repos_util/TTSpeech/Text2Speeck/Text2Speech.cpp

// Text2Speech.h  : CText2Speech 的实现;

#include "StdAfx.h"
#include "Text2Speech.h"


CText2Speech::CText2Speech(void)
{
	m_cpVoice = NULL;
	m_cpOutAudio = NULL;
	//m_sError=_T("");
}


CText2Speech::~CText2Speech(void)
{
	Destroy();
}

// 初始化函数Initialize首先初始化COM库，并调用CoCreateInstance方法初始化语音引擎。
// 然后设置必须响应的引擎事件，并指定响应事件消息的窗口句柄。
// 该窗口句柄是作为函数的参数传入的
BOOL CText2Speech::Initialize(IN HWND hWnd)
{
	HRESULT hr = ::CoInitialize(NULL);
	if (FAILED(hr))	{
		//m_sError=_T("Error intialization COM");
		return FALSE;
	}

	hr = m_cpVoice.CoCreateInstance(CLSID_SpVoice);
	if (FAILED(hr)) {
		//m_sError=_T("Error creating voice");
		return FALSE;
	}

	// Sets the type of events which will invoke a notification and become queued
	hr = m_cpVoice->SetInterest(SPFEI(SPEI_VISEME), SPFEI(SPEI_VISEME));
	if (FAILED(hr))
	{
		//m_sError=_T("Error creating interest...seriously");
		return FALSE;
	}

	if (::IsWindow(hWnd))
	{
		hr = m_cpVoice->SetNotifyWindowMessage(hWnd, WM_TTSEVENT, 0, 0);
		if (FAILED(hr))
		{
			//m_sError=_T("Error setting notification window");
			return FALSE;
		}
	}
	return TRUE;
}

// 释放函数则释放语音引擎接口和COM库，其代码如下：
void CText2Speech::Destroy()
{
	if (m_cpVoice)
		m_cpVoice.Release();
	CoUninitialize();
}

// 要朗读的文字必须位于宽字符串中;
HRESULT CText2Speech::Speak(IN const TCHAR *pText, IN DWORD dwFlags /* = SPF_DEFAULT */)
{
#ifndef UNICODE
	size_t len = strlen(pText) + 1;
	wchar_t *WStr;
	WStr = (wchar_t*)malloc(len*sizeof(wchar_t));
	MultiByteToWideChar(CP_ACP, 0, pText, -1, WStr, len*sizeof(wchar_t));
	HRESULT hr = m_cpVoice->Speak(WStr, dwFlags, NULL);
	free(WStr);
	return hr;
#else
	return m_cpVoice->Speak(pText, dwFlags, NULL);
#endif
}

/*
语言函数的实现比较复杂。
由于IspVoice接口提供的语言函数，都只与抽象的语音语言接口ISpObjectToken相关，
而我们能看到的却是语音语言的描述，比如，通过控制面板的语音程序所能见到的就是语音语言的描述。
因此，笔者设计了直接对语音语言进行操作的语言函数，包括获取系统中已安装的语音语言数目，
设置指定的语音语言，获取指定的语音语言描述（包括当前设定的语音语言）。
*/
ULONG CText2Speech::GetVoiceCount()
{
	HRESULT                             hr = S_OK;
	CComPtr<ISpObjectToken>             cpVoiceToken;
	CComPtr<IEnumSpObjectTokens>        cpEnum;
	ULONG                               ulCount = -1;

	//Enumerate the available voices 
	hr = SpEnumTokens(SPCAT_VOICES, NULL, NULL, &cpEnum);
	if (FAILED(hr))
	{
		//m_sError = _T("Error to enumerate voices");
		return -1;
	}

	//Get the number of voices
	hr = cpEnum->GetCount(&ulCount);
	if (FAILED(hr))
	{
		//m_sError = _T("Error to get voice count");
		return -1;
	}

	return ulCount;
}

HRESULT CText2Speech::GetVoice(OUT WCHAR **ppszDescription, IN ULONG lIndex)
{
	HRESULT                           hr = S_OK;
	CComPtr<ISpObjectToken>           cpVoiceToken;
	CComPtr<IEnumSpObjectTokens>      cpEnum;
	ULONG                             ulCount = 0;

	// lIndex==-1,获取当前默认的语声;
	if (lIndex == -1)
	{
		// current voice
		hr = m_cpVoice->GetVoice(&cpVoiceToken);
		if (FAILED(hr))
		{
			//m_sError = _T("Error to get current voice");
			return hr;
		}

