wchar_t SPCAT_VOICES_7[] = LR"(HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech\Voices)"; wchar_t SPCAT_VOICES_10[] = LR"(HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech_OneCore\Voices)"; std::optional> _Speak(std::wstring &Content, const wchar_t *token, int voiceid, int rate, int volume) { std::optional> ret = {}; [&]() { CO_INIT co; CHECK_FAILURE_NORET(co); CComPtr pVoice = NULL; CHECK_FAILURE_NORET(pVoice.CoCreateInstance(CLSID_SpVoice, NULL, CLSCTX_ALL)); CComPtr pSpEnumTokens = NULL; CHECK_FAILURE_NORET(SpEnumTokens(token, NULL, NULL, &pSpEnumTokens)); ULONG ulTokensNumber = 0; pSpEnumTokens->GetCount(&ulTokensNumber); ISpObjectToken *m_pISpObjectToken; pSpEnumTokens->Item(voiceid, &m_pISpObjectToken); pVoice->SetVoice(m_pISpObjectToken); pVoice->SetRate(rate); pVoice->SetVolume(volume); CComPtr cpWavStream; CComPtr cpOldStream; CSpStreamFormat originalFmt; CComPtr pMemStream; CHECK_FAILURE_NORET(pVoice->GetOutputStream(&cpOldStream)); originalFmt.AssignFormat(cpOldStream); CHECK_FAILURE_NORET(CreateStreamOnHGlobal(NULL, TRUE, &pMemStream)); CHECK_FAILURE_NORET(cpWavStream.CoCreateInstance(CLSID_SpStream)); CHECK_FAILURE_NORET(cpWavStream->SetBaseStream(pMemStream, SPDFID_WaveFormatEx, originalFmt.WaveFormatExPtr())); pVoice->SetOutput(cpWavStream, TRUE); pVoice->Speak(Content.c_str(), SPF_IS_XML, NULL); /*To verify that the data has been written correctly, uncomment this, you should hear the voice. cpVoice->SetOutput(NULL, FALSE); cpVoice->SpeakStream(cpStream, SPF_DEFAULT, NULL); */ // After SAPI writes the stream, the stream position is at the end, so we need to set it to the beginning. _LARGE_INTEGER a = {0}; CHECK_FAILURE_NORET(cpWavStream->Seek(a, STREAM_SEEK_SET, NULL)); // get the base istream from the ispstream CComPtr pIstream; cpWavStream->GetBaseStream(&pIstream); // calculate the size that is to be read STATSTG stats; pIstream->Stat(&stats, STATFLAG_NONAME); ULONG sSize = stats.cbSize.QuadPart; // size of the data to be read ULONG bytesRead; // this will tell the number of bytes that have been read std::vector datas; datas.resize(sSize + 46); auto pBuffer = datas.data(); // buffer to read the data // memcpy(pBuffer,&wavHeader,sizeof(WAV_HEADER)); int fsize = sSize + 46; int ptr = 0; memcpy(pBuffer, "RIFF", 4); ptr += 4; memcpy(pBuffer + ptr, &fsize, 4); ptr += 4; memcpy(pBuffer + ptr, "WAVEfmt ", 8); ptr += 8; memcpy(pBuffer + ptr, "\x12\x00\x00\x00", 4); ptr += 4; memcpy(pBuffer + ptr, originalFmt.WaveFormatExPtr(), sizeof(WAVEFORMATEX)); ptr += sizeof(WAVEFORMATEX); memcpy(pBuffer + ptr, "data", 4); ptr += 4; memcpy(pBuffer + ptr, &sSize, 4); ptr += 4; // read the data into the buffer pIstream->Read(pBuffer + ptr, sSize, &bytesRead); ret = std::move(datas); }(); return ret; } std::vector _List(const wchar_t *token) { std::vector ret; [&]() { CO_INIT co; CHECK_FAILURE_NORET(co); CComPtr pSpVoice = NULL; CComPtr pSpEnumTokens = NULL; CHECK_FAILURE_NORET(CoCreateInstance(CLSID_SpVoice, NULL, CLSCTX_ALL, IID_ISpVoice, (void **)&pSpVoice)); CHECK_FAILURE_NORET(SpEnumTokens(token, NULL, NULL, &pSpEnumTokens)); ULONG ulTokensNumber = 0; pSpEnumTokens->GetCount(&ulTokensNumber); CComPtr m_pISpObjectToken; for (ULONG i = 0; i < ulTokensNumber; i++) { pSpEnumTokens->Item(i, &m_pISpObjectToken); CComHeapPtr pszVoiceId; CComHeapPtr pszVoiceName; if (SUCCEEDED(m_pISpObjectToken->GetId(&pszVoiceId))) { if (SUCCEEDED(m_pISpObjectToken->GetStringValue(NULL, &pszVoiceName))) { ret.emplace_back(pszVoiceName); } } } }(); return ret; }