In other aspects, the controller is configured to detect a speech pause associated with an electrical envelope. In yet other aspects, the controller is configured to track at least one difference between a power level associated with the speech signal and a desired power level, and the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level.
- the analysis performs a computation of a maximum gain based upon a desired output divided the input and the gain of the adjusted speech signal is limited at the output to substantially the maximum gain.
- the output is coupled to an audio application in a vehicle.
- the audio application is selected from the group consisting of a speaker and a microphone.
- Other examples of audio applications and environments are possible.
- the approaches described herein are often described as being implemented in vehicular environments, they may be applied to any environment and are in no way limited to vehicular environments.
- an apparatus 100 configured to control automatic gain control for speech signals includes an interface 102 and a controller 104.
- the interface 102 includes an input 106 and an output 108 and the input 106 is configured to receive a speech signal 1 10.
- the controller 104 is coupled to the interface 102 and is configured to track at least one electrical value associated with the received speech signal 1 10.
- the controller 104 is further configured to determine a dynamic adjustment of the speech signal 1 10 and the dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal 1 10 based at least in part upon an analysis by the controller 104 of the at least one electrical value.
- the dynamic adjustment is further selected to obtain a desired output signal characteristic for the modified speech signal 1 12 to be presented at the output 108.
- the controller 104 is configured to apply the dynamic adjustment value to the speech signal 1 10 and present the adjusted speech signal 1 12 at the output 108 of the interface 102.
- the controller 104 is configured to track a power level measured associated with an electrical envelope.
- an electrical envelope it is meant
- the controller 104 is configured to detect a speech pause associated with the electrical envelope. By “speech pause,” it is meant t . In yet other aspects, the controller 104 is configured to track at least one difference between a power level associated with the speech signal 1 10 and a desired power level, and the dynamic adjustment is effective to maintain the signal at the output signal 1 12 at substantially the desired power level.
- the controller 104 computes a maximum gain based upon a desired output divided the input.
- the gain of the adjusted speech signal 1 12 is limited at the output to substantially the maximum gain.
- the output 108 is coupled to an audio application 1 14 in a vehicle.
- the audio application may be or may utilize a speaker and a microphone as used, for example, in hands-free vehicle applications.
- the speech signal 1 10 is scaled very effectively without depending on the level of input speech (e.g., approximately 40 dB/sec of gain is applied to the speech signal 1 10 where is necessary without adding any noticeable distortion to speech).
- the relative power ratio between the input speech e.g., approximately 40 dB/sec of gain is applied to the speech signal 1 10 where is necessary without adding any noticeable distortion to speech.
- instantaneous speech power level and desired speech power level may be compared.
- the AGC scaling rate is determined.
- possible AGC scaling rates may be split into multiple regions and then the actual rate selected from these regions and according to various conditions including the above- mentioned ratio.
- the AGC scaling rate is split into three regions: (a) a low AGC scaling rate; (b) a medium AGC scaling rate; and (c) a high AGC scaling rate.
- a low AGC scaling rate means that the instantaneous speech power and desired speech power levels are very close (e.g., within of each other); therefore, the AGC gain applied to signal vary very slowly from one frame to another (e.g., applying 0.5 dB per second AGC gain).
- a medium AGC scaling rate means that the instantaneous speech power and desired speech power levels are not that close (e.g., within of each other); in this case, AGC gain applied to signal is in the middle range (e.g., applying 7 dB per second AGC gain).
- a high AGC scaling rate means that the difference between the instantaneous speech power level and desired speech power level are very big (e.g., ); in this case, the AGC gain applied to signal should be in the high range (e.g., applying 40 dB per second AGC gain). In these approaches speech levels are normalized as effectively as possible in a short period of time and, hence, providing the desire speech level at the output of HF system.
- speech pauses are detected based on the slow and fast envelope tracking mechanisms. If the speech pauses are detected, then the signal is not scaled and as a result of this the noise power is not amplified.
- the degree of periodicity measure for speech signals may also be introduced and utilized. Combining both of these methods, the AGC algorithms become very robust in heavy noisy environments and eliminate (or substantially eliminate) the AGC creep problem without amplifying noise while still amplifying speech signals.
- a speech signal is received at an input.
- at least one electrical value associated with the received speech signal is tracked.
- a power level is measured that is associated with an electrical envelope is tracked.
- a speech pause associated with an electrical envelope is tracked.
- the tracking may include tracking at least one difference between a power level associated with the speech signal and a desired power level. In this case and to take one example, the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level.
- a dynamic adjustment of the speech signal is determined.
- the dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis of the at least one electrical value.
- the dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal presented at an output.
- the analysis may utilize a variety of different calculations.
- the analysis includes a computation of a maximum gain based upon a desired output divided the input and applying the dynamic adjustment value comprises limiting the gain of the adjusted speech signal at the output to substantially the maximum computed gain.
- the dynamic adjustment value is applied to the speech signal and the adjusted speech signal is presented at the output.
- the speech signal may then be transmitted to an audio application in a vehicle.
- the audio application may be a speaker associated with any application such as hands-free applications in vehicles. Other examples of audio applications are possible.
- a speech signal is received. Based upon at least one characteristic of the speech signal and at least one characteristic of the desired output signal, at step 304, a maximum gain so as to limit over amplification of the received speech signal is determined.
- the maximum gain comprises a gain that would be effective to saturate the speech signal.
- the maximum gain comprises a maximum desired output divided by a maximum magnitude of the received speech signal.
- G( max ) is Max Amplitude (o)/ Max Amplitude and the input and output signals share a common maximum dynamic range.
