以上節tts語音輸出爲例
node
下載sdk連接:http://www.xfyun.cn/sdk/dispatcher服務器
一、下載SDK,解壓;session
二、在ROS工做空間下建立一個Package;
app
catkin_create_pkg tts_voice roscpp rospy std_msgs
三、將SDK文件夾中的頭文件,即下載的SDK文件夾中include文件夾下的 .h文件拷貝至tts_voice/include/下;函數
四、將SDK文件夾中sample/tts_sample中的xf_tts.cpp文件拷貝至tts_voice/src/下;ui
五、將SDK文件夾中libs/x64/libmsc.so文件拷貝至/user/libs/下;編碼
六、打開CMakelist,spa
1)修改include_directories爲:即包括拷貝的頭文件所在的include目錄指針
include_directories(
${catkin_INCLUDE_DIRS}
include
)code
2)添加:
add_executable(tts_voice_node src/xf_tts.cpp)
target_link_libraries(tts_voice_node
${catkin_LIBRARIES} -lmsc -lrt -ldl -lpthread
)
# 生成可執行文件
# 添加連接庫,其中動態連接庫路徑按實際修改
CmakeLists.txt
cmake_minimum_required(VERSION 2.8.3) project(tts_voice) ## Find catkin macros and libraries ## if COMPONENTS list like find_package(catkin REQUIRED COMPONENTS xyz) ## is used, also find other catkin packages find_package(catkin REQUIRED COMPONENTS roscpp rospy std_msgs ) catkin_package( # INCLUDE_DIRS include # LIBRARIES tts_voice # CATKIN_DEPENDS roscpp rospy std_msgs # DEPENDS system_lib ) ########### ## Build ## ########### ## Specify additional locations of header files ## Your package locations should be listed before other locations # include_directories(include) include_directories( ${catkin_INCLUDE_DIRS} include ) ## Declare a C++ executable add_executable(tts_voice_node src/xf_tts.cpp) ## Specify libraries to link a library or executable target against target_link_libraries(tts_voice_node ${catkin_LIBRARIES} -lmsc -lrt -ldl -lpthread )
xf_tts.cpp源文件
/* * 語音合成(Text To Speech,TTS)技術可以自動將任意文字實時轉換爲連續的 * 天然語音,是一種可以在任什麼時候間、任何地點,向任何人提供語音信息服務的 * 高效便捷手段,很是符合信息時代海量數據、動態更新和個性化查詢的需求。 */ #include <ros/ros.h> #include "std_msgs/String.h" #include <stdio.h> #include <string.h> #include <stdlib.h> #include <unistd.h> #include "qtts.h" #include "msp_cmn.h" #include "msp_errors.h" const char* FileName = "/home/cam/Music/voice.wav"; const char* PlayPath = "play /home/cam/Music/voice.wav"; /* wav音頻頭部格式 */ typedef struct _wave_pcm_hdr { char riff[4]; // = "RIFF" int size_8; // = FileSize - 8 char wave[4]; // = "WAVE" char fmt[4]; // = "fmt " int fmt_size; // = 下一個結構體的大小 : 16 short int format_tag; // = PCM : 1 short int channels; // = 通道數 : 1 int samples_per_sec; // = 採樣率 : 8000 | 6000 | 11025 | 16000 int avg_bytes_per_sec; // = 每秒字節數 : samples_per_sec * bits_per_sample / 8 short int block_align; // = 每採樣點字節數 : wBitsPerSample / 8 short int bits_per_sample; // = 量化比特數: 8 | 16 char data[4]; // = "data"; int data_size; // = 純數據長度 : FileSize - 44 } wave_pcm_hdr; /* 默認wav音頻頭部數據 */ wave_pcm_hdr default_wav_hdr = { { 'R', 'I', 'F', 'F' }, 0, {'W', 'A', 'V', 'E'}, {'f', 'm', 't', ' '}, 16, 1, 1, 16000, 32000, 2, 16, {'d', 'a', 't', 'a'}, 0 }; /* 文本合成 */ int text_to_speech(const char* src_text, const char* des_path, const char* params) { int ret = -1; FILE* fp = NULL; const char* sessionID = NULL; unsigned int audio_len = 0; wave_pcm_hdr wav_hdr = default_wav_hdr; int synth_status = MSP_TTS_FLAG_STILL_HAVE_DATA; if (NULL == src_text || NULL == des_path) { printf("params is error!\n"); return ret; } fp = fopen(des_path, "wb"); if (NULL == fp) { printf("open %s error.\n", des_path); return ret; } /* 開始合成 */ sessionID = QTTSSessionBegin(params, &ret); if (MSP_SUCCESS != ret) { printf("QTTSSessionBegin failed, error code: %d.\n", ret); fclose(fp); return ret; } ret = QTTSTextPut(sessionID, src_text, (unsigned int)strlen(src_text), NULL); if (MSP_SUCCESS != ret) { printf("QTTSTextPut failed, error code: %d.