/* Copyright (C) 2003-2015 LiveCode Ltd.

This file is part of LiveCode.

LiveCode is free software; you can redistribute it and/or modify it under

the terms of the GNU General Public License v3 as published by the Free

Software Foundation.

LiveCode is distributed in the hope that it will be useful, but WITHOUT ANY

WARRANTY; without even the implied warranty of MERCHANTABILITY or

FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License

for more details.

You should have received a copy of the GNU General Public License

along with LiveCode. If not see <http://www.gnu.org/licenses/>. */

#include <windows.h>

#include <initguid.h>

#include "w32sapi5speech.h"

#define VOICE_TYPE_STRING 64

#define PITCH_XML_STRING 30

static char *path_from_native(const char *p_path)

{

char *t_path;

t_path = strdup(p_path);

char *dptr;

dptr = t_path;

if (!*dptr)

return dptr;

do {

if (*dptr == '\\')

*dptr = '/';

} while (*++dptr);

return t_path;

}

static char *path_to_native(const char *p_path)

{

char *t_path;

t_path = strdup(p_path);

char *dptr;

dptr = t_path;

if (!*dptr)

return dptr;

do {

if (*dptr == '/')

*dptr = '\\';

} while (*++dptr);

return t_path;

}

// MW-2014-08-06: [[ UnicodeSpeech ]] Utility routine to convert UTF8 to UTF16.

static wchar_t *utf8tow(const char *p_utf8)

{

wchar_t *t_utf16;

t_utf16 = new wchar_t[strlen(p_utf8) + 1];

int t_length;

t_length = MultiByteToWideChar(CP_UTF8, 0, p_utf8, strlen(p_utf8), t_utf16, strlen(p_utf8));

t_utf16[t_length] = 0;

return t_utf16;

}

// MW-2014-08-06: [[ UnicodeSpeech ]] Utility routine to convert Native (1252) to UTF16.

static wchar_t *nativetow(const char *p_native)

{

wchar_t *t_utf16;

t_utf16 = new wchar_t[strlen(p_native) + 1];

int t_length;

t_length = MultiByteToWideChar(1252, MB_PRECOMPOSED, p_native, strlen(p_native), t_utf16, strlen(p_native));

t_utf16[t_length] = 0;

return t_utf16;

}

INarrator::~INarrator(void)

{

// Nothing required.

}

// Description

// Initialises instance variables

WindowsSAPI5Narrator::WindowsSAPI5Narrator(void)

{

strcpy(m_strLastError,"");

bInited = FALSE;

isspeaking = FALSE;

m_bUseDefaultPitch = TRUE;

m_Npitch = 0;

}

// Description

// Deinitializes instance variables and frees dynamically allocated memory.

WindowsSAPI5Narrator::~WindowsSAPI5Narrator(void)

{

Finalize();

}

// Returns

// true if Initialisation sucessful, false otherwise.

// Description

// Initialises COM and creates an instance of the voice COM class.

// Registers interest for all the events associated with the voice COM class.

bool WindowsSAPI5Narrator::Initialize(void)

{

if(IsInited())

return true;

Error(L"");

//sapi init

HRESULT hr = S_OK;

if( SUCCEEDED( hr ) )

{

hr = CoCreateInstance(CLSID_SpVoice, NULL, CLSCTX_ALL, IID_ISpVoice, (void **)&m_cpVoice);

}

else

{

return false;

}

// We're interested in all TTS events

if( SUCCEEDED( hr ) )

{

hr = m_cpVoice->SetInterest( SPFEI_ALL_TTS_EVENTS, SPFEI_ALL_TTS_EVENTS );

}

else

{

return false;

}

bInited = true;

return true;

}

// Returns

// true if finalizing suceeded, false otherwise

// Description

// Finalises the Narrator class by releasing the voice COM object and

// removing the notification mechanism setup in Initialize.

bool WindowsSAPI5Narrator::Finalize(void)

{

if(!IsInited())

return false;

// Release voice, if created

if(m_cpVoice)

{

if (m_cpVoice->SetNotifySink(NULL) != S_OK)

return false;

m_cpVoice.Release();

}

m_cpVoice = NULL;

bInited = false;

return true;

}

// Returns

// true if starting suceeded, false otherwise.