- the choice of LL can be arbitrary or determined by knowledge concerning the limits of / and O. For example, if the amplitude of the output signal cannot exceed that of the input signal, LL is OdB (no gain). Thus, the choice of LL reduces under-amplification error.
- Max Amplitude (o) is pre-determined, G will vary with Max Amplitude 3 ⁇ 4i and will be bounded by LL below. It will only change (and only ever downwards) when a higher-than-previously-seen MaxAmplitude ⁇ is encountered for a given input/output set.
- the determined maximum gain is applied to the signal so as to limit over-amplification of the input signal. Since G is now bounded by the above formula, the level of over-amplification the AGC process can apply is limited, reducing its effect. The effect is most pronounced when the signal to be normalized (e.g. speech) is substantially greater in magnitude than the signals that may be misrecognised and normalized erroneously (e.g. noise).
- An example application would be the normalization of the signal received from a telephony device where the input is normalized to maximize use of the dynamic range of the output signal.
- LL will be OdB and G will be limited by the maximum amplitude of the input signal, which may range from full-scale (limiting G to OdB) to a lower (normally at most 6dB, or half) value.
- the AGC since the AGC will be operating near G (max) (to normalize speech to full scale), the AGC will be prevented from substantially over-amplifying smaller signals (such as noise) that are misrecognised as speech.
- FIG. 4 utilizes a speech power computation module 402, a background noise estimator, a slow envelope tracking module 406, a speech pause detection module 408, a fast envelope tracking module 410, a peak power estimator 412, a sudden power change detection module 414, an AGC rate module 416, an AGC gain computational module 418, and a summer 420.
- the speech power computation module 402 computes power of the speech.
- the average power, P 0 of a speech segment, s m (n) is computed for the m' h speech frame as follows:
- the slow envelope tracking module 406 provides for the slow envelope tracking, first order IIR smoothing of the average power for the input speech signal.
- the slow envelope tracking, P s (m) for the m' h frame is computed as given in the following equation:
- ⁇ 5 (m) The time variant constant, ⁇ 5 (m) for the IIR filter is chosen differentially for rising and falling signal edges as given in the following equation:
- a rising signal edge is followed faster than a falling one as:
- the fast envelope tracking module 410 performs the fast envelope tracking
- the background noise estimator module 404 provides for estimation of the background noise level, minimum statistics and power smoothing only during speech pauses is used for the AGC algorithm. For estimation of the long term average power of the signal,
- first order IIR can be utilized to smooth the instantaneous power of speech signals as given in the following:
- ⁇ is a small positive value that controls the maximum speed for increasing the estimated noise level.
- [0073] describes the maximum background noise power increase in decibels (dB) per second where f s is the sampling frequency.
- ⁇ should be chosen such that ⁇ ( ⁇ ) is within the interval 0.5 dB/second and 3 dB/second.
- the speech pause detection module 408 compares the outputs of the fast and slow envelope tracking estimators and is the first stage of a simple speech pause detector. At the beginning of speech interval, the output of the fast envelope tracking estimator rises faster than that of the slow one. Due to short term power variations during speech activity on one hand and the choices of the time constants on the other hand, the output signal of the fast envelope tracking estimator is always larger than its slow counterpart. At the end of the voice activity period, the output of the slow envelope tracking estimator does not decrease as rapidly as does that of the fast envelope tracking estimator, and now the output of the slow envelope tracking estimator exceeds that of the fast one.
- T(m) max ⁇ p s (m), ⁇ B N (m) ⁇ (13)
- the sudden power change detection module 414 and AGC rate determination module 416 are used to determine the AGC rate that will be applied during AGC gain computation. Additionally, the sudden power change detection pre-determined variety of AGC variables (i.e., peak power estimate, slow envelope tracking, fast envelope tracking, and background noise estimate) and AGC gain are used in order to more effectively eliminate the sudden power change affects in a short period of time.
- AGC variables i.e., peak power estimate, slow envelope tracking, fast envelope tracking, and background noise estimate
- the AGC gain computation module 418 performs signal scaling. During speech intervals the short term power of the input speech signal is estimated and compared to a desired value. If the speech level exceeds an upper bound, the gain factor, G(m) for the m' h frame is decreased - in the other case the gain is increased. In order to avoid fast gain variations during short speech pauses, G(m) is incremented or decremented in very small steps. The gain factors, G(m) is updated only during the periods of speech activity as:
- the scaling factor, p is computed by comparing the amplified or attenuated
- ⁇ is a small positive value that controls the maximum speed for increasing or decreasing the AGC gain in a second.
- the summer 420 sums the gain from the module 418 and the input signal and this is made available at the output. More specifically, the output speech signal of AGC algorithm for the m' frame, s 0 (m, n) is computed by multiplying the input speech signal for the m' h frame, s m (n) with the AGC gain factor, G(m) as:
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Control Of Amplification And Gain Control (AREA)
Abstract
A speech signal is received at an input. At least one electrical value associated with the received speech signal is tracked. A dynamic adjustment of the speech signal is determined. The dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis of the at least one electrical value. The dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal presented at an output. The dynamic adjustment value is applied to the speech signal and the adjusted speech signal is presented at the output. The gain of the signal can also be limited to prevent over-amplification.
Description
APPARATUS AND METHOD FOR AUTOMATIC GAIN CONTROL
FIELD OF THE INVENTION
[0001] The invention relates generally to providing voice signals to users, and, more specifically to providing automatic gain control for these signals.
