CN101221767A - Voice boosting device and method used on the same - Google Patents
Voice boosting device and method used on the same Download PDFInfo
- Publication number
- CN101221767A CN101221767A CNA2008100085184A CN200810008518A CN101221767A CN 101221767 A CN101221767 A CN 101221767A CN A2008100085184 A CNA2008100085184 A CN A2008100085184A CN 200810008518 A CN200810008518 A CN 200810008518A CN 101221767 A CN101221767 A CN 101221767A
- Authority
- CN
- China
- Prior art keywords
- voice
- signal
- frequency
- voice signal
- carried out
- 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.)
- Granted
Links
Images
Landscapes
- Stereophonic System (AREA)
- Telephone Function (AREA)
Abstract
The invention relates to a speech strengthening device and a method applied to the device, wherein the device comprises a frequency demultiplier, a speech strengthening calculator and a frequency-booster; the method comprises the following steps: a sound signal is processed by frequency reduction to form a frequency-reduction sound signal; the frequency-reduction sound signal is processed by speech strengthening calculation to form a speech strengthening sound signal; the speech strengthening sound signal is processed by frequency raising to form a frequency raising sound signal. The invention can uses idiomatic speech strengthening technology and relevant processing technology of signal mixing, low pass filtering, frequency reduction and frequency boost to make an obvious and clear strengthening effect on the speech frequency band of human voice in voice signals, and effectively improves the problems of energy consumption in arithmetic processing and resource consumption of a memorizer.
Description
Technical field
The present invention relates to a kind of voice boosting device and the method that is applied thereon, relate in particular to and a kind ofly utilize voice reinforcement technique and coherent signal treatment technology, and can make the apparatus and method of stiffening effect the voice voice in the voice signal.
Background technology
Use in the acoustic processing of general voice output interface or output effect, film for example, TV, computing machine, the micropkonic voice output of sound equipment, or mobile phone, phone, the loudspeaker voice output of microphone radio reception etc., the sound of its output has comprised the sound waveform that various frequency bands are arranged, and comprised that the dialogue of main contents voice is arranged, background sound, hum or other sound etc., and can change its effect acoustically in order to want for the output of some sound, or to strengthen the importance of some sound especially, need that just its voice signal is carried out further audio and handle.
Specifically, just on the sound effect that will export, strengthen we the content of the voice voice that require emphasis especially, for example: the dialogue frequency range of strengthening the leading role personage in the film, or the people audio frequency range of reinforcement in telephone speech output, with background sound or the hum that comparatively takes second place with respect to other importance in a certain environment or the signal, can have and comparatively significantly contrast and the output of the property emphasized performance, present and the purpose of hearing identification clearly to reach clear, this is technical very important technology of acoustic processing and subject under discussion.
From the above, these voice voice are strengthened or the technology of voice reinforcements (Speech Enhancement), have using and using of each side conventional techniques at present, as shown in Figure 1, be the waveform synoptic diagram that a conventional techniques is strengthened special frequency channel, wherein the oscillogram above this figure is original voice output waveform, and its transverse axis is expressed as the size of frequency, the longitudinal axis then is expressed as the power of waveform output, and the oscillogram below this figure then is treated waveform.Because the sound display frequency of general voice arrives between 6K or the 7K hertz (promptly 6000 arriving 7000Hz) about 500 hertz (Hz), so if the sound frequency that surpasses this scope frequency range of non-general voice voice then, thus shown in the figure, the general technology of strengthening the voice voice is directly strengthened output for the 1K that directly captures wherein to the frequency band signals of 3K hertz (Hz) in the frequency range of its voice output, or can be via the wave filter of a time domain (Time Domain) a certain special frequency channel of signal is strengthened its output with logical Filtering Processing, though so can reach the purpose that required voice voice band is partly strengthened, but the sound of non-main contents such as wherein existing background sounds or noise then also can be strengthened in the lump, thereby the effect that causes contrasting can't be especially obvious and clear.The numeral of part and simulated television can adopt this kind mode or similar processing mode to strengthen its voice output.
In addition, as shown in Figure 2, for another conventional techniques is carried out system's operation synoptic diagram that the voice voice are strengthened, wherein this technology is under frequency domain (Frequency Domain) voice signal of monophony input to be handled on handling, and the frequency sample that need change signal is than (Frequency Sample rate, abbreviate FS rate as) or so-called sampling frequency carry out digital processing, and general voice signal frequency sample commonly used ratio or sampling frequency include: 44.1K, 48K, 32K hertz etc., be with " fast fourier transform " (Fast Fourier Transform in the computing with signal, abbreviate FFT as) mode capture whole frequency spectrum, strengthening computing 10 via the voice shown in the figure again handles the back and can carry out various algorithms to each sampling value of specified resolution under frequency field, to subdue non-staple background sound frequency, the hum frequency, or strengthen processing procedures such as needed voice speech frequency, and through manage from here process getable result the voice voice band output that accounts for very big ratio can be arranged, and again via turnaround time territory, " anti-fast fourier transform " (Inverse FFT abbreviates IFFT as) back to carry out voice output.
