讯飞C/C++语音合成基础篇

首先到讯飞上注册一个帐号(感觉讯飞官网和阿里云官网的差不多)

随后创建一个新应用(Windows平台下)

最后我们要的是Appid如下图所示

下面是讯飞SDK(含说明文档)下载地址：

链接：http://pan.baidu.com/s/1i45xEo1 密码：m8il

因为此程序涉及WAV语音知识，未了解WAV的朋友请百度补充，或者在下面这个链接初略的学习下

http://blog.csdn.net/qq78442761/article/details/53385561

下面先来认识下几个API(从文档提取出来)

1.int MSPAPI MSPLogin (const char *usr, const char *pwd, const char *params) //初始化msc，用户登录。

usr[in] 此参数保留，传入NULL即可。
pwd[in] 此参数保留，传入NULL即可。

官方例子

const char* usr = NULL;
const char* pwd = NULL;
const char* lgi_param = "appid = ********";
int ret = MSPLogin(usr, pwd, lgi_param);
if( MSP_SUCCESS != ret )
{
    printf( "MSPLogin failed, error code is: %d", ret );
}

2.const char *MSPAPI QTTSSessionBegin (const char *params, int *errorCode) //开始一次语音合成，分配语音合成资源。

(信息有点多,参数具体见文档qtts.h 文件参考)

下面是官方例子

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

例子

FILE* fp = fopen("tts.pcm", "wb");
while (1)
{
    const void * data = QTTSAudioGet(sessionID, &audio_len, &synth_status, &ret);
    if (NULL != data)
    {
        fwrite(data, audio_len, 1, fp);
    }
    if (MSP_TTS_FLAG_DATA_END == synth_status || MSP_SUCCESS != ret)
    {
        break;
    }
}
fclose(fp);

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

几个个文件1.stdafx.h。2.targetver.h。3.QTTSDemo.cpp。4.stdafx.cpp。

除了QTTSDemo.cpp外，其他文件都是vs2013创建控制台程序自带的，在此只给出QTTSDemo.cpp代码

但注意在stdafx.h中加入#define _CRT_SECURE_NO_WARNINGS

QTTSDemo.cpp

// QTTSDemo.cpp : 定义控制台应用程序的入口点。
//

#include "stdafx.h"
#include <msp_cmn.h>
#include <msp_errors.h>
#include <qtts.h>
#include <string.h>
#include <windows.h>

#pragma  comment(lib,"WinMM.lib")

#ifdef _WIN64
#pragma comment(lib,"msc_x64.lib")
#else
#pragma comment(lib,"msc.lib")
#endif

//wav音频头部格式
typedef struct _wave_pcm_hdr
{
	char			riff[4];			//="RIFF"
	int				size_8;				//=FileSize=8
	char			wave[4];			//="WAVE"
	char			fmt[4];				//="fmt"
	int				fmt_size;			//=下一个结构体的大小:16

	short int		format_tag;			//=PCM:1
	short int		channels;			//=通道数:1
	int				samples_per_sec;	//=采样率：8000|6000|16000
	int				avg_bytes_per_sec;	//=每秒字节数：samples_per_sec*bit_per_sample
	short int		block_align;		//=每采样点字节数：wBitsPerSample/8
	short int		bits_per_sample;	//=量化比特数:8|16

	char			data[4];			//="data";
	int				data_size;			//=纯数据长度：FileSize-44
}wave_pcm_hdr;

/*默认wav音频头部数据*/
wave_pcm_hdr default_wav_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[])
{
	//登录
	const char* usr = NULL;
	const char* pwd = NULL;
	const char* lgi_param = "appid = 583aea17";
	int ret = MSPLogin(usr, pwd, lgi_param);
	if (MSP_SUCCESS != ret)
	{
		printf("MSPLogin failed, error code is: %d", ret);
	}

	//开始合成
	const char * ssb_param = "voice_name = xiaorong, aue = speex-wb;7, sample_rate = 16000, speed = 50, volume = 80, 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);
	}

	//设置待合成文本
	char src_text[1024];
	printf("请输入一段文字(中文注意逗号,句话)：n");
	gets(src_text);
	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);
	}

	//获取合成的音频wav
	FILE* fp = fopen("Demo.wav", "wb");	//一定是二进制模式
	fwrite(&default_wav_hdr, sizeof(default_wav_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_wav_hdr.data_size += audio_len;
		}
		if (MSP_TTS_FLAG_DATA_END == synth_status || MSP_SUCCESS != ret)
		{
			break;
		}
	}

	default_wav_hdr.size_8 += default_wav_hdr.data_size + (sizeof(default_wav_hdr)-8);
	fseek(fp, 4, 0);
	fwrite(&default_wav_hdr.size_8, sizeof(default_wav_hdr.size_8), 1, fp);	//写入size_8的值
	fseek(fp, 40, 0);	//将文件指针偏移到存储data_size值的位置
	fwrite(&default_wav_hdr.data_size, sizeof(default_wav_hdr.data_size), 1, fp);//写入data_size的值
	fclose(fp);
	PlaySoundA("Demo.wav", NULL, SND_ASYNC);
	ret = QTTSSessionEnd(sessionID, "normal end");
	if (MSP_SUCCESS != ret)
	{
		printf("QTTSSessionEnd failed, error code is : %d", ret);
	}
	//退出
	ret = MSPLogout();
	if (MSP_SUCCESS != ret)
	{
		printf("MSPLogout failed, error code is: %d", ret);
	}

	system("pause");
	return 0;
}

最后

以上就是热心魔镜为你收集整理的讯飞C/C++语音合成基础篇的全部内容，希望文章能够帮你解决讯飞C/C++语音合成基础篇所遇到的程序开发问题。

如果觉得靠谱客网站的内容还不错，欢迎将靠谱客网站推荐给程序员好友。