// Description

// Starts speaking the string p_string. If the pitch has been set to a non default

// value using SetPitch then an xml fragment is generated specifying the speach pitch, and

// containing the text to speak. The string to speak is converted to WCHAR before being rendered.

// The voice COM object created in Initialize() is used to speak the text. Sets the instance

// variable isspeaking to true if succesful.

bool WindowsSAPI5Narrator::Start(const char* p_string, bool p_wants_utf8)

{

if (!bInited)

return false;

HRESULT hr;

// MW-2014-08-06: [[ UnicodeSpeech ]] Reworked to convert p_string based on wants_utf8.

WCHAR *t_wchar_string;

if(!m_bUseDefaultPitch)

{

char *pstrTotal = new char[ strlen(p_string) + PITCH_XML_STRING + 10 ];

char strXMLtag[PITCH_XML_STRING];

sprintf( strXMLtag, "<pitch absmiddle = \'%d\'/> ", m_Npitch);

strcpy( pstrTotal, strXMLtag);

strcat( pstrTotal, p_string);

if (p_wants_utf8)

t_wchar_string = utf8tow(pstrTotal);

else

t_wchar_string = nativetow(pstrTotal);

delete pstrTotal;

}

else

{

if (p_wants_utf8)

t_wchar_string = utf8tow(p_string);

else

t_wchar_string = nativetow(p_string);

}

if (t_wchar_string != NULL)

hr = m_cpVoice->Speak(t_wchar_string, SPF_ASYNC | SPF_IS_XML, NULL );

else

hr = -1;

delete t_wchar_string;

if( SUCCEEDED( hr ) )

{

isspeaking = true;

return true;

}

else

{

return false;

}

}

// Test, speaks the string p_string into the wav file p_file.

bool WindowsSAPI5Narrator::SpeakToFile(const char* p_string, const char* p_file)

{

if (!bInited)

return false;

HRESULT hr;

USES_CONVERSION;

WCHAR* szWTextString;

if(!m_bUseDefaultPitch)

{

char *pstrTotal = new char[ strlen(p_string) + PITCH_XML_STRING + 10 ];

char strXMLtag[PITCH_XML_STRING];

sprintf( strXMLtag, "<pitch absmiddle = \'%d\'/> ", m_Npitch);

strcpy( pstrTotal, strXMLtag);

strcat( pstrTotal, p_string);

szWTextString = new WCHAR[ strlen(pstrTotal) + 20 ];

wcscpy( szWTextString, A2W(pstrTotal));

}

else

{

int t_length;

t_length = strlen(p_string);

szWTextString = new WCHAR[ t_length + 1 ];

wcscpy( szWTextString, A2W(p_string));

}

bool t_success;

t_success = true;

// Setup the wave format structure - 44kHz, 16bit, stereo

WAVEFORMATEX t_waveformat;

t_waveformat.wFormatTag = WAVE_FORMAT_PCM;

t_waveformat.nChannels = 2;

t_waveformat.nBlockAlign = 4;

t_waveformat.nSamplesPerSec = 44100;

t_waveformat.wBitsPerSample = 16;

t_waveformat.nAvgBytesPerSec = t_waveformat.nSamplesPerSec * t_waveformat.nBlockAlign;

t_waveformat.cbSize = 0;

// unicode string containg file path

WCHAR* t_file;

t_file = NULL;

// pointer to the stream object returned

CComPtr<ISpStream> t_cstream;

if (t_success)

{

// Convert the path to native format and then to unicode

char *t_native_path;

t_native_path = path_to_native(p_file);

if (t_native_path == NULL)

{

t_success = false;

}

if (t_success)

{

t_file = new WCHAR[strlen(t_native_path) + 1];

wcscpy(t_file, A2W(t_native_path));

// Bind the output to the stream

hr = CoCreateInstance(CLSID_SpStream, NULL, CLSCTX_ALL, __uuidof(t_cstream), (void**)&t_cstream);

t_success = SUCCEEDED(hr);

if (t_success)