And above-mentioned technology comprises that this voice strengthen computing 10 etc., and widespread and is main function application with the mobile phone of GSM form in the voice output of phone or mobile phone especially; At present known tupe or the disposal route of this technology includes: frequency spectrum subdues that (Spectral Subtraction) approaches, signal subspace (Signal Subspace) approaches, energy suppresses that signal subspace (Energy ConstrainedSignal Subspace) approaches, improved frequency spectrum subdues that (Modified Spectral Subtraction) approaches, linear prediction residual method tupe or methods such as (Linear Prediction Residual Method); And on such as general stereosonic voice output, the mode of two sound channel separate processes is finished the function that its voice are strengthened about most of the employing.
In the mode of above-mentioned Fig. 1, just can finish its voice reinforcement though can not need carry out time-consuming conversion process computing, shortcoming is that the processing of being done not is very obviously and gives prominence to, and can't effectively voice and other sound be done tangible difference reinforcement or filtering.And the employed technology of its Fig. 2, though then can effectively utilize the computing of fourier transform, thereby can capture people's acoustic frequency or background sound frequency at each sampling value of specified resolution under frequency field, and carry out corresponding required reinforcement or filtering, yet, during processing about this technology is used in respectively on two sound channels, on the processing procedure of computing, can comparatively consume its system storage () resource for example: DRAM or SRAM for system, and from FFT with after under frequency field, providing this voice to strengthen computing 10 to deal with, could under time domain, export its result after need remaking IFFT, and this kind also can consume the resource of system storage by the calculating process that FFT remakes IFFT very much, and can take a large amount of calculation resources of processor and usefulness.
Summary of the invention
Therefore, how to solve the problem of this conventional techniques, just become fundamental purpose of the present invention.
The object of the present invention is to provide a kind of voice boosting device and the method that is applied thereon, can utilize the treatment technology of habitual voice reinforcement technique and relevant signal mixing, low-pass filtering, frequency reducing and raising frequency, and can make obviously and clearly stiffening effect, and can improve the problem that power consumption on the calculation process and memory resource consume effectively to the voice voice band in the voice signal.
The present invention adds strong method for a kind of voice, is applied on the voice boosting device, and this method comprises the following step: receive a voice signal, the sampling frequency of this voice signal is a first frequency; This voice signal is carried out a down conversion process, and then form a frequency reducing voice signal, the sampling frequency of this frequency reducing voice signal is a second frequency, and this second frequency is lower than this first frequency; This frequency reducing voice signal is carried out voice strengthen computing, and then form voice reinforcement voice signal; And voice signal strengthened in these voice carry out raising frequency processing, and then forming a raising frequency voice signal, the sampling frequency of this raising frequency voice signal is this first frequency.
Another aspect of the present invention is a kind of voice and add strong method, be applied on the voice boosting device, this method comprises the following step: a L channel voice signal and a right channel sound signal are carried out one first signal hybrid processing, and then form a voice signal; This voice signal is carried out voice strengthen computing, and then form voice reinforcement voice signal; And after this voice are strengthened voice signal and carry out a secondary signal hybrid processing and one the 3rd signal hybrid processing with this L channel voice signal and this right channel sound signal respectively, carry out signal output.
Another aspect of the present invention is a kind of voice boosting device, this device includes: a frequency demultiplier, in order to being that a voice signal of a first frequency carries out a down conversion process to sampling frequency, and then form a frequency reducing voice signal, the sampling frequency of this frequency reducing voice signal is a second frequency, and this second frequency is lower than this first frequency; Arithmetical unit strengthened in one voice, and signal is connected in this frequency demultiplier, strengthens computing in order to this frequency reducing voice signal is carried out voice, and then form voice reinforcement voice signal; And a frequency-booster, signal is connected in this voice and strengthens arithmetical unit, carries out a raising frequency and handles in order to these voice are strengthened voice signal, and then form a raising frequency voice signal, and the sampling frequency of this raising frequency voice signal is this first frequency.
Another aspect of the present invention is a kind of voice boosting device, this device includes: one first mixer in order to a L channel voice signal and a right channel sound signal are carried out one first signal hybrid processing, and then forms a voice signal; Arithmetical unit strengthened in one voice, strengthens computing in order to this voice signal is carried out voice, and then form voice reinforcement voice signal; And one second mixer and one the 3rd mixer, in order to after these voice being strengthened voice signal and being carried out a secondary signal hybrid processing and one the 3rd signal hybrid processing with this L channel voice signal and this right channel sound signal respectively, carry out signal output.
Description of drawings
Fig. 1 is the synoptic diagram that a conventional techniques is strengthened special frequency channel.
Fig. 2 carries out system's operation synoptic diagram that the voice voice are strengthened for another conventional techniques.
Fig. 3 is the synoptic diagram that can move out of the multimedia playing apparatus of various audio processing capacities.
Fig. 4 is the synoptic diagram of voice boosting device 30 of the present invention.
