windows下sapi做TTS转化（修复）总结

TTS转换有很多第三方库，比如科大讯飞的sdk就很好用，但是对于一些简单的需求，用windows（win7)的sapi就可以了。

sapi运用是很简单，但是在很多电脑上就上放不出声音，原因是这些电脑系统不是官网下载的，TTS功能被阉割了，本文记录的主要就是修复这个问题。

由于不同电脑缺少的语音组件不同，所以试了一堆修复包，也不知道这些修复包谁有用，总之不行全执行一遍就行。

为了以后下载方便，把他们存在了百度云盘中，地址：链接: https://pan.baidu.com/s/1PbryiE0gaW835zqwd52mCg 提取码: haee

自带一个示例程序tts_helloworld.

#include <array>
#include <list>
#include <algorithm>
#include "stdafx.h"
#include "../Include/sapi.h"
#include "../Include/sphelper.h"
#define _CRT_SECURE_NO_WARNINGS

#pragma comment(lib,"../Lib/sapi.lib")

int Speak(const char *Text)
{
	ISpVoice *pVoice = NULL;
	WCHAR  Wchar[1024];
	MultiByteToWideChar(CP_ACP, 0, Text, strlen(Text) + 1, Wchar, 1024 / sizeof(Wchar[0]));
	//初始化COM接口
	if (FAILED(::CoInitialize(NULL)))
		return -1;

	//获取SpVoice接口
	HRESULT hr = CoCreateInstance(CLSID_SpVoice, NULL, CLSCTX_ALL, IID_ISpVoice, (void**)&pVoice);
	if (SUCCEEDED(hr))
	{
		ISpObjectToken *p = NULL;
		hr = SpFindBestToken(SPCAT_VOICES, L"language=804", L"Name=VM Lily", &p);
		if (SUCCEEDED(hr))//说话人控制 
		{
			pVoice->SetVoice(p);
		}
		pVoice->SetRate(-2);
		pVoice->SetVolume(100);

		hr = pVoice->Speak(Wchar, 0, NULL);
		pVoice->Release();
		pVoice = NULL;
	}
	//释放com资源
	::CoUninitialize();
	return 0;
}

int SaveVoice(const char *Text, const char* FileName)
{
	CComPtr<ISpVoice> m_cpVoice;
	WCHAR  Wchar[1024];
	MultiByteToWideChar(CP_ACP, 0, Text, strlen(Text) + 1, Wchar, 1024 / sizeof(Wchar[0]));

	//初始化COM接口
	if (FAILED(::CoInitialize(NULL)))
		return -1;

	HRESULT hr = m_cpVoice.CoCreateInstance(CLSID_SpVoice);

	if (SUCCEEDED(hr))
	{
		USES_CONVERSION;
		CComPtr<ISpStream> cpWavStream;
		CComPtr<ISpStreamFormat>    cpOldStream;
		CSpStreamFormat OriginalFmt;

		hr = m_cpVoice->GetOutputStream(&cpOldStream);
		if (hr == S_OK)
			hr = OriginalFmt.AssignFormat(cpOldStream);
		else
			hr = E_FAIL;

		if (SUCCEEDED(hr))
			hr = SPBindToFile(FileName, SPFM_CREATE_ALWAYS, &cpWavStream, &OriginalFmt.FormatId(), OriginalFmt.WaveFormatExPtr());

		if (SUCCEEDED(hr))
			hr = m_cpVoice->SetOutput(cpWavStream, TRUE);

		if (SUCCEEDED(hr))
		{
			ISpObjectToken *p = NULL;
			hr = SpFindBestToken(SPCAT_VOICES, L"language=804", L"Name=Microsoft lili", &p);
			if (SUCCEEDED(hr))//说话人控制 
			{
				m_cpVoice->SetVoice(p);
			}
			m_cpVoice->SetRate(-2);
			m_cpVoice->SetVolume(100);
			hr = m_cpVoice->Speak(Wchar, SPF_ASYNC | SPF_IS_NOT_XML, 0);
		}
		m_cpVoice->WaitUntilDone(INFINITE);
		cpWavStream.Release();

		m_cpVoice->SetOutput(cpOldStream, FALSE);
		m_cpVoice.Release();
		m_cpVoice = NULL;
	}
	::CoUninitialize();
	return 0;
}