{

hr = t_cstream->BindToFile(t_file, SPFM_CREATE_ALWAYS, &SPDFID_WaveFormatEx, &t_waveformat, SPFEI_ALL_EVENTS);

t_success = SUCCEEDED(hr);

if (!t_success)

{

t_cstream->Release();

t_cstream = NULL;

}

}

}

}

// Set the voice object's output to the stream

if (t_success)

{

hr = m_cpVoice->SetOutput(t_cstream, TRUE);

if (!SUCCEEDED(hr))

{

t_success = false;

}

}

// Render the text into the file

if (t_success)

{

hr = m_cpVoice->Speak(szWTextString, SPF_IS_XML, NULL );

if (!SUCCEEDED(hr))

t_success = false;

}

// Clean up

if (t_success)

{

delete[] szWTextString;

delete[] t_file;

t_cstream->Close();

m_cpVoice -> SetOutput(NULL, true);

}

if(t_success)

{

isspeaking = true;

return true;

}

else

{

return false;

}

}

// Returns

// true if stop speaking suceeded, false otherwise.

// Description

// Stops any speech that is currently rendering by calling the Speak method

// of the voice COM class with NULL as the string to speak. Sets the instance

// variable isspeaking to false if successful.

bool WindowsSAPI5Narrator::Stop(void)

{

if (!bInited)

return false;

// Stop current rendering with a PURGEBEFORESPEAK...

HRESULT hr = m_cpVoice->Speak( NULL, SPF_PURGEBEFORESPEAK, 0 );

if( SUCCEEDED( hr ) )

{

isspeaking = false;

return true;

}

else

{

return false;

}

}

// Returns

// true if speech is currently being rendered, false otherwise.

// Description

// Uses the GetStatus method of the voice COM object to determine whether

// speech rendering is in progress. Updates the isspeaking instance variable

// to reflect the current status of the COM object.

bool WindowsSAPI5Narrator::Busy(void)

{

if (!bInited)

return false;

SPVOICESTATUS spstat;

HRESULT hr = m_cpVoice->GetStatus(&spstat,NULL);

if( SUCCEEDED( hr ) )

{

isspeaking = (spstat.dwRunningState & SPRS_IS_SPEAKING) != 0;

return isspeaking;

}

else

{

return false;

}

}

// Parameters

// p_volume : The value to set the volume to.

// Returns

// true if setting the volume suceeded, false otherwise.

// Description

// Uses the SetVolume method of the voice COM object to attempt to

// set the volume to p_volume.

bool WindowsSAPI5Narrator::SetVolume(int p_volume)

{

if (!bInited)

return false;

HRESULT hr = m_cpVoice->SetVolume(p_volume);

if ( SUCCEEDED( hr ) )

return true;

else

{

return false;

}

}

// Parameters

// p_volume : The volume is placed into this variable

// Returns

// true if getting the volume succeeded, false otherwise.

// Description

// Uses the GetVolume method of the voice COM object to get the volume.

// Sets the value of p_volume to the volume if successful, otherwise leaves

// p_volume with the value it originally had.

bool WindowsSAPI5Narrator::GetVolume(int& p_volume)

{

if (!bInited)

return false;

USHORT t_volume;

HRESULT hr = m_cpVoice->GetVolume(&t_volume);

if ( SUCCEEDED( hr ) )

{

p_volume = (int)t_volume;

return true;

}

else

{

return false;

}

}

bool _GetCategoryFromId(const WCHAR *p_id, ISpObjectTokenCategory** r_category)

{

bool t_success = true;

CComPtr<ISpObjectTokenCategory> t_token_category = NULL;

t_success = SUCCEEDED(t_token_category.CoCreateInstance(CLSID_SpObjectTokenCategory));

if (t_success)

t_success = SUCCEEDED(t_token_category->SetId(p_id, FALSE));

if (t_success)

*r_category = t_token_category.Detach();

return t_success;

}

bool _EnumTokens(const WCHAR *p_category_id, const WCHAR *p_req_attribs, const WCHAR *p_opt_attribs, IEnumSpObjectTokens **r_enum)