Fig. 5 is the process flow diagram of first preferred embodiment of the invention.
Fig. 6 is the synoptic diagram of half frequency band filter of FIR form.
The operation synoptic diagram that covers sampling value and filtering HFS that Fig. 7 (a) handles for raising frequency to Fig. 7 (c).
Fig. 8 is the process flow diagram of second preferred embodiment of the invention.
Fig. 9 is the synoptic diagram of the folded biquadratic filter of string of an IIR form.
Wherein, description of reference numerals is as follows:
Computing 20 audio digital signal processors strengthened in 10 voice
201～204 sound figures are handled audio channel one～four
211～215 signals are imported 200 multiplexers
23 system storages, 30 voice boosting devices
301 first mixers, 302 second mixers
303 the 3rd mixers, 311 first delayers
312 second delayers, 32 first low-pass filters
320～3222 delayers, 3200 totalizers
Arithmetical unit strengthened in 33 frequency demultipliers, 34 voice
35 frequency-boosters, 36 second low-pass filters
37 multiplier (-icator) V1 voice signals
V2 filtering high-frequency sound signal V3 frequency reducing voice signal
Voice signal V5 raising frequency voice signal strengthened in the V4 voice
The V6 voice are strengthened and filtering high-frequency sound signal
F1, f2, f3 curve
S0～S5, S0 '～S4 ', S0 "～S4 " sampling value
Embodiment
The present invention is by following accompanying drawing and explanation, to obtain a more deep understanding:
As described in the prior art, the existing technology of strengthening at the frequency range of voice voice in conventional techniques, and be applied on the relevant apparatus or equipment with sound-playing function, TV for example, computing machine, mobile phone etc., and the objective of the invention is to improve the calculating process of strengthening function for voice in the conventional techniques the problem that consumes of the power consumption processing that can cause and system storage, the present invention still continues to utilize habitual voice to strengthen voice existing in (Speech Enhancement) technology and strengthens calculation function in addition, just via using voice to strengthen computing module or voice reinforcement arithmetical unit, the function of utilizing the computing of fourier transform and under the channel field, strengthening or subdue to specific frequency range, its purpose is energy and other background sound except being the voice voice to be strengthened, hum has outside the obviously and clearly contrast, can also improve a large amount of consumption processor resources and the problems such as usefulness and system memory resources consumption of conventional techniques effectively.
See also Fig. 3, be one can move out of the synoptic diagram of the multimedia playing apparatus of various audio processing capacities, this multimedia playing apparatus can be a digital television, and the user can set on the menu and carry out relevant setting control of harmony sound display effect or preference adjustment via relevant user's interface or in a screen display (On Screen Display (being called for short OSD)).This device mainly is to utilize an audio digital signal processor 20, come multiple voice signal is carried out digital processing, its voice signal input 20 treatable types of visual this processor or form and the signal input of varying number is arranged, voice signal input 211～215 as icon can include: by the signal input of voice decoder (Audio Decoder), new power/philips digital interface (SONY/PHILIPS Digital Interface, abbreviation SPDIF) the signal input of form, high resolution multimedia interface (High Definition Multimedia Interface, abbreviation HDMI) the signal input of form, the signal input of chip chamber sound (Inter-IC Sound is called for short I2S) form, the signal input of analog digital conversion (Analog Digital Change) form etc.One system storage 23 then can provide the memory resource on the calculation process.
These signals can be digital format signal, or are to import behind the digital format by analog-converted, and transport to wherein multiple sound figure by a multiplexer 200 and handle in the audio channel one～4 201～204 and handle and export.Wherein each audio channel can include according to the difference of processing capacity: volume control (VolumeControl), bass adjustment (Bass Adjustment), high pitch adjustment (Treble Adjustment), around (Surround), lamprophonia (Superior Voice) etc., and user control or adjust this setting menu after just can start the corresponding sound effect processing capacity, in like manner, the number of this audio channel is by deciding according to these processor 20 treatable functions.
Voice of the present invention add strong method and just can be applicable on the above-mentioned multimedia playing apparatus, furthermore, method of the present invention and application are that above-mentioned multiple sound figure is handled the audio channel relevant channel of this lamprophonia (Superior Voice) function that neutralizes, just can carry out the channel that its voice are strengthened function and handled, improved calculation process mode wherein, thereby can make the user after the voice that the present invention is designed add the pairing audio channel startup of strong method, just can obtain obviously and clearly output effect of voice.
See also Fig. 4, be the preferred embodiment synoptic diagram of a voice boosting device 30 of the present invention.From the above, this voice boosting device 30 can be applicable in these audio channels to add in one of them relevant channel of powerful and the corresponding input structure with voice, and the voice signal after this voice boosting device 30 of the present invention is handled also can be exported by this structure shown in Figure 3.As shown in Figure 4, this voice boosting device 30 mainly has been provided with: an arithmetical unit 34 and a frequency-booster 35 strengthened in three mixers 301～303, two delayers 311～312, two low-pass filter (Low Pass Filter) 32 and 36, one frequency demultipliers 33, voice, and also shown each unit signal annexation to each other in this figure.
