混音算法
转载：http://blog.csdn.net/dancing_night/article/details/53080819
Wav文件直接反映了一个声音在每个时刻的大小值，比如说以下一段波形：
我们按每人0.1秒取一点，得到的wav文件数值就是0,1,1,-1,0,1。因此，假如我们能把许多Wav文件的数据直接相加，你听到的就是所有的声音，这就是混音器的原理。
Step 1, Get the Raw data of the two files, (Example, of the sample 8bit and 8Kh, means one sample is of 8bit)
Step 2 Let the two audio signal be A and B respectively, the range is between 0 and 255. Where A and B are the Sample Values (Each raw data) And store the resultant into the Y
If Both the samples Values are possitv Y = A + B - A * B / 255
Where Y is the resultant signal which contains both signal A and B, merging two audio streams into single
stream by this method solves the problem of overflow and information loss to an extent.
If the range of 8-bit sampling is between -127 to 128
If both A and B are negative Y = A +B - (A * B / (-127))
Else Y = A + B - A * B / 128
Similarly for the nbit (ex 16bit data)
For n-bit sampling audio signal
If both A and B are negative Y = A + B - (A * B / (-(2 pow(n-1) -1)))
Else Y = A + B - (A * B / (2 pow(n-1))
Step 3.
Add the Header to the Resultant (mixed) data and play back.
If some thing is unclear and ambigious let me know.
Regards
Ranjeet Gupta.
还有简单C程序示意代码，但是其中包含了核心算法：
#include
#include
#include
#include
int main(int argc,char *argv[]) {
char mixname[255];
FILE *pcm1, *pcm2, *mix;
char sample1, sample2;
int value;
pcm1 = fopen(argv[1],"r");
pcm2 = fopen(argv[2],"r");
strcpy (mixname, argv[1]);
strcat (mixname, "_temp.wav");
mix = fopen(mixname, "w");
while(!feof(pcm1)) {
sample1 = fgetc(pcm1);
sample2 = fgetc(pcm2);
if ((sample1 < 0) && (sample2 < 0)) {
value = sample1 + sample2 - (sample1 * sample2 / -(pow(2,16-1)-1));
}else{
value = sample1 + sample2 - (sample1 * sample2 / (pow(2,16-1)-1));
}
fputc(value, mix);
}
fclose(pcm1);
fclose(pcm2);
fclose(mix);
return 0;
}
自己的混音(混音麦克风和扬声器)：16位的数据，双声道
//将PCM叠加
for (int i = 0; i < oAcc->frame_size*2; i=i+2)
{
uint8_t* pMicOut = frame_audioMicOut->extended_data[0] + i;
uint8_t* pMicIn = frame_audioMicIn->extended_data[0] + i;
short tempMicOut = *(short*)pMicOut;
short tempMicIn = *(short*)pMicIn;
int tempOut = 0;
if (tempMicOut < 0 && tempMicIn < 0)
tempOut = tempMicOut + tempMicIn - tempMicOut*tempMicIn / (-(pow(2, 15) - 1));
else if (tempMicOut > 0 && tempMicIn > 0)
tempOut = tempMicOut + tempMicIn - tempMicOut*tempMicIn / (pow(2, 15));
pMicIn = (uint8_t*)tempOut;
}
线性叠加后求平均
优点：不会产生溢出，噪音较小；
缺点：衰减过大，影响通话质量；
short remix(short buffer1,short buffer2)
{
int value = buffer1 + buffer2;
return (short)(value/2);
}- 1
- 2
- 3
- 4
- 5
- 1
- 2
- 3
- 4
- 5
归一化混音(自适应加权混音算法)
思路：
使用更多的位数(32 bit)来表示音频数据的一个样本，混完音后在想办法降低其振幅，使其仍旧分布在16 bit所能表示的范围之内，这种方法叫做归一法.
