一千萬個為什麽

搜索

在Windows上使用Espeak SAPI/dll?

問題:我正在嘗試使用espeak文本到語音引擎。 因此,我讓它在Linux上運行得很好(下面的代碼)。 現在我想將這個基本程序移植到Windows上,但這幾乎是不可能的......

部分問題是windows dll只允許AUDIO_OUTPUT_SYNCHRONOUS,這意味著它需要回調,但我無法弄清楚如何從回調中播放音頻......首先它崩潰了,然後我意識到,我需要一個回調函數,現在我在回調函數中獲取數據,但我不知道如何播放它...因為它既不是一個wav文件,也不像Linux那樣自動播放。

sourceforge站點相當無用,因為它基本上說使用SAPI版本,但是沒有關於如何使用sapi espeak dll的示例...

無論如何,這是我的代碼,任何人都可以幫忙嗎?

#ifdef __cplusplus
#include 
#include 
#include 
#else
#include 
#include 
#include 
#endif

#include 
#include 

//#include "speak_lib.h"
#include "espeak/speak_lib.h"

// libespeak-dev: /usr/include/espeak/speak_lib.h
// apt-get install libespeak-dev
// apt-get install libportaudio-dev

// g++ -o mine mine.cpp -lespeak
// g++ -o mine mine.cpp -I/usr/include/espeak/ -lespeak
// gcc -o mine mine.cpp -I/usr/include/espeak/ -lespeak


char voicename[40];
int samplerate;
int quiet = 0;
static char genders[4] = {' ','M','F',' '};

//const char *data_path = "/usr/share/";  ///usr/share/espeak-data/
const char *data_path = NULL;  //use default path for espeak-data


int strrcmp(const char *s, const char *sub)
{
int slen = strlen(s);
int sublen = strlen(sub);
return memcmp(s + slen - sublen, sub, sublen);
}


char * strrcpy(char *dest, const char *source)
{
// Pre assertions
assert(dest != NULL);
assert(source != NULL);
assert(dest != source);

// tk: parentheses
while((*dest++ = *source++))
    ;
return(--dest);
}

const char* GetLanguageVoiceName(const char* pszShortSign)
{
#define LANGUAGE_LENGTH 30
static char szReturnValue[LANGUAGE_LENGTH] ;
memset(szReturnValue, 0, LANGUAGE_LENGTH);

for (int i = 0; pszShortSign[i] != '\0'; ++i)
    szReturnValue[i] = (char) tolower(pszShortSign[i]);

const espeak_VOICE **voices;
espeak_VOICE voice_select;
voices = espeak_ListVoices(NULL);

const espeak_VOICE *v;
for(int ix=0; (v = voices[ix]) != NULL; ix++)
{
    if( !strrcmp( v->languages, szReturnValue) )
    {
        strcpy(szReturnValue, v->name);
        return szReturnValue;
    }
}//End for

strcpy(szReturnValue, "default");
return szReturnValue;
}//End function getvoicename


void ListVoices()
{
const espeak_VOICE **voices;
espeak_VOICE voice_select;
voices = espeak_ListVoices(NULL);

const espeak_VOICE *v;
for(int ix=0; (v = voices[ix]) != NULL; ix++)
{
    printf("Shortsign: %s\n", v->languages);
    printf("age: %d\n", v->age);
    printf("gender: %c\n", genders[v->gender]);
    printf("name: %s\n", v->name);
    printf("\n\n");
}//End for
}//End function getvoicename


int main()
{
printf("Hello World!\n");
const char* szVersionInfo = espeak_Info(NULL);

printf("Espeak version: %s\n", szVersionInfo);
samplerate = espeak_Initialize(AUDIO_OUTPUT_PLAYBACK,0,data_path,0);

strcpy(voicename, "default");
// espeak --voices
strcpy(voicename, "german");
strcpy(voicename, GetLanguageVoiceName("DE"));

if(espeak_SetVoiceByName(voicename) != EE_OK)
{
    printf("Espeak setvoice error...\n");
}

static char word[200] = "Hello World" ;
strcpy(word, "TV-fäns aufgepasst, es ist 20 Uhr 15. Zeit für Rambo 3");
strcpy(word, "Unnamed Player wurde zum Opfer von GSG9");
int speed = 220;
int volume = 500;//volume in range 0-100    0=silence
int pitch = 50;// base pitch, range 0-100.  50=normal