And at first, we utilize this first mixer 301 to carry out one first signal hybrid processing and form a voice signal V1 L channel voice signal of input in this voice boosting device 30 and a right channel sound signal (can be in these signals inputs 211～215 simultaneously with about the signal input that transmits of two sound channels), and this voice signal V1 just will carry out the calculation process object of voice reinforcement for the present invention.
In this, processing about being applied in respectively compared to the voice signal with monophony input of prior art on two sound channels, the present invention can with on the calculating process may consume this system storage 23 (can be DRAM or SRAM) a resource shrinkage half, this is because if when this L channel and right channel sound signal carried out calculation process individually, this system storage 23 just needs respectively to provide the memory space of a part to carry out computing to this two signal respectively, and this processor 20 has also comprised L channel and right channel sound Signal Processing process in the required operation efficiency of computing, but the present invention only needs that single this voice signal V1 is carried out calculation process to get final product, and this first signal hybrid processing can be directly will become this voice signal V1 again divided by 2 behind this L channel and the right channel sound signal plus, so still has complete signal content after its mixing.So except required operation efficiency of the consumption of the memory resource on the calculating process or 20 computings of this processor etc., just all only be half of conventional techniques, thereby can effectively solve the problem of conventional techniques.
In addition, we will will make the signal down of voice intensive treatment, specifically under the condition that does not influence its voice output effect, reduce the sampling frequency of pending signal, said method is not only at the most of energy that still comprises the voice voice after the frequency reducing on frequency range, can not influence the quality of voice voice, can also reduce its operand further, and can significantly improve the problem of storer and the consumption of processor calculating usefulness, its embodiment is described as follows.
Consult Fig. 5 simultaneously, be the process flow diagram of first preferred embodiment of the invention, step S11 wherein just is the process of above-mentioned this first signal hybrid processing.And this L channel and right channel sound signal are when importing, its frequency sample is a first frequency than (FS rate) or so-called sampling frequency, as described in the prior art, the frequency sample ratio of strengthening at voice can be 44.1K, 48K, 32K hertz (Hz) etc., and this voice signal V1 that is produced also has this identical first frequency; And in this embodiment, we design this first frequency that this left and right sound channels voice signal and this voice signal V1 have, for have the sampling frequency of n sampling value in a unit interval.
Yet step S12 is a down conversion process flow process of the present invention, and we carry out low-pass filtering treatment to this voice signal V1 earlier, remake the processing of frequency reducing.In this example, we utilize this first low-pass filter 32 that this voice signal V1 is carried out one first low-pass filtering treatment, and form a filtering high-frequency sound signal V2, and only with the HFS filtering of this voice signal V1 and do not change its sampling frequency, therefore, this filtering high-frequency sound signal V2 still has n sampling value in the unit interval.
Then, this filtering high-frequency sound signal V2 is carried out down conversion process,, be reduced to n/2 sampling value, and form a frequency reducing voice signal V3 the n of original unit in a time sampling value by this frequency demultiplier 33; For instance, in this first preferred embodiment, our design with sampling frequency to be processed reduce to half of former sampling frequency, and this first low-pass filter 32 just can be selected half frequency band filter (Half-Band Filter) for use, and can carry out sampling frequency at this frequency demultiplier 33 processing procedure of half is fallen, in order to prevent high-frequency signal influence (alias) down conversion process.In Fig. 6, demonstrate this first low-pass filter 32 and be the synoptic diagram of half frequency band filter, as shown in the figure, this wave filter includes 23 delayers 320～3222 and a totalizer 3200, because these delayers have the design factor of half (one the coefficient of promptly being separated by that is 0, only Zhong Yang coefficient is not 0), so can effectively reduce operand, and the result of the product of 23 delayers and its coefficient and addition just is the result of its low-pass filtering.
The process flow diagram that holds, we use this frequency demultiplier 33 that sampling frequency can be fallen half that this filtering high-frequency sound signal V2 is carried out a down conversion process and forms this frequency reducing voice signal V3 in step S12, the sampling frequency of this frequency reducing voice signal V3 is a second frequency, this second frequency that we design after the frequency reducing is m/one of original this first frequency, and be that m is taken as 2 in this embodiment, just fall half, thereby made formed this frequency reducing voice signal V3 in this unit interval, have n/2 sampling value.
In this embodiment, this first frequency that we use is the 48K hertz, so the second frequency after the frequency reducing just is the 24K hertz, this down conversion process also will be subdued (m-1) individual sampling value in every m sampling value in n sampling value of script simultaneously, for instance, we are taken as 2 with m, are to subdue 1 sampling value in per 2 sampling values, if hypothesis n originally is 1024, then new sampling value has m/n sampling value just to reduce in this unit interval becomes 512 sampling values.Therefore, the sampling value number of being got when doing the fourier transform computing that voice strengthen is the same with its sampling frequency has also done the disposal that reduces by half, so the resolution of its frequency domain (Frequency Resolution) (frequency that is correspondence is divided by its sampling value number) is still identical; Hereat, still can possess and the performance of signal same frequency resolution originally via the processing of frequency reducing and the reduction of sampling value number.