方法：
为避免发生溢出，使用一个可变的衰减因子对语音进行衰减。这个衰减因子也就代表语音的权重，衰减因子随着音频数据的变化而变化，所以称为自适应加权混音。当溢出时，衰减因子较小，使得溢出的数据在衰减后能够处于临界值以内，而在没有溢出时，又让衰减因子慢慢增大，使数据较为平缓的变化.
代码:
void Mix(char sourseFile[10][SIZE_AUDIO_FRAME],int number,char *objectFile)
{
//归一化混音
int const MAX=32767;
int const MIN=-32768;
double f=1;
int output;
int i = 0,j = 0;
for (i=0;i2;i++)
{
int temp=0;
for (j=0;jshort*)(sourseFile[j]+i*2);
}
output=(int)(temp*f);
if (output>MAX)
{
f=(double)MAX/(double)(output);
output=MAX;
}
if (outputdouble)MIN/(double)(output);
output=MIN;
}
if (f<1)
{
f+=((double)1-f)/(double)32;
}
*(short*)(objectFile+i*2)=(short)output;
}
} - 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
下面是我从newlc上找到的一个关于PCM脉冲编码的音频信号的混音实现，其中包含了一个关键的混音算法！
if( data1 < 0 && data2 < 0)
date_mix = data1+data2 - (data1 * data2 / -(pow(2,16-1)-1));
else
date_mix = data1+data2 - (data1 * data2 / (pow(2,16-1)-1)); - 1
- 2
- 3
- 4
- 1
- 2
- 3
- 4
切割时间片，重采样算法
可以把各个通道的声音叠到一起，让声音的采样率按倍增加，如果提高声音的播放频率，声音可以正常的播放，声音实现了叠加；如果不想修改声音的播放输出频率，可以通过声音的重采样后输出自己想要的输出频率；
下面是上面的混音的测试代码：
#include
#include
#include
#define IN_FILE1 "1.wav"
#define IN_FILE2 "2.wav"
#define OUT_FILE "remix.pcm"
#define SIZE_AUDIO_FRAME (2)
void Mix(char sourseFile[10][SIZE_AUDIO_FRAME],int number,char *objectFile)
{
//归一化混音
int const MAX=32767;
int const MIN=-32768;
double f=1;
int output;
int i = 0,j = 0;
for (i=0;i2;i++)
{
int temp=0;
for (j=0;jshort*)(sourseFile[j]+i*2);
}
output=(int)(temp*f);
if (output>MAX)
{
f=(double)MAX/(double)(output);
output=MAX;
}
if (outputdouble)MIN/(double)(output);
output=MIN;
}
if (f<1)
{
f+=((double)1-f)/(double)32;
}
*(short*)(objectFile+i*2)=(short)output;
}
}
int main()
{
FILE * fp1,*fp2,*fpm;
fp1 = fopen(IN_FILE1,"rb");
fp2 = fopen(IN_FILE2,"rb");
fpm = fopen(OUT_FILE,"wb");
short data1,data2,date_mix;
int ret1,ret2;
char sourseFile[10][2];
while(1)
{
ret1 = fread(&data1,2,1,fp1);
ret2 = fread(&data2,2,1,fp2);
*(short*) sourseFile[0] = data1;
*(short*) sourseFile[1] = data2;
if(ret1>0 && ret2>0)
{
Mix(sourseFile,2,(char *)&date_mix);
/*
if( data1 < 0 && data2 < 0)
date_mix = data1+data2 - (data1 * data2 / -(pow(2,16-1)-1));
else
date_mix = data1+data2 - (data1 * data2 / (pow(2,16-1)-1));*/
if(date_mix > pow(2,16-1) || date_mix < -pow(2,16-1))
printf("mix error\n");
}
else if( (ret1 > 0) && (ret2==0))
{
date_mix = data1;
}
else if( (ret2 > 0) && (ret1==0))
{
date_mix = data2;
}
else if( (ret1 == 0) && (ret2 == 0))
{
break;
}
fwrite(&date_mix,2,1,fpm);
}
fclose(fp1);
fclose(fp2);
fclose(fpm);
printf("Done!\n");
}