// espeak.cpp 625
espeak_SetParameter(espeakRATE, speed, 0);
espeak_SetParameter(espeakVOLUME,volume,0);
espeak_SetParameter(espeakPITCH,pitch,0);
// espeakRANGE:   pitch range, range 0-100. 0-monotone, 50=normal
// espeakPUNCTUATION:  which punctuation characters to announce:
   //value in espeak_PUNCT_TYPE (none, all, some), 
espeak_VOICE *voice_spec = espeak_GetCurrentVoice();
voice_spec->gender=2;//0=none 1=male, 2=female,
//voice_spec->age = age;

espeak_SetVoiceByProperties(voice_spec);


espeak_Synth( (char*) word, strlen(word)+1, 0, POS_CHARACTER, 0, espeakCHARS_AUTO, NULL, NULL);
espeak_Synchronize();

strcpy(voicename, GetLanguageVoiceName("EN"));
espeak_SetVoiceByName(voicename);
strcpy(word, "Geany was fragged by GSG9 Googlebot");
strcpy(word, "Googlebot");

espeak_Synth( (char*) word, strlen(word)+1, 0, POS_CHARACTER, 0, espeakCHARS_AUTO, NULL, NULL);
espeak_Synchronize();


espeak_Terminate();
printf("Espeak terminated\n");
return EXIT_SUCCESS; 
}

/*
if(espeak_SetVoiceByName(voicename) != EE_OK)
{
    memset(&voice_select,0,sizeof(voice_select));
    voice_select.languages = voicename;
    if(espeak_SetVoiceByProperties(&voice_select) != EE_OK)
    {
        fprintf(stderr,"%svoice '%s'\n",err_load,voicename);
        exit(2);
    }
}
*/

以上代碼適用於Linux。 下面的代碼大約是我在Vista x64(32位emu)上得到的:

#ifdef __cplusplus
#include 
#include 
#include 
#else
#include 
#include 
#include 
#endif

#include 
#include 

#include "speak_lib.h"
//#include "espeak/speak_lib.h"

// libespeak-dev: /usr/include/espeak/speak_lib.h
// apt-get install libespeak-dev
// apt-get install libportaudio-dev

// g++ -o mine mine.cpp -lespeak
// g++ -o mine mine.cpp -I/usr/include/espeak/ -lespeak
// gcc -o mine mine.cpp -I/usr/include/espeak/ -lespeak


char voicename[40];
int iSampleRate;
int quiet = 0;
static char genders[4] = {' ','M','F',' '};

//const char *data_path = "/usr/share/";  ///usr/share/espeak-data/
//const char *data_path = NULL;  //use default path for espeak-data
const char *data_path = "C:\\Users\\Username\\Desktop\\espeak-1.43-source\\espeak-1.43-source\\";


int strrcmp(const char *s, const char *sub)
{
int slen = strlen(s);
int sublen = strlen(sub);
return memcmp(s + slen - sublen, sub, sublen);
}


char * strrcpy(char *dest, const char *source)
{
// Pre assertions
assert(dest != NULL);
assert(source != NULL);
assert(dest != source);

// tk: parentheses
while((*dest++ = *source++))
    ;
return(--dest);
}

const char* GetLanguageVoiceName(const char* pszShortSign)
{
#define LANGUAGE_LENGTH 30
static char szReturnValue[LANGUAGE_LENGTH] ;
memset(szReturnValue, 0, LANGUAGE_LENGTH);

for (int i = 0; pszShortSign[i] != '\0'; ++i)
    szReturnValue[i] = (char) tolower(pszShortSign[i]);

const espeak_VOICE **voices;
espeak_VOICE voice_select;
voices = espeak_ListVoices(NULL);

const espeak_VOICE *v;
for(int ix=0; (v = voices[ix]) != NULL; ix++)
{
    if( !strrcmp( v->languages, szReturnValue) )
    {
        strcpy(szReturnValue, v->name);
        return szReturnValue;
    }
}//End for

strcpy(szReturnValue, "default");
return szReturnValue;
}//End function getvoicename


void ListVoices()
{
const espeak_VOICE **voices;
espeak_VOICE voice_select;
voices = espeak_ListVoices(NULL);

const espeak_VOICE *v;
for(int ix=0; (v = voices[ix]) != NULL; ix++)
{
    printf("Shortsign: %s\n", v->languages);
    printf("age: %d\n", v->age);
    printf("gender: %c\n", genders[v->gender]);
    printf("name: %s\n", v->name);
    printf("\n\n");
}//End for
}//End function getvoicename