BACKGROUND OF THE INVENTION
[0002] Automatic gain control (AGC) systems are generally designed to provide a normalized speech level at the output of the system by applying a gain to voice signals received at an input. It is typically desirable that this gain is made as a function of time and that this gain is varied very slowly from one frame to another. However, if the gain that will be applied to speech signal is high, then it will take very long time for AGC to settle to the normalized desired speech level. Unfortunately, many previous AGC algorithms suffered in effectiveness by providing normalized speech very quickly without considering the level of speech signal at the input.
[0003] Another problem with previous AGC systems involves noisy environments such as automobile applications. In these applications, where background noise is often high, communications sometimes become impossible when AGC algorithms amplify noise signals more than speech signals. This is a considerable problem in vehicle-to-vehicle
communications, which results in an increased noise loop back problem from one vehicle to another, often making effective communications difficult or impossible to achieve.
[0004] Still another problem associated with previous AGC approaches is the tendency to over or under amplify the input signal under certain conditions, creating a non- desirable output. More specifically and when normalizing audio signals, unintentional over- amplification can produce a noticeably unpleasant (e.g., noisy) output. When used in vehicle -based hands-free applications where an AGC algorithm is intended to normalize speech, over amplified signals can be unintentionally produced by noise and other sounds common to the environment in the input signal that are misrecognised as speech.
[0005] The problems described above result in the limited use of AGC algorithms in automobile and other noisy applications. And when used, user dissatisfaction with these approaches has been high.
BRIEF DESCRIPTION OF THE DRAWINGS
[0006] The present invention is illustrated, by way of example and not limitation, in the accompanying figures, in which like reference numerals indicate similar elements, and in which:
[0007] FIG. 1 comprises a block diagram of an apparatus that provides automatic gain control according to various embodiments of the present invention;
[0008] FIG. 2 comprises a flowchart of an approach for providing automatic gain control according to various embodiments of the present invention;
[0009] FIG. 3 comprises a flowchart of an approach for providing automatic gain control according to various embodiments of the present invention;
[0010] FIG. 4 comprises a flowchart showing an approach for providing automatic gain control according to various embodiments of the present invention;
[0011] FIG. 5 comprises a graph showing the provision of gain control according to various embodiments of the present invention.
[0012] Skilled artisans will appreciate that elements in the figures are illustrated for simplicity and clarity and have not necessarily been drawn to scale. For example, the dimensions and/or relative positioning of some of the elements in the figures may be exaggerated relative to other elements to help to improve understanding of various embodiments of the present invention. Also, common but well-understood elements that are useful or necessary in a commercially feasible embodiment are often not depicted in order to facilitate a less obstructed view of these various embodiments of the present invention. It will further be appreciated that certain actions and/or steps may be described or depicted in a particular order of occurrence while those skilled in the art will understand that such specificity with respect to sequence is not actually required. It will also be understood that the terms and expressions used herein have the ordinary meaning as is accorded to such
terms and expressions with respect to their corresponding respective areas of inquiry and study except where specific meanings have otherwise been set forth herein.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] In the approaches described herein, noise-adaptive Automatic Gain Control
(AGC) algorithms that are particularly applicable and effective in noisy environments (e.g., vehicular applications) are provided. Some of these approaches normalize the level speech signals instantaneously without adding any or substantially any distortion.
[0014] In noisy environments, AGC approaches tend to increase the noise power level much more than speech signals which is not desired. Consequently, many of the present approaches employ a robust noise adaptive mechanism that controls the amplification of noise and, as a result, the noise power level is not increased more than the speech power level.
[0015] The approaches described herein are also applicable to a wide range of applications such as hands-free communications in vehicles. Many of these approaches are based upon calculations involving the background noise estimate, slow and fast envelope tracking of speech signals, speech pause detection, and speech peak power estimate, among other factors.
[0016] Many of these approaches prevent over-amplification by an Automatic Gain
Control (AGC) algorithm by pre-determining the maximum allowed amplitude of the output signal. If an absolute limit can be established for the output signal, over-amplification can be limited or eliminated, thereby reducing its negative impact on the system.
[0017] In many of these embodiments, a speech signal is received at an input. At least one electrical value associated with the received speech signal is tracked. A dynamic adjustment of the speech signal is determined. The dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis of the at least one electrical value. The dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal presented at an output. The dynamic adjustment value is applied to the speech signal
and the adjusted speech signal is presented at the output.
[0018] In some aspects, a power level is measured that is associated with an electrical envelope. In other aspects, a speech pause associated with an electrical envelope is detected.
[0019] The tracking may include tracking at least one difference between a power level associated with the speech signal and a desired power level. In this case and to take one example, the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level. Other tracking approaches may also be used.
[0020] The analysis may include a variety of different calculations. In one aspect, the analysis includes a computation of a maximum gain based upon a desired output divided the input and applying the dynamic adjustment value includes limiting the gain of the adjusted speech signal at the output to substantially the maximum computed gain.
[0021] In other aspects, the speech signal is transmitted to an audio application in a vehicle. The audio application may be a speaker associated with any application such as hands-free applications in vehicles.
[0022] In others of these embodiments, a speech signal is received. Based upon at least one characteristic of the speech signal and at least one characteristic of the desired output signal, a maximum gain so as to limit over amplification of the received speech signal is determined. The determined maximum gain is applied to the signal so as to limit over- amplification of the input signal.
[0023] In some aspects, the maximum gain is the gain that would be effective to saturate the speech signal. By "saturation", it is meant . In other aspects, the maximum gain is the maximum desired output divided by a maximum magnitude of the received speech signal.
[0024] In still others of these embodiments, an apparatus configured to control automatic gain control for speech signals includes an interface and a controller. The interface includes an input and an output and the input is configured to receive a speech signal.