Then, be to utilize these voice to strengthen arithmetical unit 34 this frequency reducing voice signal V3 is carried out voice reinforcement computing and form voice strengthening voice signal V4 in step S13.And in this embodiment, these voice are strengthened these voice that arithmetical unit 34 carried out and are strengthened the technology of computing for habitually practising at present, for example: computing strengthened in these voice adopt a kind of frequency spectrum of digital signal processing to subdue the voice reinforcement computing that (Spectral Subtraction) approaches, this frequency reducing voice signal V3 that is imported is dealt with; Because the down conversion process of previous step, we can be effectively strengthen operand that arithmetical unit 34 will carry out with these voice and the resource space that will use for this system storage 23 etc., can reach and reduce to the situation of half originally, thereby can improve problems such as storer and the consumption of processor calculating usefulness.
In addition, the processing that computing strengthened in these voice does not change the frequency of this frequency reducing voice signal V3, has this identical second frequency so voice signal V4 strengthened in these voice of being exported with this frequency reducing voice signal V3.Further add in the left and right acoustic channels voice signal that comprises voice and background sound originally with output correctly and strengthen voice signal V4, make processing procedures such as the raising frequency of correspondence and low-pass filtering and then in step S14, these voice are strengthened voice signal V4 for these voice that will be handled well.Therefore then just utilizing 35 pairs of these voice of this frequency-booster to strengthen voice signal V4 earlier carries out raising frequency processing and forms a raising frequency voice signal V5, and in this embodiment because we have done the processing of frequency halving earlier before, therefore this raising frequency of relative this moment is handled just and is upgraded to twice for the frequency with its signal, make the sampling frequency of this raising frequency voice signal V5 become this original first frequency, make this raising frequency voice signal V5 in this unit interval, can have an original n sampling value simultaneously.
In this embodiment, the second frequency (24K hertz) that we strengthen voice signal V4 with these voice rises the first frequency (48K hertz) that twice (because of m is taken as 2) becomes this raising frequency voice signal V5, this raising frequency processing simultaneously is that zero sampling value becomes an original n sampling value with covering (m-1) individual numerical value between per two sampling values also, 512 sampling values after promptly will reducing in this example cover 1 sampling value and become 1024 original sampling values between per two sampling values, and utilize this can finish its raising frequency process once the practice that covers the sampling value number.
Then, be to utilize this second low-pass filter 36 that this raising frequency voice signal V5 is carried out one second low-pass filtering treatment again and form voice and strengthen and filtering high-frequency sound signal V6, wherein second low-pass filter 36 of this in this example can equally with this first low-pass filter 32 adopt identical this half frequency band filter (Half-Band Filter), and formed these voice strengthen just having an original n sampling value with filtering high-frequency sound signal V6, be 1024 sampling values among this embodiment, as step S14.
And represented that at Fig. 7 (a) the practice that above-mentioned utilization covers sampling value number and filtering high frequency finishes this raising frequency and handle and this second low-pass filtering treatment to the synoptic diagram of (c), it is the frequency of a low frequency that a curve f1 wherein can show, and a curve f2 is the frequency of a high frequency, 6 sampling value S0～S5 are arranged on this curve f1, when we will rise to the frequency of high frequency with the frequency of low frequency, cover its value between can per two sampling values and be sampling value S0 '～S4 ' of 0 to this curve f1, and form this curve f2, shown in Fig. 7 (a), then just can obtain to augment sampling value number S0 via these second low-pass filter, 36 computings "～S4 ", as Fig. 7 (b), at last, augment sampling value S0 in conjunction with these sampling values S0～S5 and this "～S4 ", can return back to a curve f3 of original samples frequency (being first frequency), as Fig. 7 (c).
And we have also designed a multiplier (-icator) 37 in the step S15 of this embodiment, in order to this voice being strengthened carrying out with filtering high-frequency sound signal V6 the processing of signal gain control, and these voice can be strengthened being adjusted with filtering high-frequency sound signal V6.For example, we can utilize this multiplier (-icator) 37 again its signal to be adjusted amplification or adjustment is dwindled, and wherein 37 of this multiplier (-icator)s signal that can produce strengthens being enlarged into a kind of positive signal gain value (gain), just the volume that shown of the voice phonological component that we will be able to be added is back controlled the ratio of its amplification again, and the effect that can make the voice voice strengthen is more obvious.
And the last step that the signal of handling is added back signal, owing to strengthen to cause phase delay (Group Delay) in the calculating process in above-mentioned filtering and voice, therefore we can use this first delayer 311 and second delayer 312 respectively this original L channel and right channel sound signal to be carried out processing of one first signal delay and secondary signal delay processing, and in this embodiment, these signal delays are treated to and postpone this L channel and right channel sound signal to be exported after identical time again, and use this second mixer 302 and the 3rd mixer 303 with adjusted these voice of signal strengthen with filtering high-frequency sound signal V6 respectively with postpone after this L channel voice signal and after this right channel sound signal carries out a secondary signal hybrid processing and one the 3rd signal hybrid processing, after just directly above-mentioned frequency range of carrying out the reinforcement of perfect person's sound voice being distinguished among this L channel of add-back and the right channel sound signal, just required audio result can be carried out signal output and reach described purpose, as step S15.