{

bool t_success = true;

CComPtr<ISpObjectTokenCategory> t_category;

t_success = _GetCategoryFromId(p_category_id, &t_category);

if (t_success)

t_success = SUCCEEDED(t_category->EnumTokens(p_req_attribs, p_opt_attribs, r_enum));

return t_success;

}

bool _GetDescription(ISpObjectToken *p_token, WCHAR **r_description)

{

typedef HRESULT (WINAPI *_RegLoadMUIStringWPtr)(HKEY, LPCWSTR, LPWSTR, DWORD, LPDWORD, DWORD, LPCWSTR);

static HMODULE s_advapi32 = NULL;

static _RegLoadMUIStringWPtr s_RegLoadMUIStringW = NULL;

if (s_advapi32 == NULL)

{

// Attempt to load function RegLoadMUIStringW which should be available if we're running on Vista & later

s_advapi32 = LoadLibraryA("advapi32.dll");

if(s_advapi32 != NULL)

s_RegLoadMUIStringW = (_RegLoadMUIStringWPtr)GetProcAddress(s_advapi32, "RegLoadMUIStringW");

}

bool t_success = true;

if (r_description == NULL)

{

return false;

}

WCHAR *t_description = NULL;

if(s_RegLoadMUIStringW != NULL)

{

WCHAR* t_path = NULL;

WCHAR* t_reg_key = NULL;

HKEY t_hkey = NULL;

t_success = SUCCEEDED(p_token->GetId(&t_reg_key));

if (t_success)

{

// Split the path & base of the registry key - path is invalid if there is no separator

t_path = wcschr(t_reg_key, L'\\');

t_success = t_path != NULL;

}

if(t_success)

{

*t_path++ = L'\0';

if (wcscmp(t_reg_key, L"HKEY_LOCAL_MACHINE") == 0)

t_success = ERROR_SUCCESS == RegOpenKeyExW(HKEY_LOCAL_MACHINE, t_path, 0, KEY_QUERY_VALUE, &t_hkey);

else if (wcscmp(t_reg_key, L"HKEY_CURRENT_USER") == 0)

t_success = ERROR_SUCCESS == RegOpenKeyExW(HKEY_CURRENT_USER, t_path, 0, KEY_QUERY_VALUE, &t_hkey);

else

t_success = false;

DWORD t_size = 0;

if (t_success)

t_success = ERROR_MORE_DATA == s_RegLoadMUIStringW(t_hkey, L"Description", NULL, 0, &t_size, 0, NULL);

if (t_success)

t_success = NULL != (t_description = (WCHAR*) CoTaskMemAlloc(t_size));

if (t_success)

t_success = ERROR_SUCCESS == s_RegLoadMUIStringW(t_hkey, L"Description", t_description, t_size, NULL, 0, NULL);

}

// Cleanup

if(t_hkey)

RegCloseKey(t_hkey);

if(t_reg_key)

CoTaskMemFree(t_reg_key);

if (!t_success && t_description != NULL)

{

CoTaskMemFree(t_description);

t_description = NULL;

}

}

// fetch from the registry - if all else fails, fallback to the default attribute

if (s_RegLoadMUIStringW == NULL || !t_success)

{

WCHAR t_lang_id[10];

if (_ultow_s(GetUserDefaultUILanguage(), t_lang_id, 9, 16))

{

t_lang_id[0] = L'0';

t_lang_id[1] = 0;

}

HRESULT hr;

hr = p_token->GetStringValue(t_lang_id, &t_description);

if (hr == SPERR_NOT_FOUND)

{

hr = p_token->GetStringValue(NULL, &t_description);

}

t_success = SUCCEEDED(hr);

}

if (t_success)

*r_description = t_description;

return t_success;

}

// Helper function, frees tokens allocated by ListVoices.

void ManageCleanup( WCHAR** ppszTokenIds, CSpDynamicString* ppcDesciptionString, ULONG ulNumTokens)

{

ULONG ulIndex;

// Free all allocated token ids

if ( ppszTokenIds )

{

for ( ulIndex = 0; ulIndex < ulNumTokens; ulIndex++ )

{

if ( NULL != ppszTokenIds[ulIndex] )

{

CoTaskMemFree( ppszTokenIds[ulIndex] );

}

}

delete [] ppszTokenIds;

}

if ( ppcDesciptionString )

{

delete [] ppcDesciptionString;

}

}

// Returns

// true if listing voices suceeeded, false otherwise.