[0025] The controller is coupled to the interface and configured to track at least one electrical value associated with the received speech signal. The controller is further configured to determine a dynamic adjustment of the speech signal and the dynamic
adjustment is selected at least in part so as to minimize a distortion and minimize an over- amplification of the speech signal based at least in part upon an analysis by the controller of the at least one electrical value. The dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal to be presented at the output. The controller is configured to apply the dynamic adjustment value to the speech signal and present the adjusted speech signal at the output of the interface.
[0026] In some aspects, the controller is configured to track a power level measured associated with an electrical envelope. In other aspects, the controller is configured to detect a speech pause associated with an electrical envelope. In yet other aspects, the controller is configured to track at least one difference between a power level associated with the speech signal and a desired power level, and the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level.
[0027] In some examples, the analysis performs a computation of a maximum gain based upon a desired output divided the input and the gain of the adjusted speech signal is limited at the output to substantially the maximum gain.
[0028] In some aspects, the output is coupled to an audio application in a vehicle. In other aspects, the audio application is selected from the group consisting of a speaker and a microphone. Other examples of audio applications and environments are possible. In this regard, it will be appreciated that although the approaches described herein are often described as being implemented in vehicular environments, they may be applied to any environment and are in no way limited to vehicular environments.
[0029] Referring now to FIG. 1 , an apparatus 100 configured to control automatic gain control for speech signals includes an interface 102 and a controller 104. The interface 102 includes an input 106 and an output 108 and the input 106 is configured to receive a speech signal 1 10.
[0030] The controller 104 is coupled to the interface 102 and is configured to track at least one electrical value associated with the received speech signal 1 10. The controller 104 is further configured to determine a dynamic adjustment of the speech signal 1 10 and the dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal 1 10 based at least in part upon an analysis by the
controller 104 of the at least one electrical value. The dynamic adjustment is further selected to obtain a desired output signal characteristic for the modified speech signal 1 12 to be presented at the output 108. The controller 104 is configured to apply the dynamic adjustment value to the speech signal 1 10 and present the adjusted speech signal 1 12 at the output 108 of the interface 102.
[0031] In some aspects, the controller 104 is configured to track a power level measured associated with an electrical envelope. By an electrical envelope, it is meant
. In other aspects, the controller 104 is configured to detect a speech pause associated with the electrical envelope. By "speech pause," it is meant t . In yet other aspects, the controller 104 is configured to track at least one difference between a power level associated with the speech signal 1 10 and a desired power level, and the dynamic adjustment is effective to maintain the signal at the output signal 1 12 at substantially the desired power level.
[0032] In some examples, the controller 104 computes a maximum gain based upon a desired output divided the input. The gain of the adjusted speech signal 1 12 is limited at the output to substantially the maximum gain.
[0033] In some other aspects, the output 108 is coupled to an audio application 1 14 in a vehicle. In other aspects, the audio application may be or may utilize a speaker and a microphone as used, for example, in hands-free vehicle applications.
[0034] In the above-described approaches, the speech signal 1 10 is scaled very effectively without depending on the level of input speech (e.g., approximately 40 dB/sec of gain is applied to the speech signal 1 10 where is necessary without adding any noticeable distortion to speech). In order to realize this, the relative power ratio between the
instantaneous speech power level and desired speech power level may be compared.
[0035] Based on this ratio, then the AGC scaling rate is determined. In some of these approaches, possible AGC scaling rates may be split into multiple regions and then the actual rate selected from these regions and according to various conditions including the above- mentioned ratio. In one example, the AGC scaling rate is split into three regions: (a) a low AGC scaling rate; (b) a medium AGC scaling rate; and (c) a high AGC scaling rate. In this example, a low AGC scaling rate means that the instantaneous speech power and desired
speech power levels are very close (e.g., within of each other); therefore, the AGC gain applied to signal vary very slowly from one frame to another (e.g., applying 0.5 dB per second AGC gain). A medium AGC scaling rate means that the instantaneous speech power and desired speech power levels are not that close (e.g., within of each other); in this case, AGC gain applied to signal is in the middle range (e.g., applying 7 dB per second AGC gain). A high AGC scaling rate means that the difference between the instantaneous speech power level and desired speech power level are very big (e.g., ); in this case, the AGC gain applied to signal should be in the high range (e.g., applying 40 dB per second AGC gain). In these approaches speech levels are normalized as effectively as possible in a short period of time and, hence, providing the desire speech level at the output of HF system.
[0036] These approaches also improve the robustness of AGC in extremely noisy environments. More specifically, the above-mentioned AGC algorithms are immune or substantially immune to amplifying noise power and, hence, eliminate or substantially eliminate AGC creep for noise signals under heavy noise environments such as in automobile applications. "AGC creep" as used herein means .
[0037] For this purpose, speech pauses are detected based on the slow and fast envelope tracking mechanisms. If the speech pauses are detected, then the signal is not scaled and as a result of this the noise power is not amplified. In order to increase the robustness of these approaches, the degree of periodicity measure for speech signals may also be introduced and utilized. Combining both of these methods, the AGC algorithms become very robust in heavy noisy environments and eliminate (or substantially eliminate) the AGC creep problem without amplifying noise while still amplifying speech signals.
[0038] Referring now to FIG. 2, one example of an approach for providing AGC is described. At step 202, a speech signal is received at an input. At step 204, at least one electrical value associated with the received speech signal is tracked. In some aspects, a power level is measured that is associated with an electrical envelope is tracked. In other aspects, a speech pause associated with an electrical envelope is tracked. The tracking may include tracking at least one difference between a power level associated with the speech signal and a desired power level. In this case and to take one example, the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power
level.