In sum, we except can be earlier with about two sound channels mix and only single voice signal handled to reduce a large amount of calculation resources of its processor to improve usefulness and to reduce its memory resource demand, also can further make down conversion process again and can reduce operand again, thereby can more strengthen above-mentioned usefulness, simultaneously, such processing mode can't influence the background sound after last voice voice are strengthened, and can be normally in voice output originally, strengthen the energy of voice voice effectively, thereby can successfully solve and improve the mentioned problem of prior art.
In addition, be to handle with the raising frequency that the down conversion process of frequency halving and corresponding frequency increase twice to illustrate in first preferred embodiment of the present invention, yet, we can also be further with frequency be kept to three/once (raising frequency of follow-up correspondence is handled to increasing three times) or frequency be kept to four/once the processing of (raising frequency of follow-up correspondence is handled to increasing four times), reducing more processor calculating amount and memory resource consumes, that is to say that we can be taken as this m value among the present invention the positive integer (m and n are positive integer in notion of the present invention) greater than 1, for example: 2,3,4 etc., carry out calculation process in various degree, yet be noted that if this m value get big more the time, then the high-frequency band of required filtering is also just big more, and may influence the voice voice band; Therefore, the m value is taken as 4 the more possible actual operation condition that is designed at most.
And in second preferred embodiment of the present invention, we just adopt and the frequency that will make signal Processing are reduced to original 1/3rd, and corresponding raising frequency handles and then increase three times and illustrate, and its process flow diagram as shown in Figure 8; Step S11, S13, the S15 of step S21, the S23 in this second preferred embodiment, S25 and first preferred embodiment are identical, the difference of second preferred embodiment and first preferred embodiment only is among the step S22 down conversion process is carried out to be kept to 1/3rd mode, and in step S24 raising frequency is handled accordingly in the mode that increases three times and carry out.
In addition, employed low-pass filter also need be adjusted; Using the folded biquadratic filter (IIR Cascade Bi-Quad Filter) of a kind of string by the IIR form in this second preferred embodiment is a main sampling filter (Decimation Filter) that is constituted or an interpolation filter (Interpolation Filter) and can show preferable effect; Shown in Figure 9 is the synoptic diagram of this kind wave filter, and partly just folds the structure (employed coefficient when wherein coefficient a0～a2, b1～b2 are computing) of biquadratic filter as shown in phantom in FIG. for the string of main IIR form; We so just can reach this kind wave filter use these low- pass filters 32,36 in above-mentioned Fig. 4 the frequency reducing specified in this second preferred embodiment and the disposal of raising frequency effectively.
Therefore, in sum, utilize the voice reinforcement computing of conventional techniques to strengthen to the voice phonological component in the voice signal of related sound output interface, and construct and processing mode by the signal Processing that signal mixing of the present invention, filtering and frequency reducing are formed, can further reduce the consumption of the operand and the system storage of processor, increase performance of entire system effectively, and can improve and the problem that solves conventional techniques, thereby can successfully reach fundamental purpose of the present invention.
Anyly be familiar with those skilled in the art, can be under the prerequisite of utilization and the identical purpose of the present invention, the basis of using disclosed notion and embodiment variation to be used as designing and improving some other method.These change, substitute and improve and can not deviate from protection scope of the present invention that claim limits.Therefore, the present invention is appointed by those of ordinary skills and executes that the craftsman thinks and be to modify as all, yet does not break away from the scope of the claim institute desire protection of enclosing.
Claims (12)
1. voice add strong method, are applied on the voice boosting device, and this method comprises the following step:
Receive a voice signal, the sampling frequency of this voice signal is a first frequency;
This voice signal is carried out a down conversion process, and then form a frequency reducing voice signal, the sampling frequency of this frequency reducing voice signal is a second frequency, and this second frequency is lower than this first frequency;
This frequency reducing voice signal is carried out voice strengthen computing, and then form voice reinforcement voice signal; And
Voice signal strengthened in these voice carry out raising frequency processing, and then form a raising frequency voice signal, the sampling frequency of this raising frequency voice signal is this first frequency.
2. voice as claimed in claim 1 add strong method, it is characterized in that this method also comprises the following step:
An one L channel voice signal and a right channel sound signal are carried out one first signal hybrid processing, and then form this voice signal; And
After this raising frequency voice signal carried out a secondary signal hybrid processing and one the 3rd signal hybrid processing with this L channel voice signal and this right channel sound signal respectively, carry out signal output.