		SpGetDescription(cpVoiceToken, ppszDescription);
		if (FAILED(hr))
		{
			//m_sError = _T("Error to get current voice description");
			return hr;
		}
	}
	else
	{
		// else other voices, we should enumerate the voice list first
		//Enumerate the available voices 
		hr = SpEnumTokens(SPCAT_VOICES, NULL, NULL, &cpEnum);
		if (FAILED(hr))
		{
			//m_sError = _T("Error to enumerate voices");
			return hr;
		}
		//Get the number of voices
		hr = cpEnum->GetCount(&ulCount);
		if (FAILED(hr))
		{
			//m_sError = _T("Error to voice count");
			return hr;
		}

		// range control
		ASSERT(lIndex >= 0);
		ASSERT(lIndex < ulCount);
		// Obtain specified voice id
		ULONG l = 0;
		while (SUCCEEDED(hr))
		{
			cpVoiceToken.Release();
			hr = cpEnum->Next(1, &cpVoiceToken, NULL);
			if (FAILED(hr))
			{
				//m_sError = _T("Error to get voice token");
				return hr;
			}

			if (l == lIndex)
			{
				hr = SpGetDescription(cpVoiceToken, ppszDescription);
				if (FAILED(hr))
				{
					//m_sError = _T("Error to get voice description");
					return hr;
				}
				break;
			}
			l++;
		}
	}
	return hr;
}

HRESULT CText2Speech::SetVoice(IN WCHAR **ppszDescription)
{
	HRESULT                             hr = S_OK;
	CComPtr<ISpObjectToken>             cpVoiceToken;
	CComPtr<IEnumSpObjectTokens>        cpEnum;
	ULONG                               ulCount = 0;

	//Enumerate the available voices 
	hr = SpEnumTokens(SPCAT_VOICES, NULL, NULL, &cpEnum);
	if (FAILED(hr))
	{
		//m_sError = _T("Error to enumerate voices");
		return hr;
	}
	//Get the number of voices
	hr = cpEnum->GetCount(&ulCount);
	if (FAILED(hr))
	{
		//m_sError = _T("Error to voice count");
		return hr;
	}
	// Obtain specified voice id
	while (SUCCEEDED(hr) && ulCount--)
	{
		cpVoiceToken.Release();
		hr = cpEnum->Next(1, &cpVoiceToken, NULL);
		if (FAILED(hr))
		{
			//m_sError = _T("Error to voice token");
			return hr;
		}
		WCHAR *pszDescription1;
		hr = SpGetDescription(cpVoiceToken, &pszDescription1);
		if (FAILED(hr))
		{
			//m_sError = _T("Error to get voice description");
			return hr;
		}
		if (!wcsicmp(pszDescription1, *ppszDescription))
		{
			hr = m_cpVoice->SetVoice(cpVoiceToken);
			if (FAILED(hr))
			{
				//m_sError = _T("Error to set voice");
				return hr;
			}

			break;
		}
	}

	return hr;
}

HRESULT CText2Speech::SetVoice(IN const ULONG &uIndex)
{
	HRESULT hr;
	CComPtr<ISpObjectToken>             cpVoiceToken; // 用下面的函数获取当前正在使用的语音;
	CComPtr<IEnumSpObjectTokens>        cpEnum;
	ULONG                               ulCount = 0;

	hr = SpEnumTokens(SPCAT_VOICES, NULL, NULL, &cpEnum);
	if (FAILED(hr)){
		return hr;
	}

	// Get the number of voices
	// 获取语音语言数量;
	hr = cpEnum->GetCount(&ulCount);
	if (FAILED(hr)) {
		return hr;
	}

	// 根据参数 nLang ,设置为所需要的语音语言;
	hr = cpEnum->Item(uIndex, &cpVoiceToken);
	if (SUCCEEDED(hr)){
		m_cpVoice->SetVoice(cpVoiceToken);
	}
	else
	{
		return hr;
	}

	return hr;
}

HRESULT CText2Speech::GetOutAudio(OUT SPSTREAMFORMAT &eFmt)
{
	CComPtr<ISpStreamFormat> cpStream;
	HRESULT hr = m_cpVoice->GetOutputStream(&cpStream);
	//if (hr == S_OK)
	if (SUCCEEDED(hr))
	{
		CSpStreamFormat Fmt;
		HRESULT hStreamFmt = Fmt.AssignFormat(cpStream);
		if (SUCCEEDED(hStreamFmt))
		{
			eFmt = Fmt.ComputeFormatEnum();
		}
		else
		{
			return hStreamFmt;
		}
	}

	return hr;
}

HRESULT CText2Speech::SetOutAudio(IN const SPSTREAMFORMAT &eFmt)
{
	HRESULT hr = S_OK;
	if (!m_cpOutAudio)
	{
		hr = SpCreateDefaultObjectFromCategoryId(SPCAT_AUDIOOUT, &m_cpOutAudio); //创建接口
		if ( FAILED(hr) )
			return hr;
	}

