876 lines
18 KiB
Markdown
876 lines
18 KiB
Markdown
|
## linux-sdk 使用说明
|
|||
|
|
|
|||
|
|
### 接口说明
|
|||
|
|
|
|||
|
|
**sdk语音相关的接口**
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sdk 初始化函数，主要包括sdk参数初始化和授权两个过程，这个函数是阻塞的，
|
|||
|
|
授权的超时时间为 15 秒。
|
|||
|
|
|
|||
|
|
struct dds_client *dds_client_init (const char *config_json);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
@ config_json: 配置选项，json 格式的字符串；
|
|||
|
|
|
|||
|
|
返回值:
|
|||
|
|
|
|||
|
|
错误情况下返回NULL, 否则返回 struct dds_client 实例指针；
|
|||
|
|
```
|
|||
|
|
```
|
|||
|
|
运行 sdk 函数，调用此函数之后就可以语音交互了。
|
|||
|
|
|
|||
|
|
int dds_client_start(struct dds_client *ds, ddsLintener cb,
|
|||
|
|
void *user);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
@ ds: 由 dds_client_init 返回的 struct dds_client 实例指针；
|
|||
|
|
@ cb: 监听 sdk 事件的回调函数；
|
|||
|
|
@ user: 用户参数；
|
|||
|
|
|
|||
|
|
返回值:
|
|||
|
|
|
|||
|
|
出错返回 -1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
释放 sdk 实例的函数:
|
|||
|
|
|
|||
|
|
void dds_client_release(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
@ ds: 由 dds_client_init 返回的 struct dds_client 实例指针；
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
**下面的接口必须在 `dds_client_init ` 和 `dds_client_start` 正确返回之后才能正确执行。**
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
向 sdk 内部发送消息:
|
|||
|
|
|
|||
|
|
int dds_client_publish(struct dds_client *ds, int ev,
|
|||
|
|
const char *data);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
@ ds: 由 dds_client_init 返回的 struct dds_client 实例指针；
|
|||
|
|
@ ev: 发送的消息事件；
|
|||
|
|
@ data: 此消息附带的数据, json 格式；
|
|||
|
|
|
|||
|
|
|
|||
|
|
返回值:
|
|||
|
|
|
|||
|
|
只有当 sdk 完成初始化并且授权成功之后才能正确返回，否则返回 -1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
返回 nativeAPI 查询结果的接口
|
|||
|
|
|
|||
|
|
int dds_client_resp_nativeapi(struct dds_client *ds,
|
|||
|
|
const char *native_api, const char *native_api_data_json);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
ds: sdk 实例指针；
|
|||
|
|
|
|||
|
|
native_api: 这个 nativeAPI topic 名；
|
|||
|
|
|
|||
|
|
native_api_data_json: 查询数据结果，json string，格式如下，
|
|||
|
|
"duiWidget" 字段必须包含， 且目前取值为 "text"。 用户自定义的数据必须放在
|
|||
|
|
extra 字段中。
|
|||
|
|
|
|||
|
|
{
|
|||
|
|
"duiWidget": "text",
|
|||
|
|
"extra": {
|
|||
|
|
"xx": "11"
|
|||
|
|
}
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
返回值: 如果 sdk 没有初始化完成或者授权成功则返回-1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
内部合成的接口:
|
|||
|
|
|
|||
|
|
int dds_client_speak(struct dds_client *ds, const char *text);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
|
|||
|
|
text: 需要合成的文本
|
|||
|
|
|
|||
|
|
返回值: 如果 sdk 没有初始化完成或者授权成功则返回-1
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
外部 feed 音频接口:
|
|||
|
|
|
|||
|
|
int dds_client_feed_audio(struct dds_client *ds, char *data, int len);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
|
|||
|
|
data: 录音机数据
|
|||
|
|
|
|||
|
|
len: 数据长度
|
|||
|
|
|
|||
|
|
返回值： 出错返回 -1 ，此接口只有在 recorder 配置为外部方式才会生效。
|
|||
|
|
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
停止当前对话，包括停止合成，取消识别等。
|
|||
|
|
|
|||
|
|
int dds_client_stop_dialog(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
|
|||
|
|
text: 需要合成的文本
|
|||
|
|
|
|||