/* Callback from espeak.  Directly speaks using AudioTrack. */
#define LOGI(x) printf("%s\n", x)
static int AndroidEspeakDirectSpeechCallback(short *wav, int numsamples, espeak_EVENT *events) 
{
    char buf[100];
    sprintf(buf, "AndroidEspeakDirectSpeechCallback: %d samples", numsamples);
    LOGI(buf);

    if (wav == NULL) 
{
        LOGI("Null: speech has completed");
    }

    if (numsamples > 0)
{
        //audout->write(wav, sizeof(short) * numsamples);
        sprintf(buf, "AudioTrack wrote: %d bytes", sizeof(short) * numsamples);
        LOGI(buf);
    }

    return 0; //continue synthesis (1 is to abort)
}


static int AndroidEspeakSynthToFileCallback(short *wav, int numsamples,espeak_EVENT *events) 
{
    char buf[100];
    sprintf(buf, "AndroidEspeakSynthToFileCallback: %d samples", numsamples);
    LOGI(buf);

    if (wav == NULL) 
{
        LOGI("Null: speech has completed");
    }

   //The user data should contain the file pointer of the file to write to
    //void* user_data = events->user_data;
FILE* user_data = fopen ( "myfile1.wav" , "ab" );

    FILE* fp = static_cast(user_data);

   //Write all of the samples
    fwrite(wav, sizeof(short), numsamples, fp);
    return 0; //continue synthesis (1 is to abort)
}



int main()
{
printf("Hello World!\n");
const char* szVersionInfo = espeak_Info(NULL);

printf("Espeak version: %s\n", szVersionInfo);

iSampleRate = espeak_Initialize(AUDIO_OUTPUT_SYNCHRONOUS, 4096, data_path, 0);
if (iSampleRate <= 0) 
{
    printf("Unable to initialize espeak");
    return EXIT_FAILURE;
}

//samplerate = espeak_Initialize(AUDIO_OUTPUT_PLAYBACK,0,data_path,0);

//ListVoices();

strcpy(voicename, "default");
// espeak --voices
//strcpy(voicename, "german");
//strcpy(voicename, GetLanguageVoiceName("DE"));

if(espeak_SetVoiceByName(voicename) != EE_OK)
{
    printf("Espeak setvoice error...\n");
}

static char word[200] = "Hello World" ;
strcpy(word, "TV-fäns aufgepasst, es ist 20 Uhr 15. Zeit für Rambo 3");
strcpy(word, "Unnamed Player wurde zum Opfer von GSG9");
int speed = 220;
int volume = 500;//volume in range 0-100    0=silence
int pitch = 50;// base pitch, range 0-100.  50=normal


// espeak.cpp 625
espeak_SetParameter(espeakRATE, speed, 0);
espeak_SetParameter(espeakVOLUME,volume,0);
espeak_SetParameter(espeakPITCH,pitch,0);
// espeakRANGE:   pitch range, range 0-100. 0-monotone, 50=normal
// espeakPUNCTUATION:  which punctuation characters to announce:
   //value in espeak_PUNCT_TYPE (none, all, some), 
//espeak_VOICE *voice_spec = espeak_GetCurrentVoice();
//voice_spec->gender=2;//0=none 1=male, 2=female,
//voice_spec->age = age;

//espeak_SetVoiceByProperties(voice_spec);

//espeak_SetSynthCallback(AndroidEspeakDirectSpeechCallback);
espeak_SetSynthCallback(AndroidEspeakSynthToFileCallback);

unsigned int unique_identifier;
espeak_ERROR err = espeak_Synth( (char*) word, strlen(word)+1, 0, POS_CHARACTER, 0, espeakCHARS_AUTO, &unique_identifier, NULL);

err = espeak_Synchronize();



/*
strcpy(voicename, GetLanguageVoiceName("EN"));
espeak_SetVoiceByName(voicename);
strcpy(word, "Geany was fragged by GSG9 Googlebot");
strcpy(word, "Googlebot");

espeak_Synth( (char*) word, strlen(word)+1, 0, POS_CHARACTER, 0, espeakCHARS_AUTO, NULL, NULL);
espeak_Synchronize();
*/

// espeak_Cancel();
espeak_Terminate();
printf("Espeak terminated\n");
system("pause");
return EXIT_SUCCESS; 
}

最佳答案

Several changes in source code are needed to make the windows library have the same functionality as the one on Linux. I listed the changes here. The ready to use binary is also available.

所有補丁和描述也被發送到espeak維護者(公開,通過郵件列表和補丁跟蹤器),所以將來可能會直接提供。

轉載註明原文: 在Windows上使用Espeak SAPI/dll?