CN1790901A - Automatic gain control method - Google Patents

Abstract

The disclosed AGC method to voice flow from different communication sending end comprises: a. determining the current treatment block in voice flow; according to current volume of receive end, calculating gain value of current block as the value opposite to the last sample in current block, and using the gain value of last sample point on former treatment block as the first gain value of current block; b. according to the gain values of first and last sample points in current block, calculating other gain values; c. multiplying the amplitude value of all sample points in current block with opposite gain value to output; then ,deciding whether there is other block achieved, if yes, returning to step a; or else, ending the process.

Description

A kind of auto gain control method

Technical field

The present invention relates to the voice communication technical field, relate in particular to a kind of auto gain control method.

Background technology

Current, voice communication has obtained very using widely, utilize voice communication, not only can realize intercommunication, also can further realize multi-party communication, for example, in video conferencing system, communication parties can realize synchronous AC by multipartite voice communication, each communication party can be in communication process a plurality of communication parties' of uppick voice.In the multipartite voice communication process, because the volume of each voice communication transmitting terminal is provided with difference, therefore, can causes the voice communication receiving terminal to receive the voice of different volumes simultaneously, thereby can cause the discontinuous of voice, influence the service quality of voice communication.In order to address this problem, need carry out balance to voice at the voice communication receiving terminal from each voice communication transmitting terminal, be adjusted to unified volume with volume, thereby eliminate the discontinuous of voice these voice.

In the prior art, carry out balance, specifically comprise by the achieve a butt joint voice of each voice transmitting terminal of receiving of auto gain control method:

For the voice flow that each voice transmitting terminal is sent, all handle respectively in the following manner:

Voice flow is divided into several processing blocks (block) according to fixing duration, comprise several sample points in each processing block, each sample point has amplitude separately respectively, communication receiver receives after the processing block at every turn, calculate each included sample point amplitude average absolute among this processing block, with this mean value the average energy value as processing block, then, with the energy value of the current broadcast sound volume correspondence of receiving terminal divided by this average energy value, obtain a multiplication factor, with the yield value of this multiplication factor as this processing block, again this yield value be multiply by the amplitude and the output of each sample point, thereby reach the purpose that the energy value of this processing block is amplified to the energy value of current broadcast sound volume correspondence; All carry out said process for each processing block in the voice flow, thereby realize the integral body of voice flow is amplified, the energy value of each processing block correspondence equals the energy value of current broadcast sound volume correspondence respectively in the voice flow behind feasible the amplification; Wherein, because the amplitude of sample point has nothing in common with each other in each processing block, therefore, for the yield value that each processing block calculated also just has nothing in common with each other.

Though the prior art can be by amplifying the purpose that reaches unified speech volume to voice flow,, have following shortcoming:

In the prior art, each processing block adopts identical yield value to amplify, thereby can guarantee that the unified of voice amplifies, still, because the yield value of adjacent processing block has nothing in common with each other, and the energy value of first or last sample point is different with the average energy value of processing block in the processing block, therefore, can cause the joining place of adjacent processing block the yield value cataclysm to occur, instantaneous great changes will take place thereby cause voice, makes voice discontinuous, influences the service quality of voice communication.

Summary of the invention

In view of this, main purpose of the present invention is to provide a kind of auto gain control method, utilizes this method, can realize that adjacent processing block joining place the yield value cataclysm can not occur in the voice flow, thereby guarantees the service quality of voice communication.

For achieving the above object, the invention provides a kind of auto gain control method, this method is to all carrying out following steps from the voice flow of each communication transmitting terminal:

Step a: according to the arrival order of processing block in the voice flow, determine current processing block, current volume according to communication receiver, calculate the yield value of current processing block, with the yield value of this yield value as last sample point correspondence in the current processing block, with, with the yield value of last sample point correspondence of the last processing block of current processing block yield value as first sample point correspondence in the current processing block;

Step b: according to the yield value of first sample point in the current processing block and last sample point correspondence, calculate the yield value of all the other sample point correspondences in the current processing block, make the yield value of first sample point correspondence can be smoothly transitted into the yield value of last sample point correspondence;

Step c: the amplitude of each sample point multiply by each self-corresponding yield value and output respectively in the current processing block, then, judges whether that other processing block arrives communication receiver after current processing block in addition, if, then return step a, otherwise, process ends.