// Description

// Iterates through the appropriate available voices and calls p_callback to return each voice.

// If p_gender is kNarratorGenderMale then only male voices are returned, if p_gender is kNarratorGenderFemale

// then only female voices are returned, otherwise all voices are returned. Can return false either if unable

// to get a token enumerator for the voices or if out of memory when listing voices.

bool WindowsSAPI5Narrator::ListVoices(NarratorGender p_gender, NarratorListVoicesCallback p_callback, void* p_context)

{

bool t_success = true;

USES_CONVERSION;

if(!bInited)

return false;

ISpObjectToken *pToken = NULL; // Token interface pointer

CComPtr<IEnumSpObjectTokens> cpEnum;// Pointer to token enumerator

ULONG ulIndex = 0;

CSpDynamicString* ppcDesciptionString;

ppcDesciptionString = NULL;

WCHAR** ppszTokenIds;

ppszTokenIds = NULL;

ULONG ulNumTokens;

WCHAR *szRequiredAttributes = NULL;

// Set the required attributes field for the enum if we have special needs

// based on our LPARAM in

if (p_gender == kNarratorGenderMale)

{

szRequiredAttributes = L"Gender=Male";

}

else if (p_gender == kNarratorGenderFemale)

{

szRequiredAttributes = L"Gender=Female";

}

else if (p_gender == kNarratorGenderNeuter)

{

szRequiredAttributes = L"Gender=Neutral";

}

// Get a token enumerator for tts voices available

t_success = _EnumTokens(SPCAT_VOICES, szRequiredAttributes, NULL, &cpEnum);

if (t_success)

{

t_success = SUCCEEDED(cpEnum->GetCount( &ulNumTokens ));

if ( t_success && 0 != ulNumTokens )

{

ppcDesciptionString = new CSpDynamicString [ulNumTokens];

if ( NULL == ppcDesciptionString )

{

/* TODO - CLEANUP */

return false;

}

ppszTokenIds = new WCHAR* [ulNumTokens];

if ( NULL == ppszTokenIds )

{

/* TODO - CLEANUP */

return false;

}

ZeroMemory( ppszTokenIds, ulNumTokens*sizeof( WCHAR* ) );

// Get the next token in the enumeration

// State is maintained in the enumerator

while (t_success && cpEnum->Next(1, &pToken, NULL) == S_OK)

{

// Get a string which describes the token, in our case, the voice name

t_success = SUCCEEDED(_GetDescription( pToken, &ppcDesciptionString[ulIndex] ));

// Get the token id, for a low overhead way to retrieve the token later

// without holding on to the object itself

if (t_success)

t_success = SUCCEEDED(pToken->GetId( &ppszTokenIds[ulIndex] ));

ulIndex++;

// Release the token itself

pToken->Release();

pToken = NULL;

}

}

// if we've failed to properly initialize, then we should completely shut-down

if ( !t_success )

{

if ( pToken )

{

pToken->Release();

}

ppszTokenIds = NULL;

ppcDesciptionString = NULL;

ulNumTokens = 0;

}

for ( ulIndex = 0; ulIndex < ulNumTokens; ulIndex++ )

{

const char* string = W2A(ppcDesciptionString[ulIndex]);

p_callback(p_context, p_gender, (const char *)string);

}

ManageCleanup( ppszTokenIds, ppcDesciptionString, ulNumTokens );

}

return t_success;

}

// Parameters

// p_voice : The string to set the voice to

// Returns

// true if setting the voice succeeded, false otherwise.

// Description

// Lists the available voice tokens and iterates through them comparing each to p_voice.