	CSpStreamFormat Fmt;
	hr = Fmt.AssignFormat(eFmt);
	if (SUCCEEDED(hr))
	{
		hr = m_cpOutAudio->SetFormat(Fmt.FormatId(), Fmt.WaveFormatExPtr());
	}
	else
	{
		return hr;//hr = E_FAIL;
	}

	if (SUCCEEDED(hr)){
		m_cpVoice->SetOutput(m_cpOutAudio, FALSE);
	}

	return hr;
}

// 播放WAV,通过ISpStream接口实现; 
HRESULT CText2Speech::PlayWav(IN const TCHAR *szWavFileName)
{
	CComPtr<ISpStream>       cpWavStream;
#ifndef UNICODE
	WCHAR                    szwWavFileName[MAX_PATH] = L"";
	USES_CONVERSION;
	wcscpy(szwWavFileName, T2W(szWavFileName));//从ANSI将WAV文件的名字转换成宽字符串
	//使用sphelper.h 提供的这个函数打开 wav 文件，并得到一个 IStream 指针
	HRESULT hr = SPBindToFile(szwWavFileName, SPFM_OPEN_READONLY, &cpWavStream);
#else
	HRESULT hr = SPBindToFile(szWavFileName, SPFM_OPEN_READONLY, &cpWavStream);
#endif
	if (SUCCEEDED(hr))
	{
		m_cpVoice->SpeakStream(cpWavStream, SPF_ASYNC, NULL);//播放WAV文件
	}

	return hr;
}

HRESULT CText2Speech::Save2Wav(IN const TCHAR *pText, IN const TCHAR *szWavFileName)
{
#ifndef UNICODE
	//TCHAR szFileName[256];//假设这里面保存着目标文件的路径
	USES_CONVERSION;
	WCHAR szWFileName[MAX_PATH];
	wcscpy(szWFileName, T2W(szWavFileName));	// ANSI转换成宽字符串; 
#endif
	//创建一个输出流，绑定到wav文件;
	CSpStreamFormat				OriginalFmt;
	CComPtr<ISpStream>			cpWavStream;
	CComPtr<ISpStreamFormat>    cpOldStream;

	HRESULT hr = m_cpVoice->GetOutputStream(&cpOldStream);
	if (FAILED(hr)) {
		return hr;
	}
	hr = OriginalFmt.AssignFormat(cpOldStream);

	// 使用sphelper.h中提供的函数创建 wav 文件;
	if (SUCCEEDED(hr))
	{
		OriginalFmt.m_pCoMemWaveFormatEx->wFormatTag = WAVE_FORMAT_PCM;
#ifndef UNICODE
		hr = SPBindToFile(szWFileName, SPFM_CREATE_ALWAYS, &cpWavStream,
			&OriginalFmt.FormatId(), OriginalFmt.WaveFormatExPtr());
#else
		hr = SPBindToFile(szWavFileName, SPFM_CREATE_ALWAYS, &cpWavStream,
			&OriginalFmt.FormatId(), OriginalFmt.WaveFormatExPtr());
#endif
	}
	else
	{
		return hr;
	}

	if (SUCCEEDED(hr))
	{
		//设置声音的输出到 wav 文件，而不是 speakers;
		hr = m_cpVoice->SetOutput(cpWavStream, TRUE);
	}
	else
	{
		return hr;
	}

	// 开始朗读;
	hr = Speak(pText, SPF_ASYNC | SPF_IS_NOT_XML);

	// 等待朗读结束;
	m_cpVoice->WaitUntilDone(INFINITE);
	cpWavStream.Release();

	// 把输出重新定位到原来的流;
	m_cpVoice->SetOutput(cpOldStream, FALSE);

	return hr;
}

/* 
如果不想处理有关语言的细节问题，
只想显示和选择系统可提供的语音引擎，
则可以直接调用Speech SDK提供的两个帮助函数SpInitTokenComboBox和
SpInitTokenListBox来实现语音语言的显示和选择，其代码如下：*/
#if 0
HRESULT SpInitTokenComboBox(	
							HWND		hwnd,
							const WCHAR* pszCatName,
							const WCHAR* pszRequiredAttrib = NULL,
							const WCHAR* pszOptionalAttrib = NULL
							);	
HRESULT SpInitTokenListBox(
						   HWND		hwnd,
						   const WCHAR* pszCatName,
						   const WCHAR* pszRequiredAttrib = NULL,
						   const WCHAR* pszOptionalAttrib = NULL
						   );
#endif