Wherein, the described current volume according to communication receiver of the step a yield value that calculates current processing block comprises:

Calculate to deserve each included sample point amplitude average absolute among the pre-treatment piece, with this mean value the average energy value as processing block, then, with the energy value of the current broadcast sound volume correspondence of communication receiver divided by this average energy value, obtain a multiplication factor, with the yield value of this multiplication factor as current processing block.

Wherein, if current processing block is first processing block in the voice flow, the yield value of last sample point correspondence of a described last processing block with current processing block as the yield value of first sample point correspondence in the current processing block is:

With the yield value of a predefined yield value as described first sample point correspondence.

Wherein, described predefined yield value is 1.

Wherein, among the step b, the yield value of all the other sample point correspondences comprises in the current processing block of described calculating:

Calculate the yield value of all the other sample point correspondences in the current processing block according to formula (1);

Formula (1): ${\mathrm{Gain}}_i=\mathrm{Gain}\_\mathrm{first}+\frac{\mathrm{Gain}\_\mathrm{last}\_\mathrm{Gain}\_\mathrm{first}}{\mathrm{Block}\_\mathrm{size}-1}\×i$

Wherein, in the formula (1), i represents the sample point sequence number, is natural number, Gain _iThe yield value of i sample point correspondence in the expression processing block, Gain_first represents the yield value of first sample point correspondence in the processing block, Gain_last represents the yield value of last sample point correspondence in the processing block, and Block_size represents the sample point number that comprises in the sample point.

As seen, the present invention calculates corresponding yield value respectively to each sample point in the processing block, and make the yield value of first sample point of a processing block equate with the yield value of last sample point of a last processing block of this processing block, thereby the situation of yield value cataclysm appears in the joining place of avoiding adjacent processing block, thereby guarantee not take place the incoherent phenomenon of voice, guarantee the service quality of voice communication.

Description of drawings

Fig. 1 is for realizing flow chart of the present invention.

Embodiment

The present invention is a kind of auto gain control method, in the method, all carry out gain process for voice flow from each communication transmitting terminal, thereby with these voice flows unifications is the volume of current broadcast, in gain process process to each voice flow, this method is that last sample point and first sample point in the current processing block determined corresponding yield value respectively, so that two pairing yield values of sample point that adjacent processing block is connected mutually are identical, thereby the joining place of avoiding adjacent processing block goes out the situation of realize voice cataclysm, thereby guarantees the service quality of voice communication.

Describe the present invention below in conjunction with accompanying drawing.

Referring to Fig. 1, in the present invention, the voice flow from each communication transmitting terminal is all carried out gain process according to following steps:

Step 101: according to the arrival order of each processing block in the voice flow, determine current processing block, according to the current volume of communication receiver and the average energy value of current processing block, calculate the yield value of current processing block, with the yield value of this yield value as last sample point correspondence in the current processing block; Wherein, the yield value that can adopt method described in the prior art to calculate the average energy value of current processing block and calculate current processing block does not influence realization of the present invention;

Step 102: obtain the yield value of last sample point correspondence in the last processing block of current processing block, with the yield value of this yield value as first sample point correspondence in the current processing block; Wherein, if current processing block is first processing block in the voice flow, then with the yield value of a predefined yield value as first sample point correspondence, in embodiments of the present invention, this predefined yield value is 1; Wherein, the execution sequence of step 101 and step 102 does not influence realization of the present invention;

Step 103: according to the yield value of step 101 and 102 determined first sample points and last sample point correspondence, calculate the yield value of all the other sample point correspondences in the current processing block, make the yield value of first sample point correspondence in the current processing block can be smoothly transitted into the yield value of last sample point correspondence; Wherein, in embodiments of the present invention, calculate the yield value of all the other sample point correspondences in the current processing block according to formula (1);

Formula (1): ${\mathrm{Gain}}_i=\mathrm{Gain}\_\mathrm{first}+\frac{\mathrm{Gain}\_\mathrm{last}\_\mathrm{Gain}\_\mathrm{first}}{\mathrm{Block}\_\mathrm{size}-1}\×i$