|
|
返回值: 如果 sdk 没有初始化完成或者授权成功则返回-1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
关闭唤醒，如果在语音对话过程中调用此接口，会在这条对话自然结束之后才会禁止唤醒。
|
|||
|
|
|
|||
|
|
int dds_client_disable_wakeup(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
|
|||
|
|
返回值: 如果 sdk 没有初始化完成或者授权成功则返回-1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
打开唤醒
|
|||
|
|
|
|||
|
|
int dds_client_enable_wakeup(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
|
|||
|
|
返回值: 如果 sdk 没有初始化完成或者授权成功则返回-1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
设置用户唤醒词
|
|||
|
|
|
|||
|
|
int dds_client_update_customword(struct dds_client *ds,
|
|||
|
|
const char *word);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
word: 唤醒词配置，格式是 json string，说明如下:
|
|||
|
|
|
|||
|
|
{
|
|||
|
|
"greetingFile":"path:../res/tts/help.mp3", 可选
|
|||
|
|
"greeting": "我在，有什么可以帮你", 可选
|
|||
|
|
"pinyin": "ni hao xiao chi", 必选
|
|||
|
|
"name": "你好小驰", 必选
|
|||
|
|
"threshold": 0.127 必选
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
此函数成功返回后，唤醒词的相关配置会更新到 config.json 文件。
|
|||
|
|
对于客户端异常断电可能导致 config.json 文件破坏的话， 需要开发者自己来避免，
|
|||
|
|
比如采用备份文件的机制。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
获取当前的唤醒词
|
|||
|
|
|
|||
|
|
char* dds_client_get_wakeupwords(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
参数说明:
|
|||
|
|
ds: sdk 实例指针
|
|||
|
|
|
|||
|
|
此函数返回字符串指针， 开发者需要主动释放内存。 返回字符串为json格式，如下:
|
|||
|
|
|
|||
|
|
{
|
|||
|
|
"majorword": [{
|
|||
|
|
"greetingFile": "path:../res/tts/help.mp3",
|
|||
|
|
"greeting": "我在，有什么可以帮你",
|
|||
|
|
"pinyin": "ni hao xiao le",
|
|||
|
|
"name": "你好小乐",
|
|||
|
|
"threshold": 0.144000
|
|||
|
|
}],
|
|||
|
|
"minorword": [{
|
|||
|
|
"greetingFile": "path:../res/tts/help.mp3",
|
|||
|
|
"greeting": "我在，有什么可以帮你",
|
|||
|
|
"pinyin": "ni hao xiao chi",
|
|||
|
|
"name": "你好小驰",
|
|||
|
|
"threshold": 0.127000
|
|||
|
|
}],
|
|||
|
|
"cmdword": [{
|
|||
|
|
"pinyin": "jiang di yin liang",
|
|||
|
|
"threshold": 0.100000,
|
|||
|
|
"action": "decrease.volume",
|
|||
|
|
"name": "降低音量"
|
|||
|
|
}],
|
|||
|
|
"customword": [{
|
|||
|
|
"pinyin": "ni hao tiam mao",
|
|||
|
|
"name": "你好天猫",
|
|||
|
|
"threshold": 0.200000
|
|||
|
|
}]
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
majorword 为主唤醒词，minorword 为副唤醒词， cmdword 为命令词，
|
|||
|
|
customword 为用户定义唤醒词。 其实就是 config.json 文件里面的配置。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
// 获取当前的 tts 发音人，出错返回 NULL
|
|||
|
|
char *dds_client_get_speaker(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
// 获取当前的 tts 播报速度，为 float 型， 在 0 ~ 1 之间，越大表示速度越慢。
|
|||
|
|
float dds_clent_get_speed(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
// 获取当前的 tts 的播报音量大小， 为 int 型， 在 0 ~ 100 之间。
|
|||
|
|
int dds_client_get_volume(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
// 设置当前的 tts 的播报音色人，出错返回 -1
|
|||
|
|
int dds_client_set_speaker(struct dds_client *ds, char *speaker);
|
|||
|
|
|
|||
|
|
// 设置当前的 tts 的播报速度大小，出错返回 -1
|
|||
|
|
int dds_client_set_speed(struct dds_client *ds, float speed);
|
|||
|
|
|
|||
|
|
// 设置当前的 tts 的播报音量大小，出错返回 -1
|
|||
|
|
int dds_client_set_volume(struct dds_client *ds, int vol);
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