3. voice as claimed in claim 2 add strong method, it is characterized in that this method also comprises the following step: with this raising frequency voice signal respectively with this L channel voice signal and this right channel sound signal carry out this secondary signal hybrid processing and the 3rd signal hybrid processing with the step of carrying out signal output in, this L channel voice signal and this right channel sound signal carried out respectively earlier one first signal delay is handled and secondary signal delay processing.
4. voice as claimed in claim 2 add strong method, it is characterized in that this method also comprises the following step: with this raising frequency voice signal respectively with this L channel voice signal and this right channel sound signal carry out this secondary signal hybrid processing and the 3rd signal hybrid processing with the step of carrying out signal output in, this raising frequency voice signal is carried out signal gain control earlier.
5. voice as claimed in claim 1 add strong method, it is characterized in that this method also comprises the following step:
Before this down conversion process, this voice signal is carried out one first low-pass filtering treatment, and then form a filtering high-frequency sound signal; And
After this raising frequency is handled, this raising frequency voice signal is carried out one second low-pass filtering treatment, and then form voice reinforcement and filtering high-frequency sound signal.
6. voice add strong method, are applied on the voice boosting device, and this method comprises the following step:
An one L channel voice signal and a right channel sound signal are carried out one first signal hybrid processing, and then form a voice signal;
This voice signal is carried out voice strengthen computing, and then form voice reinforcement voice signal; And
After this voice are strengthened voice signal and carry out a secondary signal hybrid processing and one the 3rd signal hybrid processing with this L channel voice signal and this right channel sound signal respectively, carry out signal output.
7. voice boosting device, this device includes:
One frequency demultiplier in order to being that a voice signal of a first frequency carries out a down conversion process to sampling frequency, and then forms a frequency reducing voice signal, and the sampling frequency of this frequency reducing voice signal is a second frequency, and this second frequency is lower than this first frequency;
Arithmetical unit strengthened in one voice, and signal is connected in this frequency demultiplier, strengthens computing in order to this frequency reducing voice signal is carried out voice, and then form voice reinforcement voice signal; And
One frequency-booster, signal are connected in these voice and strengthen arithmetical unit, carry out raising frequency processing in order to these voice are strengthened voice signal, and then form a raising frequency voice signal, and the sampling frequency of this raising frequency voice signal is this first frequency.
8. voice boosting device as claimed in claim 7 is characterized in that this device also comprises:
One first mixer in order to a L channel voice signal and a right channel sound signal are carried out one first signal hybrid processing, and then forms this voice signal; And
One second mixer and one the 3rd mixer after this raising frequency voice signal is carried out a secondary signal hybrid processing and one the 3rd signal hybrid processing with this L channel voice signal and this right channel sound signal respectively, carry out signal output.
9. voice boosting device as claimed in claim 8, it is characterized in that this device also includes one first delayer and one second delayer, after respectively this L channel voice signal and this right channel sound signal being carried out processing of one first signal delay and secondary signal delay processing, input to this second mixer and the 3rd mixer.
10. voice boosting device as claimed in claim 8 is characterized in that this device also includes a multiplier (-icator), controls in order to this raising frequency voice signal is carried out a signal gain, and then inputs to this second mixer and the 3rd mixer.
11. voice boosting device as claimed in claim 7 is characterized in that this device also comprises:
One first low-pass filter in order to before this down conversion process, carries out one first low-pass filtering treatment to this voice signal, and then forms a filtering high-frequency sound signal; And
One second low-pass filter in order to after this raising frequency is handled, carries out one second low-pass filtering treatment to this raising frequency voice signal, and then forms voice reinforcement and filtering high-frequency sound signal.
12. a voice boosting device, this device includes:
One first mixer in order to a L channel voice signal and a right channel sound signal are carried out one first signal hybrid processing, and then forms a voice signal;
Arithmetical unit strengthened in one voice, strengthens computing in order to this voice signal is carried out voice, and then form voice reinforcement voice signal; And
One second mixer and one the 3rd mixer after this voice reinforcement voice signal is carried out a secondary signal hybrid processing and one the 3rd signal hybrid processing with this L channel voice signal and this right channel sound signal respectively, carry out signal output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008100085184A CN101221767B (en) | 2008-01-23 | 2008-01-23 | Voice boosting device and method used on the same |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008100085184A CN101221767B (en) | 2008-01-23 | 2008-01-23 | Voice boosting device and method used on the same |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101221767A true CN101221767A (en) | 2008-07-16 |