[0039] At step 206, a dynamic adjustment of the speech signal is determined. The dynamic adjustment is selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis of the at least one electrical value. The dynamic adjustment is further selected to obtain a desired output signal characteristic for the speech signal presented at an output. The analysis may utilize a variety of different calculations. In one aspect, the analysis includes a computation of a maximum gain based upon a desired output divided the input and applying the dynamic adjustment value comprises limiting the gain of the adjusted speech signal at the output to substantially the maximum computed gain.
[0040] At step 208, the dynamic adjustment value is applied to the speech signal and the adjusted speech signal is presented at the output. The speech signal may then be transmitted to an audio application in a vehicle. The audio application may be a speaker associated with any application such as hands-free applications in vehicles. Other examples of audio applications are possible.
[0041] Referring now to FIG. 3, another approach for providing AGC is described.
At step 302, a speech signal is received. Based upon at least one characteristic of the speech signal and at least one characteristic of the desired output signal, at step 304, a maximum gain so as to limit over amplification of the received speech signal is determined. In some aspects, the maximum gain comprises a gain that would be effective to saturate the speech signal. In other aspects, the maximum gain comprises a maximum desired output divided by a maximum magnitude of the received speech signal.
[0042] For a given input signal / and a desired output signal O, there exists a bounded gain relationship (G), with a predetermined Lower Limit {LL) that will ensure that O does not exceed a given amplitude:
[0043] LL >= G >= Max Amplitude(0) / Max Amplitude(i)
[0044] Where G(max) is Max Amplitude (o)/ Max Amplitude and the input and output signals share a common maximum dynamic range.
[0045] The choice of LL can be arbitrary or determined by knowledge concerning the limits of / and O. For example, if the amplitude of the output signal cannot exceed that of the
input signal, LL is OdB (no gain). Thus, the choice of LL reduces under-amplification error.
[0046] Since Max Amplitude (o) is pre-determined, G will vary with Max Amplitude ¾i and will be bounded by LL below. It will only change (and only ever downwards) when a higher-than-previously-seen MaxAmplitude^ is encountered for a given input/output set.
[0047] At step 306, the determined maximum gain is applied to the signal so as to limit over-amplification of the input signal. Since G is now bounded by the above formula, the level of over-amplification the AGC process can apply is limited, reducing its effect. The effect is most pronounced when the signal to be normalized (e.g. speech) is substantially greater in magnitude than the signals that may be misrecognised and normalized erroneously (e.g. noise).
[0048] An example application would be the normalization of the signal received from a telephony device where the input is normalized to maximize use of the dynamic range of the output signal. In this case, LL will be OdB and G will be limited by the maximum amplitude of the input signal, which may range from full-scale (limiting G to OdB) to a lower (normally at most 6dB, or half) value.
[0049] In each case, since the AGC will be operating near G(max) (to normalize speech to full scale), the AGC will be prevented from substantially over-amplifying smaller signals (such as noise) that are misrecognised as speech.
[0050] Referring now to FIG. 4, an example of an approach at AGC is described.
This approach describes various modules that can be used to implement the functions described with respect to FIGs. 1 and 2 above. It will be appreciated that these functions may be implemented by any combination of hardware and/or software (e.g., computer instructions stored on computer media). The approach of FIG. 4 utilizes a speech power computation module 402, a background noise estimator, a slow envelope tracking module 406, a speech pause detection module 408, a fast envelope tracking module 410, a peak power estimator 412, a sudden power change detection module 414, an AGC rate module 416, an AGC gain computational module 418, and a summer 420.
[0051] The speech power computation module 402 computes power of the speech.
[0053] The slow envelope tracking module 406 provides for the slow envelope tracking, first order IIR smoothing of the average power for the input speech signal. The slow envelope tracking, Ps (m) for the m'h frame is computed as given in the following equation:
[0055] The time variant constant, β5 (m) for the IIR filter is chosen differentially for rising and falling signal edges as given in the following equation:
[0057] A rising signal edge is followed faster than a falling one as:
[0058] 0 « β^ < β5 Ι < 1 (4)
[0059] The fast envelope tracking module 410 performs the fast envelope tracking,
PF ( ) for the m'h frame and is computed in the same manner as the slow envelope tracking as given in the following equation:
[0061] Where
\ F r ; > PF {m - i)
[0062] βΑ>") = \ 0 ' ° ^
\ pF f ; Otherwise
[0063] One difference between slow and fast envelope tracking filters, is that the different time varying constants are used which provide the following conditions:
[0065] The background noise estimator module 404 provides for estimation of the background noise level, minimum statistics and power smoothing only during speech pauses is used for the AGC algorithm. For estimation of the long term average power of the signal,
PN (m) for the m'h frame, first order IIR can be utilized to smooth the instantaneous power
of speech signals as given in the following:
[0066] PN (m) = βΝ (m)PN (m - 1) + (l - βΝ (m))P0 (8)
[0067] To detect rising speech signal powers very rapidly, different smoothing constants for rising and falling signal edges are used as given:
\βΝ τ ; P0 > PN (m - \)
[0068] fiN (m) =
\ pN' 0 N (9)
j Otherwise
[0069] Where 0 < βΝ r < fiNJ < 1 . For estimating the background noise level, PBNE (m) for the m'h frame, the minimum of the current average short term power and the output of the background noise level for the previous frame are compared. To avoid freezing at a global minimum, the result of the minimum operator is multiplied by a constant slightly larger than one as:
[0070] + ε) (10)
[0071] Where the constant, ε , is a small positive value that controls the maximum speed for increasing the estimated noise level. The quantity:
[0072] A(£) = /_ 101og10 (l + £) (1 1)
[0073] describes the maximum background noise power increase in decibels (dB) per second where fs is the sampling frequency. Depending on the type of background noise, ε should be chosen such that Α(ε) is within the interval 0.5 dB/second and 3 dB/second.