声纹的相关接口
|
|||
|
|
|
|||
|
|
// 获取当前的声纹详细信息
|
|||
|
|
|
|||
|
|
char *dds_client_vprint_get_detail(struct dds_client *ds);
|
|||
|
|
|
|||
|
|
返回的是一个 json 格式字符串，格式为:
|
|||
|
|
|
|||
|
|
{
|
|||
|
|
"vNum": 2,
|
|||
|
|
"detail": [
|
|||
|
|
{
|
|||
|
|
"name":"test1"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"name":"test2"
|
|||
|
|
}
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
|
|||
|
|
// 开始进入声纹注册的接口
|
|||
|
|
|
|||
|
|
int dds_client_vprint_regist(struct dds_client *ds, char *name);
|
|||
|
|
|
|||
|
|
name 参数表示声纹注册人的姓名标识。
|
|||
|
|
|
|||
|
|
出错返回-1，通常是格式错误。而声纹注册过程中的详细错误将会通过回调抛出。
|
|||
|
|
|
|||
|
|
|
|||
|
|
// 删除注册人信息
|
|||
|
|
|
|||
|
|
int dds_client_vprint_unregist(struct dds_client *ds, char *name);
|
|||
|
|
|
|||
|
|
出错返回-1，通常是格式错误。而声纹删除过程中的详细错误将会通过回调抛出。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
能量接口
|
|||
|
|
|
|||
|
|
int dds_client_energy_estimate(struct dds_client* ds, int second);
|
|||
|
|
|
|||
|
|
second 参数表示接下来所要计算的音频时间长度，单位为秒。计算结果通过回调给出。
|
|||
|
|
|
|||
|
|
出错返回 -1
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
客户端信息上传的相关接口
|
|||
|
|
|
|||
|
|
int dds_client_upload_city(struct dds_client *dc, char *city);
|
|||
|
|
|
|||
|
|
city 为当前设备所在的城市名，比如苏州，上海等
|
|||
|
|
|
|||
|
|
出错返回 -1
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
**sdk回调消息接口**
|
|||
|
|
|
|||
|
|
<table>
|
|||
|
|
<tr>
|
|||
|
|
<th>回调函数</th>
|
|||
|
|
<th>消息</th>
|
|||
|
|
<th>含义</th>
|
|||
|
|
<th>参数</th>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> local_wakeup.result </td>
|
|||
|
|
<td> 唤醒事件 </td>
|
|||
|
|
<td> json string，形如
|
|||
|
|
{"type":"major","greeting":"好的",
|
|||
|
|
"word":"你好小驰"} </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> doa.result </td>
|
|||
|
|
<td> doa事件 </td>
|
|||
|
|
<td> json string, 形如 {"dao": 100} </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.vad.begin </td>
|
|||
|
|
<td> vad开始的事件 </td>
|
|||
|
|
<td> 无 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.vad.end </td>
|
|||
|
|
<td> vad结束的事件 </td>
|
|||
|
|
<td> 无 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.tts.begin </td>
|
|||
|
|
<td> 合成音开始的事件 </td>
|
|||
|
|
<td> 无 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.tts.end </td>
|
|||
|
|
<td> 合成音结束的事件，播放结束 </td>
|
|||
|
|
<td> 无 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.asr.begin </td>
|
|||
|
|
<td> sdk内部开始做识别 </td>
|
|||
|
|
<td> 无 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> asr.speech.text </td>
|
|||
|
|
<td> 实时的语音识别结果反馈 </td>
|
|||
|
|
<td> json string </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> asr.speech.result </td>
|
|||
|
|
<td> 最终的语音识别结果反馈 </td>
|
|||
|
|
<td> json string </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> dm.output </td>
|
|||
|
|
<td> 对话的输出结果 </td>
|
|||
|
|
<td> json string </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.dm.end </td>
|
|||
|
|
<td> 表示结束对话 </td>
|
|||
|
|
<td> 无 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> device.mode.return </td>
|
|||
|
|
<td> 表示设置设备状态的回复消息 </td>
|
|||
|
|
<td> json string "{"result":"success"}"</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> sys.client.error </td>