Wherein, in the formula (1), i represents the sample point sequence number, is natural number, Gain _iThe yield value of i sample point correspondence in the expression processing block, Gain_first represents the yield value of first sample point correspondence in the processing block, Gain_last represents the yield value of last sample point correspondence in the processing block, and Block_size represents the sample point number that comprises in the sample point;

In step 103, by calculate each self-corresponding yield value of each sample point according to formula (1), make since first sample point of a processing block, the yield value of sample point correspondence is incremented to the pairing yield value of last sample point of this processing block gradually, thereby realized the seamlessly transitting of yield value of sample point correspondence in the processing block, therefore, avoided occurring the cataclysm of the yield value of sample point correspondence in the processing block, and then avoided the appearance of the discontinuous situation of voice that produced thus;

Step 104: the amplitude of each sample point multiply by each self-corresponding yield value and output respectively in the current processing block, thereby the speech volume of this processing block is amplified to the current volume of receiving terminal; Then, judge whether that other processing block arrives communication receiver after current processing block in addition, if, then return step 101, until sign off, otherwise, process ends.

By each processing block in the voice flow is distinguished the described content of execution in step 101～step 104, the speech volume of each processing block in the voice flow can be amplified to the current volume of receiving terminal, thereby realize with the current volume of whole voice flow unification, to finish gain process to voice flow to receiving terminal.

The present invention is applicable in the voice flow of any sample rate and quantification gradation that implementation is with mode is consistent as mentioned above.

The above only is preferred embodiment of the present invention, and is in order to restriction the present invention, within the spirit and principles in the present invention not all, any modification of being done, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (5)

1, a kind of auto gain control method is characterized in that, this method is to all carrying out following steps from the voice flow of each communication transmitting terminal:

Step a: according to the arrival order of processing block in the voice flow, determine current processing block, current volume according to communication receiver, calculate the yield value of current processing block, with the yield value of this yield value as last sample point correspondence in the current processing block, with, with the yield value of last sample point correspondence of the last processing block of current processing block yield value as first sample point correspondence in the current processing block;

Step b: according to the yield value of first sample point in the current processing block and last sample point correspondence, calculate the yield value of all the other sample point correspondences in the current processing block, make the yield value of first sample point correspondence can be smoothly transitted into the yield value of last sample point correspondence;

Step c: the amplitude of each sample point multiply by each self-corresponding yield value and output respectively in the current processing block, then, judges whether that other processing block arrives communication receiver after current processing block in addition, if, then return step a, otherwise, process ends.

2, method according to claim 1 is characterized in that, the yield value that the described current volume according to communication receiver of step a is calculated current processing block comprises:

Calculate to deserve each included sample point amplitude average absolute among the pre-treatment piece, with this mean value the average energy value as processing block, then, with the energy value of the current broadcast sound volume correspondence of communication receiver divided by this average energy value, obtain a multiplication factor, with the yield value of this multiplication factor as current processing block.

3, method according to claim 1, it is characterized in that, if current processing block is first processing block in the voice flow, the yield value of last sample point correspondence of a described last processing block with current processing block as the yield value of first sample point correspondence in the current processing block is:

With the yield value of a predefined yield value as described first sample point correspondence.

4, method according to claim 3 is characterized in that, described predefined yield value is 1.

5, method according to claim 1 is characterized in that, among the step b, the yield value of all the other sample point correspondences comprises in the current processing block of described calculating:

Calculate the yield value of all the other sample point correspondences in the current processing block according to formula (1);

Formula (1): ${\mathrm{Gain}}_i=\mathrm{Gain}\_\mathrm{first}+\frac{\mathrm{Gain}\_\mathrm{last}-\mathrm{Gain}\_\mathrm{first}}{\mathrm{Block}\_\mathrm{size}-1}\×i$

Wherein, in the formula (1), i represents the sample point sequence number, is natural number, Gain _iThe yield value of i sample point correspondence in the expression processing block, Gain_first represents the yield value of first sample point correspondence in the processing block, Gain_last represents the yield value of last sample point correspondence in the processing block, and Block_size represents the sample point number that comprises in the sample point.

Images