[0074] The speech pause detection module 408 compares the outputs of the fast and slow envelope tracking estimators and is the first stage of a simple speech pause detector. At the beginning of speech interval, the output of the fast envelope tracking estimator rises faster than that of the slow one. Due to short term power variations during speech activity on one hand and the choices of the time constants on the other hand, the output signal of the fast envelope tracking estimator is always larger than its slow counterpart. At the end of the voice activity period, the output of the slow envelope tracking estimator does not decrease as rapidly as does that of the fast envelope tracking estimator, and now the output of the slow envelope tracking estimator exceeds that of the fast one.
[0075] In order to avoid wrong decisions during speech pauses or voice activity, the
simple comparison is enhanced by bounding the output of the slow envelope tracking estimator to a constant, BN (m) . This constant should be chosen a few decibels above
( 101og10 («) ) the background noise level estimate as described above and given as follows:
[0076] BN (m) = aPBNE (m) (12)
[0077] The threshold for speech pause or speech activity detector, T(m) , is then computed as:
[0078] T(m) = max{ps (m), ^BN (m) } (13)
[0079] The speech pauses (SVAD (m) = θ) or speech activity (SVAD (m) = l) is detected as follows:
fl; PF (m) > T(m)
[0080] Sr D (m) = \ ' . (14)
[0; Otherwise
[0081] The peak power estimator module 412 estimates the peak power of speech signals. During voice intervals, the peak amplitude level is tracked. This is performed again by first order IIR filtering as:
[βΡ (m)PP (m - 1) + (l - βΡ (m))PF (m); SVAD (m) = 1
[0082] PP (m) = p P HP , F VAD (15)
Pp (m - l); Otherwise
[0083] The filter is updated only if there is speech activity; otherwise the filter keeps its old output values. The fast envelope tracking estimator is used as an input. In order to avoid an excessive rapid decreases PP (m) during wrong decision of speech pause detection, a very long time constant for falling signal edges was chosen as:
P t erwise
[0085] Where
[0086] 0 « βΝ„ < βΝ / < \ (17)
[0087] The sudden power change detection module 414 and AGC rate determination module 416 are used to determine the AGC rate that will be applied during AGC gain computation. Additionally, the sudden power change detection pre-determined variety of AGC variables (i.e., peak power estimate, slow envelope tracking, fast envelope tracking,
and background noise estimate) and AGC gain are used in order to more effectively eliminate the sudden power change affects in a short period of time.
[0088] The AGC gain computation module 418 performs signal scaling. During speech intervals the short term power of the input speech signal is estimated and compared to a desired value. If the speech level exceeds an upper bound, the gain factor, G(m) for the m'h frame is decreased - in the other case the gain is increased. In order to avoid fast gain variations during short speech pauses, G(m) is incremented or decremented in very small steps. The gain factors, G(m) is updated only during the periods of speech activity as:
[pG{m - X) SVAri (m) = \
[0089] G(m) = \ ^ vad (18)
[ G(m - 1); Otherwise
[0090] The scaling factor, p is computed by comparing the amplified or attenuated
(depending on G(m - 1) ) speech level estimator, Pp (m) with the desired level, DL . If the speech level after scaling with G(m - 1) is still small, the gain factor is increased as:
[0092] Where
Ginc = d + )
[0093] = J_ (20)
dec ~
inc
[0094] Where the constant, μ , is a small positive value that controls the maximum speed for increasing or decreasing the AGC gain in a second. The quantity:
[0095] Δ( ) = /.201ο8 ιο (1 + ) (21)
[0096] describes the maximum AGC Gain increase/decrease in decibels (dB) per second where fs is the sampling frequency, μ should be chosen such that Α(μ) is to provide the desired AGC gain increase/decrease in decibels (dB) per second. For example, if μ = 0.0001 , then the Α(μ) = 7 dB per second.
[0097] The summer 420 sums the gain from the module 418 and the input signal and this is made available at the output. More specifically, the output speech signal of AGC
algorithm for the m' frame, s0 (m, n) is computed by multiplying the input speech signal for the m'h frame, sm (n) with the AGC gain factor, G(m) as:
[0098] s0 (m, n) = G(m)sm (n) (22)
[0099] Referring now to FIG. 5, an example graph showing the change in G(max) in response to increasingly larger input signals is described. For each successively higher input sample, G(max) is reduced accordingly. With the upper bound on G shrinking in this fashion, it can be seen that the maximum possible over amplification error (which is often applied to smaller input signals) has been limited by the larger input signals. Thus, over-amplification is prevented and signal quality is improved.
[00100] It will be understood that the functions described herein may be implemented by computer instructions stored on a computer media (e.g., in a memory) and executed by a processing device (e.g., a microprocessor, controller, or the like).
[00101] It is understood that the implementation of other variations and modifications of the present invention and its various aspects will be apparent to those of ordinary skill in the art and that the present invention is not limited by the specific embodiments described. It is therefore contemplated to cover by the present invention any modifications, variations or equivalents that fall within the spirit and scope of the basic underlying principles disclosed and claimed herein.
Claims
1. A method of controlling automatic gain control for speech signals, the method comprising:
receiving a speech signal at an input;
tracking at least one electrical value associated with the received speech signal;
determining a dynamic adjustment of the speech signal, the dynamic adjustment selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis of the at least one electrical value, the dynamic adjustment further selected to obtain a desired output signal characteristic for the speech signal presented at an output;
applying the dynamic adjustment value to the speech signal;
presenting the adjusted speech signal at the output.