| CN101221767B CN101221767B (en) | 2012-05-30 |
Family
ID=39631550
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008100085184A Active CN101221767B (en) | 2008-01-23 | 2008-01-23 | Voice boosting device and method used on the same |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101221767B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103489451A (en) * | 2012-06-13 | 2014-01-01 | 百度在线网络技术（北京）有限公司 | Voice processing method of mobile terminal and mobile terminal |
| CN104464746A (en) * | 2013-09-12 | 2015-03-25 | 索尼公司 | Voice filtering method and device and electron equipment |
| CN104978955A (en) * | 2014-04-14 | 2015-10-14 | 美的集团股份有限公司 | Voice control method and system |
| CN105049932A (en) * | 2014-01-20 | 2015-11-11 | 联想（新加坡）私人有限公司 | Detection noise or object interruption in audio video viewing and altering presentation based thereon |
| CN113782043A (en) * | 2021-09-06 | 2021-12-10 | 北京捷通华声科技股份有限公司 | Voice acquisition method and device, electronic equipment and computer readable storage medium |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2163032C2 (en) * | 1995-09-14 | 2001-02-10 | Эрикссон Инк. | System for adaptive filtration of audiosignals for improvement of speech articulation through noise |
| JP2001526482A (en) * | 1997-12-12 | 2001-12-18 | トゥルー・ディメンショナル・サウンド・インコーポレーテッド | Apparatus and method for enhancing electronic audio signals |
| JP4089020B2 (en) * | 1998-07-09 | 2008-05-21 | ソニー株式会社 | Audio signal processing device |
| CN100433938C (en) * | 2002-08-22 | 2008-11-12 | 联发科技股份有限公司 | Sound effect treatment method for microphone and its device |
| KR100739762B1 (en) * | 2005-09-26 | 2007-07-13 | 삼성전자주식회사 | Apparatus and method for cancelling a crosstalk and virtual sound system thereof |
| KR100636248B1 (en) * | 2005-09-26 | 2006-10-19 | 삼성전자주식회사 | Apparatus and method for cancelling vocal |
-
2008
- 2008-01-23 CN CN2008100085184A patent/CN101221767B/en active Active
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103489451A (en) * | 2012-06-13 | 2014-01-01 | 百度在线网络技术（北京）有限公司 | Voice processing method of mobile terminal and mobile terminal |
| CN103489451B (en) * | 2012-06-13 | 2016-11-23 | 百度在线网络技术（北京）有限公司 | The method of speech processing of mobile terminal and mobile terminal |
| CN104464746A (en) * | 2013-09-12 | 2015-03-25 | 索尼公司 | Voice filtering method and device and electron equipment |
| CN105049932A (en) * | 2014-01-20 | 2015-11-11 | 联想（新加坡）私人有限公司 | Detection noise or object interruption in audio video viewing and altering presentation based thereon |
| US10073671B2 (en) | 2014-01-20 | 2018-09-11 | Lenovo (Singapore) Pte. Ltd. | Detecting noise or object interruption in audio video viewing and altering presentation based thereon |
| CN105049932B (en) * | 2014-01-20 | 2019-09-24 | 联想（新加坡）私人有限公司 | It is a kind of to change equipment, the method and apparatus that audio-video is presented |
| CN104978955A (en) * | 2014-04-14 | 2015-10-14 | 美的集团股份有限公司 | Voice control method and system |
| CN113782043A (en) * | 2021-09-06 | 2021-12-10 | 北京捷通华声科技股份有限公司 | Voice acquisition method and device, electronic equipment and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101221767B (en) | 2012-05-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2517315C2 (en) | Method and device for audio signal processing | |
| CN102074246B (en) | Dual-microphone based speech enhancement device and method | |
| CN106658284B (en) | Addition of virtual bass in the frequency domain | |
| CN101505443A (en) | Virtual supper bass enhancing method and system | |
| US20130044896A1 (en) | Virtual Bass Synthesis Using Harmonic Transposition | |
| CN101577848B (en) | Supper bass boosting method and system | |
| CN101221767B (en) | Voice boosting device and method used on the same | |
| CN111739544B (en) | Voice processing method, device, electronic equipment and storage medium | |
| US8498429B2 (en) | Acoustic correction apparatus, audio output apparatus, and acoustic correction method | |
| CN112259116B (en) | Noise reduction method and device for audio data, electronic equipment and storage medium | |
| EP2476115A1 (en) | Method and apparatus for processing audio signals | |
| US9633667B2 (en) | Adaptive audio signal filtering | |
| CN113299313A (en) | Audio processing method and device and electronic equipment | |
| TWI351683B (en) | Speech enhancement device and method for the same | |
| EP2720477B1 (en) | Virtual bass synthesis using harmonic transposition | |
| CN109074813B (en) | Processing high definition audio data | |
| US20230051509A1 (en) | Method and system for implementing a modal processor | |
| US9075697B2 (en) | Parallel digital filtering of an audio channel | |
| CN101930736B (en) | Audio frequency equalizing method of decoder based on sub-band filter frame | |
| US20130177161A1 (en) | Sine wave generating device, digital signal processor and audio output device | |
| CN117409801A (en) | Audio processing method and device, storage medium and electronic equipment | |
| CN109167587A (en) | Signal bandpass filtering treatment method and system based on FPGA | |
| JP2019169840A (en) | Digital audio signal processing device and program thereof | |
| JP2001358562A (en) | Signal processing device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211027 Address after: China Taiwan Hsinchu Science Park Hsinchu city Dusing Road No. 1 Patentee after: MEDIATEK Inc. Address before: Hsinchu County, Taiwan, China Patentee before: MSTAR SEMICONDUCTOR Inc. |
|
| TR01 | Transfer of patent right |