科大讯飞 文字转语音

1、wav音频文件的格式

wav文件由文件头和采样数据2部分组成。

文件头又分为RIFF（Resource Interchange File Format）、WAVE文件标识段 和 声音数据格式说明段组成。

各段的起始地址分别由RIFF标识符、WAVE标识符、以及波形格式标识符（FMT）标定。

（1）文件头格式

注意：下面的地址是连续的

虽然上图给出的数据标识符起始地址刚好是文件头的末地址+1，但并不代表总是这样。

因此，我们在读取数据时最好是找到数据标识符，该标识符的4个字节刚好是'd'、‘a’、‘t’、‘a’。

（2）数据格式

 

 

wav结构体定义:

typedef struct _wave_pcm_hdr
{
    char    riff[4]; //资源交换文件标志
    int     size_8;  //从下个地址开始到文件结尾的字节数  
    char    wave[4]; //wave文件标识  
    char    fmt[4];  //波形格式标识  
    int     fmt_size;//过滤字节(一般为00000010H)  

    short int formate_tag;      //格式种类，值为1时，表示PCM线性编码  
    short int channels;         //通道数，单声道为1，双声道为2  
    int       samples_per_sec;  //采样频率  
    int       avg_bytes_per_sec;//数据传输率 (每秒字节＝采样频率×每个样本字节数)  
    short int block_align;      //块对齐字节数 = channles * bit_samp / 8  
    short int bits_per_sample;  //bits per sample (又称量化位数) 

    char  data[4];
    int   data_size;
}_wave_pcm_hdr;
_wave_pcm_hdr default_wave_hdr=
{
    { 'R', 'I', 'F', 'F' },
    0,
    { 'W', 'A', 'V', 'E' },
    { 'f', 'm', 't', ' ' },
    16,
    1,
    1,
    16000,
    32000,
    2,
    16,
    { 'd', 'a', 't', 'a' },
    0

};

 

代码：

#include "stdafx.h"
#include "msp_cmn.h"
#include "msp_errors.h"
#include "qtts.h"
#include <string>
#include <Windows.h>
#ifdef _WIN64
#pragma comment (lib,"msc_x64.lib")
#else
#pragma comment (lib,"msc.lib")
#endif // _WIN64
 

#pragma comment(lib,"WinMM.lib")
typedef struct _wave_pcm_hdr
{
    char    riff[4]; //资源交换文件标志
    int     size_8;  //从下个地址开始到文件结尾的字节数  
    char    wave[4]; //wave文件标识  
    char    fmt[4];  //波形格式标识  
    int     fmt_size;//过滤字节(一般为00000010H)  

    short int formate_tag;      //格式种类，值为1时，表示PCM线性编码  
    short int channels;         //通道数，单声道为1，双声道为2  
    int       samples_per_sec;  //采样频率  
    int       avg_bytes_per_sec;//数据传输率 (每秒字节＝采样频率×每个样本字节数)  
    short int block_align;      //块对齐字节数 = channles * bit_samp / 8  
    short int bits_per_sample;  //bits per sample (又称量化位数) 

    char  data[4];
    int   data_size;
}_wave_pcm_hdr;
_wave_pcm_hdr default_wave_hdr=
{
    { 'R', 'I', 'F', 'F' },
    0,
    { 'W', 'A', 'V', 'E' },
    { 'f', 'm', 't', ' ' },
    16,
    1,
    1,
    16000,
    32000,
    2,
    16,
    { 'd', 'a', 't', 'a' },
    0

};

int _tmain(int argc, _TCHAR* argv[])
{
    const char* usr = NULL;
    const char* pwd = NULL;
    const char* lgi_param = "appid = 58610d7f";
    int ret = MSPLogin(usr, pwd, lgi_param);
    if (MSP_SUCCESS != ret)
    {
        printf("MSPLogin failed, error code is: %d", ret);
    }


    const char * ssb_param = "voice_name = xiaoyan, aue = speex-wb;7, sample_rate = 16000, speed = 50, volume = 50, pitch = 50, rdn = 2";
     ret = -1;
    const char * sessionID = QTTSSessionBegin(ssb_param, &ret);
    if (MSP_SUCCESS != ret)
    {
        printf("QTTSSessionBegin failed, error code is : %d", ret);
    }
    const char* src_text;
    char ch[1000];
    src_text=gets(ch);

   
    unsigned int text_len = strlen(src_text); //textLen参数为合成文本所占字节数
     ret = QTTSTextPut(sessionID, src_text, text_len, NULL);
    if (MSP_SUCCESS != ret)
    {
        printf("QTTSTextPut failed, error code is : %d", ret);
    }



    FILE* fp = fopen("112.wav", "wb");
    fwrite(&default_wave_hdr, sizeof(default_wave_hdr),1,fp);
    unsigned int audio_len = 0;
    int synth_status = 0;
    while (1)
    {
        const void * data = QTTSAudioGet(sessionID, &audio_len, &synth_status, &ret);
        if (NULL != data)
        {
            fwrite(data, audio_len, 1, fp);
            default_wave_hdr.data_size += audio_len;
        }
        if (MSP_TTS_FLAG_DATA_END == synth_status || MSP_SUCCESS != ret)
        {
            break;
        }
    }
    fclose(fp);
    default_wave_hdr.size_8 += default_wave_hdr.data_size + (sizeof(default_wave_hdr) - 8);
    fseek(fp, 4, 0);
    fwrite(&default_wave_hdr.size_8, sizeof(default_wave_hdr.size_8), 1, fp);
    fseek(fp, 40, 0);
    fwrite(&default_wave_hdr.data_size, sizeof(default_wave_hdr.data_size), 1, fp);
    fclose(fp);

    ret = QTTSSessionEnd(sessionID, "normal end");
    if (MSP_SUCCESS != ret)
    {
        printf("QTTSSessionEnd failed, error code is : %d", ret);
    }

    PlaySoundA("qweqwr.wav", NULL, SND_ALIAS);

    ret = MSPLogout();
    if (MSP_SUCCESS != ret)
    {
        printf("MSPLogout failed, error code is: %d", ret);
    }
    system("pause");
    return 0;
}

代码