2. The method of claim 1 wherein the tracking comprises tracking a power level measured associated with an electrical envelope.
3. The method of claim 1 wherein the tracking comprises detecting a speech pause associated with an electrical envelope.
4. The method of claim 1 wherein the tracking comprises tracking at least one difference between a power level associated with the speech signal and a desired power level, and wherein the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level.
5. The method of claim 1 wherein the analysis comprises a computation of a maximum gain based upon a desired output divided the input and wherein applying the dynamic adjustment value comprises limiting the gain of the adjusted speech signal at the output to substantially the maximum computed gain.
6. The method of claim 1 further comprising transmitting the speech signal to an audio application in a vehicle.
7. A method of providing automatic speech gain control to prevent over- amplification of speech signals, the method comprising:
receiving a speech signal;
based upon at least one characteristic of the speech signal and at least one characteristic of the desired output signal, determining a maximum gain so as to limit over amplification of the received speech signal;
applying the determined maximum gain to the signal so as to limit over-amplification of the input signal.
8. The method of claim 7 wherein the maximum gain comprises a gain that would be effective to saturate the speech signal.
9. The method of claim 7 wherein the maximum gain comprises a maximum desired output divided by a maximum magnitude of the received speech signal.
10. An apparatus configured to control automatic gain control for speech signals, the apparatus comprising:
an interface with an input and an output, the input being configured to receive a speech signal;
a controller, the controller configured to track at least one electrical value associated with the received speech signal, the controller further configured to determine a dynamic adjustment of the speech signal, the dynamic adjustment selected at least in part so as to minimize a distortion and minimize an over-amplification of the speech signal based at least in part upon an analysis by the controller of the at least one electrical value, the dynamic adjustment further selected to obtain a desired output signal characteristic for the speech signal to be presented at the output, the controller configured to apply the dynamic adjustment value to the speech signal and present the adjusted speech signal at the output of the interface.
1 1. The apparatus of claim 10 wherein the controller is configured to track a power level measured associated with an electrical envelope.
12. The apparatus of claim 10 wherein the controller is configured to detect a speech pause associated with an electrical envelope.
13. The apparatus of claim 10 wherein the controller is configured to track at least one difference between a power level associated with the speech signal and a desired power level, and wherein the dynamic adjustment is effective to maintain the signal at the output at substantially the desired power level.
14. The apparatus of claim 10 wherein the analysis comprises a computation of a maximum gain based upon a desired output divided the input and wherein applying the dynamic adjustment value comprises limiting the gain of the adjusted speech signal at the output to substantially the maximum gain.
15. The apparatus of claim 10 wherein the output is coupled to an audio application in a vehicle.
16. The apparatus of claim 15 wherein the audio application is selected from the group consisting of a speaker and a microphone.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US13/189,191 US9537460B2 (en) | 2011-07-22 | 2011-07-22 | Apparatus and method for automatic gain control |
| US13/189,191 | 2011-07-22 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2013015887A1 true WO2013015887A1 (en) | 2013-01-31 |
Family
ID=46420524
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/US2012/041024 Ceased WO2013015887A1 (en) | 2011-07-22 | 2012-06-06 | Apparatus and method for automatic gain control |
Country Status (2)
| Country | Link |
|---|---|
| US (1) | US9537460B2 (en) |
| WO (1) | WO2013015887A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107533849A (en) * | 2015-04-23 | 2018-01-02 | 华为技术有限公司 | Audio signal processing device for processing input earphone audio signal based on microphone audio signal |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8862465B2 (en) * | 2010-09-17 | 2014-10-14 | Qualcomm Incorporated | Determining pitch cycle energy and scaling an excitation signal |
| US9070371B2 (en) * | 2012-10-22 | 2015-06-30 | Ittiam Systems (P) Ltd. | Method and system for peak limiting of speech signals for delay sensitive voice communication |
| GB2547459B (en) * | 2016-02-19 | 2019-01-09 | Imagination Tech Ltd | Dynamic gain controller |
| US9749733B1 (en) | 2016-04-07 | 2017-08-29 | Harman Intenational Industries, Incorporated | Approach for detecting alert signals in changing environments |
| US10540990B2 (en) | 2017-11-01 | 2020-01-21 | International Business Machines Corporation | Processing of speech signals |
| US11164592B1 (en) * | 2019-05-09 | 2021-11-02 | Amazon Technologies, Inc. | Responsive automatic gain control |
| CN112669878B (en) * | 2020-12-23 | 2024-04-19 | 北京声智科技有限公司 | Sound gain value calculation method and device and electronic equipment |
| CN120166333B (en) * | 2025-03-17 | 2025-10-21 | 东莞阳康电子有限公司 | A gradual change loudspeaker box and loudspeaker control method thereof |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6212273B1 (en) * | 1998-03-20 | 2001-04-03 | Crystal Semiconductor Corporation | Full-duplex speakerphone circuit including a control interface |