// If a token is found that matches p_voice then the voice is set to that token. If p_voice

// doesn't match any of the voices available then returns false, and if setting the voice to

// the matching token fails then also returns false.

bool WindowsSAPI5Narrator::SetVoice(const char* p_voice)

{

USES_CONVERSION;

if(!bInited)

return false;

bool t_success = true;

const WCHAR *t_wvoice = A2W(p_voice);

CComPtr<IEnumSpObjectTokens> cpEnum;// Pointer to token enumerator

ULONG ulIndex = 0, ulCurTokenIndex = 0, ulNumTokens = 0;

// Find out which token corresponds to our voice which is currently in use

ISpObjectToken *pToken = NULL; // Token interface pointer

t_success = _EnumTokens(SPCAT_VOICES, NULL, NULL, &cpEnum);

if (t_success)

t_success = SUCCEEDED(cpEnum->GetCount( &ulNumTokens ));

if (t_success)

t_success = ulNumTokens > 0;

while (t_success && cpEnum->Next(1, &pToken, NULL) == S_OK)

{

CSpDynamicString ppcDesciptionString;

// Get a string which describes the token, in our case, the voice name

t_success = _GetDescription( pToken, &ppcDesciptionString);

// Release the token itself

pToken->Release();

pToken = NULL;

if (t_success)

{

if(wcscmp(ppcDesciptionString, t_wvoice) == 0)

{

ulCurTokenIndex = ulIndex;

break;

}

ulIndex++;

}

}

if (t_success)

t_success = ulIndex < ulNumTokens;

if (t_success)

{

cpEnum->Item( ulCurTokenIndex, &pToken );

t_success = SUCCEEDED(m_cpVoice->SetVoice( pToken ));

pToken->Release();

}

return t_success;

}

// Parameters

// p_speed : The value to set the speed to

// Returns

// true if setting the speed succeeded, false otherwise.

// Description

// Attempts to set the speed to p_speed using the voice COM object SetRate method.

bool WindowsSAPI5Narrator::SetSpeed(int p_speed)

{

if (!bInited)

return false;

int newint = (int)(-10 + ((double)((p_speed - 30)/360.0) * 20));

HRESULT hr = hr = m_cpVoice->SetRate(newint);

if ( SUCCEEDED( hr ) )

return true;

else

{

return false;

}

}

// Parameters

// p_speed : The speed is placed into this variable

// Returns

// true if getting the speed suceeded, false otherwise.

// Description

// Gets the speed using the GetRate method of the voice COM object.

// If getting the speed suceeeded then the resulting value will be placed into

// p_speed, otherwise the value of p_speed will be unchanged.

bool WindowsSAPI5Narrator::GetSpeed(int& p_speed)

{

if (!bInited)

return false;

long lRate;

HRESULT hr = m_cpVoice->GetRate( &lRate );

// initialize sliders and edit boxes with default rate

lRate = 30 + (18 * (lRate - -10));

if ( SUCCEEDED( hr ) )

{

p_speed = (int)lRate;

return true;

}

else

{

return false;

}

}

// Parameters

// p_pitch : the value to set the pitch to.

// Returns

// true if setting the pitch suceeded, false otherwise

// Description

// Translates p_pitch to a value in the correct range, then puts

// the resulting number into the instance variable m_Npitch. If the pitch

// is non-default, sets the instance variable m_bUseDefaultPitch to true.

// The new pitch will not take effect take the next call to Start.

bool WindowsSAPI5Narrator::SetPitch(int p_pitch)

{

m_Npitch = p_pitch;

m_Npitch = (int)(-10 + ((double)((p_pitch - 30)/360.0) * 20));

if(m_Npitch == 0)

m_bUseDefaultPitch = true;

else

m_bUseDefaultPitch = false;

return true;

}

// Parameters

// p_pitch : variable to place value of the pitch into

// Returns

// always returns true

// Description

// Returns the current pitch by placing it into p_pitch.

bool WindowsSAPI5Narrator::GetPitch(int& p_pitch)

{

p_pitch = (int)((((double)(m_Npitch + 10) / 20) * 360.0) + 30);

return true;

}