\n",ret); QTTSSessionEnd(sessionID, "TextPutError"); fclose(fp); return ret; } printf("正在合成 ...\n"); fwrite(&wav_hdr, sizeof(wav_hdr) ,1, fp); //添加wav音頻頭,使用採樣率爲16000 while (1) { /* 獲取合成音頻 */ const void* data = QTTSAudioGet(sessionID, &audio_len, &synth_status, &ret); if (MSP_SUCCESS != ret) break; if (NULL != data) { fwrite(data, audio_len, 1, fp); wav_hdr.data_size += audio_len; //計算data_size大小 } if (MSP_TTS_FLAG_DATA_END == synth_status) break; printf(">"); usleep(15*1000); //防止頻繁佔用CPU }//合成狀態synth_status取值請參閱《訊飛語音雲API文檔》 printf("\n"); if (MSP_SUCCESS != ret) { printf("QTTSAudioGet failed, error code: %d.\n",ret); QTTSSessionEnd(sessionID, "AudioGetError"); fclose(fp); return ret; } /* 修正wav文件頭數據的大小 */ wav_hdr.size_8 += wav_hdr.data_size + (sizeof(wav_hdr) - 8); /* 將修正過的數據寫回文件頭部,音頻文件爲wav格式 */ fseek(fp, 4, 0); fwrite(&wav_hdr.size_8,sizeof(wav_hdr.size_8), 1, fp); //寫入size_8的值 fseek(fp, 40, 0); //將文件指針偏移到存儲data_size值的位置 fwrite(&wav_hdr.data_size,sizeof(wav_hdr.data_size), 1, fp); //寫入data_size的值 fclose(fp); fp = NULL; /* 合成完畢 */ ret = QTTSSessionEnd(sessionID, "Normal"); if (MSP_SUCCESS != ret) { printf("QTTSSessionEnd failed, error code: %d.\n",ret); } return ret; } /** * 將文本發送到訊飛服務器,獲取語音文件文件 **/ int makeTextToWav(const char* text, const char* filename) { int ret = MSP_SUCCESS; const char* login_params = "appid = 57f49f64, work_dir = .";//登陸參數,appid與msc庫綁定,請勿隨意改動 /* * rdn: 合成音頻數字發音方式 * volume: 合成音頻的音量 * pitch: 合成音頻的音調 * speed: 合成音頻對應的語速 * voice_name: 合成發音人 * sample_rate: 合成音頻採樣率 * text_encoding: 合成文本編碼格式 * * 詳細參數說明請參閱《訊飛語音雲MSC--API文檔》 */ const char* session_begin_params = "voice_name = xiaowanzi, text_encoding = utf8, sample_rate = 16000, speed = 50, volume = 50, pitch = 50, rdn = 0"; // const char* filename = "tts_sample.wav"; //合成的語音文件名稱 // const char* text = "我是小丸子,今年126歲了,個人車牌號是12365,今年是2016年12月20號"; //合成文本 /* 用戶登陸 */ ret = MSPLogin(NULL, NULL, login_params);//第一個參數是用戶名,第二個參數是密碼,第三個參數是登陸參數,用戶名和密碼可在http://open.voicecloud.cn註冊獲取 if (MSP_SUCCESS != ret) { printf("MSPLogin failed, error code: %d.\n", ret); goto exit ;//登陸失敗,退出登陸 } /* 文本合成 */ printf("開始合成 ...\n"); ret = text_to_speech(text, filename, session_begin_params); if (MSP_SUCCESS != ret) { printf("text_to_speech failed, error code: %d.\n", ret); } printf("合成完畢\n"); exit: MSPLogout(); //退出登陸 return 0; } /** * play the wav file **/ void PlayWav() { system(PlayPath); } /** *接受/voice/xf_tts_topic話題的字符串的回調函數 */ void TTSCallback(const std_msgs::String::ConstPtr &msg) { std::cout<<"get topic text: " << msg->data.c_str(); makeTextToWav(msg->data.c_str(), FileName); PlayWav(); } int main(int argc, char* argv[]) { const char* start = "寶寶啓動成功了"; makeTextToWav(start,FileName); PlayWav(); ros::init(argc, argv,"xf_tts_node"); ros::NodeHandle n; ros::Subscriber sub = n.subscribe("/voice/xf_tts_topic",3,TTSCallback); ros::spin(); return 0; }
到~/catkin_ws編譯
cam@cam-Aspire-V3-572G:~/catkin_ws$ catkin_make -j4
運行
¥ roscore
$ rosrun tts_voice tts_voice_node
啓動播放 "寶寶啓動成功了"
經過話題播放
rostopic pub /voice/xf_tts_topic std_msgs/String " 韓聯社預計,鑑於黃教堅持現有外交政策,「薩德」部署相關工做可能提速。本月早些時候,一些政府消息人士透露,
因爲總統朴槿惠遭彈劾,國內政 局不肯定性增長,韓國國防部決定加快「薩德」的相關部署工做,最快可能在明年5月完成部署。根據消息人士的說法,爲了能儘早完成部署工做,
韓國國防部甚至考慮省略掉環境影響評估的環節。"
使用時須要把
const char* login_params = "appid = 57f49f64, work_dir = .";//登陸參數,appid與msc庫綁定,請勿隨意改動
appid改成本身的賬號
若是play播放音頻出錯,是由於沒有安裝sox
執行安裝便可
sudo apt-get install sox