| WO2002097977A2 (en) * | 2001-05-30 | 2002-12-05 | Intel Corporation | Enhancing the intelligibility of received speech in a noisy environment |
| EP2048659A1 (en) * | 2007-10-08 | 2009-04-15 | Harman Becker Automotive Systems GmbH | Gain and spectral shape adjustment in audio signal processing |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| IT1044353B (en) * | 1975-07-03 | 1980-03-20 | Telettra Lab Telefon | METHOD AND DEVICE FOR RECOVERY KNOWLEDGE OF THE PRESENCE E. OR ABSENCE OF USEFUL SIGNAL SPOKEN WORD ON PHONE LINES PHONE CHANNELS |
| US5301364A (en) * | 1988-11-30 | 1994-04-05 | Motorola, Inc. | Method and apparatus for digital automatic gain control in a receiver |
| US5396576A (en) * | 1991-05-22 | 1995-03-07 | Nippon Telegraph And Telephone Corporation | Speech coding and decoding methods using adaptive and random code books |
| GB2355607B (en) * | 1999-10-20 | 2002-01-16 | Motorola Israel Ltd | Digital speech processing system |
| DE60029147T2 (en) * | 2000-12-29 | 2007-05-31 | Nokia Corp. | QUALITY IMPROVEMENT OF AUDIO SIGNAL IN A DIGITAL NETWORK |
| US7155385B2 (en) * | 2002-05-16 | 2006-12-26 | Comerica Bank, As Administrative Agent | Automatic gain control for adjusting gain during non-speech portions |
| US7542899B2 (en) * | 2003-09-30 | 2009-06-02 | Alcatel-Lucent Usa Inc. | Method and apparatus for adjusting the level of a speech signal in its encoded format |
| JP2006350246A (en) * | 2005-06-20 | 2006-12-28 | Nhk Engineering Services Inc | Speech / pause section detection device |
| US20090010453A1 (en) * | 2007-07-02 | 2009-01-08 | Motorola, Inc. | Intelligent gradient noise reduction system |
| JPWO2009051197A1 (en) * | 2007-10-19 | 2011-03-03 | 日本電気株式会社 | Echo suppression method and apparatus |
| US8641595B2 (en) * | 2008-01-21 | 2014-02-04 | Cochlear Limited | Automatic gain control for implanted microphone |
| US8416964B2 (en) * | 2008-12-15 | 2013-04-09 | Gentex Corporation | Vehicular automatic gain control (AGC) microphone system and method for post processing optimization of a microphone signal |
| US8254590B2 (en) * | 2009-04-29 | 2012-08-28 | Dolby Laboratories Licensing Corporation | System and method for intelligibility enhancement of audio information |
| JP4906908B2 (en) * | 2009-11-30 | 2012-03-28 | インターナショナル・ビジネス・マシーンズ・コーポレーション | Objective speech extraction method, objective speech extraction apparatus, and objective speech extraction program |
-
2011
- 2011-07-22 US US13/189,191 patent/US9537460B2/en active Active
-
2012
- 2012-06-06 WO PCT/US2012/041024 patent/WO2013015887A1/en not_active Ceased
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6212273B1 (en) * | 1998-03-20 | 2001-04-03 | Crystal Semiconductor Corporation | Full-duplex speakerphone circuit including a control interface |
| WO2002097977A2 (en) * | 2001-05-30 | 2002-12-05 | Intel Corporation | Enhancing the intelligibility of received speech in a noisy environment |
| EP2048659A1 (en) * | 2007-10-08 | 2009-04-15 | Harman Becker Automotive Systems GmbH | Gain and spectral shape adjustment in audio signal processing |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107533849A (en) * | 2015-04-23 | 2018-01-02 | 华为技术有限公司 | Audio signal processing device for processing input earphone audio signal based on microphone audio signal |
| CN107533849B (en) * | 2015-04-23 | 2021-06-29 | 华为技术有限公司 | Audio signal processing device for processing input earphone audio signal based on microphone audio signal |
Also Published As
| Publication number | Publication date |
|---|---|
| US9537460B2 (en) | 2017-01-03 |
| US20130024193A1 (en) | 2013-01-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9537460B2 (en) | Apparatus and method for automatic gain control | |
| US8150045B2 (en) | Automatic gain control system applied to an audio signal as a function of ambient noise | |
| EP2241099B1 (en) | Acoustic echo reduction | |
| JP5678445B2 (en) | Audio processing apparatus, audio processing method and program | |
| US9401685B2 (en) | Systems and methods for adjusting automatic gain control | |
| CN112272848B (en) | Background noise estimation using gap confidence | |
| CN1867965B (en) | Voice Activity Detection Using Adaptive Noise Floor Tracking | |
| US20080008311A1 (en) | Methods and apparatus for improving voice quality in an environment with noise | |
| US20040101038A1 (en) | Systems and methods for far-end noise reduction and near-end noise compensation in a mixed time-frequency domain compander to improve signal quality in communications systems | |
| WO2006036490A1 (en) | Method of cascading noise reduction algorithms to avoid speech distortion | |
| US9712908B2 (en) | Adaptive residual feedback suppression | |
| JP3961290B2 (en) | Noise suppressor | |
| US11664040B2 (en) | Apparatus and method for reducing noise in an audio signal | |
| US8477887B2 (en) | Systems and methods providing frequency-domain automatic gain control (AGC) | |
| US20080201137A1 (en) | Method of estimating noise levels in a communication system | |
| JP4321049B2 (en) | Automatic gain controller | |
| JP3929686B2 (en) | Voice switching apparatus and method | |
| JP2009147701A (en) | Amplitude control device, mobile phone device, and amplitude limiting method | |
| EP2869600B1 (en) | Adaptive residual feedback suppression | |
| JP2009147702A (en) | Noise level estimation device, received sound volume control device, mobile phone device, and noise level estimation method | |
| JP3588555B2 (en) | Method and apparatus for automatically adjusting sound level | |
| EP2760024B1 (en) | Noise estimation control | |
| JP5040623B2 (en) | Automatic volume correction device | |
| KR101394504B1 (en) | Apparatus and method for adaptive noise processing | |
| JP4732231B2 (en) | Mute circuit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12731223 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 12731223 Country of ref document: EP Kind code of ref document: A1 |