|
|||
|
|
<td> 表示客户端出现异常情况 </td>
|
|||
|
|
<td> json string "{"error":"ttsError"}" 目前 error 字段的取值一共有: ttsError, ddsNetworkError, vadSlienceTimeout</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> command://xx </td>
|
|||
|
|
<td> 在dui平台上配置的 command 指令 </td>
|
|||
|
|
<td> json string </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> native://xx </td>
|
|||
|
|
<td> 在dui平台上配置的 native 指令 </td>
|
|||
|
|
<td> json string </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> vprint.regist.result </td>
|
|||
|
|
<td> 声纹注册的消息回调 </td>
|
|||
|
|
<td> json string 形如, {"operation":"start"} 表示声纹注册接口调用成成功，开始注册声纹。 {"operation":"nameRepeat"} 表示注册人姓名重复， {"operation":"vNumLimit"} 表示超出声纹注册上限， {"operation":"unavailable"} 表示所注册音频信噪比不够， {"operation":"continue"} 表示可以继续注册声纹，{"operation":"success"}表示声纹注册成功</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> vprint.unregist.result </td>
|
|||
|
|
<td> 声纹删除的消息回调 </td>
|
|||
|
|
<td> json string 形如 {"operation":"success"} 表示删除成功, {"operation":"noSpeaker"} 表示没有所有删除的注册人信息</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> vprint.test.result </td>
|
|||
|
|
<td> 声纹计算的消息回调 </td>
|
|||
|
|
<td> json string 形如: {"register":"nothing"} 表示当前还没有声纹模型， {"score":23.286682,"word":"qi ke kong tiao","register":"test0","time":179.679932,"speech":0.880000,"RTF":0.204188} 表示收到了正常的计算结果，其中的 register 字段表示所计算的声纹标识。 如果 register 为 others表示没有匹配到具体的声纹</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> ddsLintener </td>
|
|||
|
|
<td> energy.estimate.result </td>
|
|||
|
|
<td> 能量计算的消息回调 </td>
|
|||
|
|
<td> json string 形如, {"value":40} </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
|
|||
|
|
</table>
|
|||
|
|
|
|||
|
|
### 配置选项
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
{
|
|||
|
|
"sdk": {
|
|||
|
|
"configPath":"./config.json"
|
|||
|
|
},
|
|||
|
|
"auth": {
|
|||
|
|
"productId": "278569448",
|
|||
|
|
"deviceProfile": ""
|
|||
|
|
},
|
|||
|
|
"front": {
|
|||
|
|
"aecBinPath": "",
|
|||
|
|
"wakeupBinPath": "",
|
|||
|
|
"beamformingBinPath": "",
|
|||
|
|
"rollBack": 0
|
|||
|
|
},
|
|||
|
|
"vad": {
|
|||
|
|
"resBinPath": "",
|
|||
|
|
"pauseTime": 500,
|
|||
|
|
"slienceTimeout": 5
|
|||
|
|
},
|
|||
|
|
"cloud": {
|
|||
|
|
"productId": "278569448",
|
|||
|
|
"aliasKey": "prod"
|
|||
|
|
},
|
|||
|
|
"recorder": {
|
|||
|
|
"mode": "internal"
|
|||
|
|
"samplerate":16000,
|
|||
|
|
"bits":16,
|
|||
|
|
"channels":1,
|
|||
|
|
"device": "default"
|
|||
|
|
},
|
|||
|
|
"player": {
|
|||
|
|
"device": "default"
|
|||
|
|
},
|
|||
|
|
"tts": {
|
|||
|
|
"type": "cloud",
|
|||
|
|
"zhilingf": {
|
|||
|
|
"resBinPath":"",
|
|||
|
|
"dictPath":""
|
|||
|
|
},
|
|||
|
|
"voice": "zhilingf",
|
|||
|
|
"volume": 50,
|
|||
|
|
"speed": 0.85
|
|||
|
|
},
|
|||
|
|
"oneShot": {
|
|||
|
|
"enable": false
|
|||
|
|
},
|
|||
|
|
"abnormal": {
|
|||
|
|
"netErrorHint":"path:../res/tts/net.mp3",
|
|||
|
|
"ttsErrorHint":"path:../res/tts/tts_error.mp3"
|
|||
|
|
},
|
|||
|
|
"debug": {
|
|||
|
|
"recAudioDumpFile":"",
|
|||
|
|
"bfAudioDumpFile":""
|
|||
|
|
}
|
|||
|
|
}
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
<table>
|
|||
|
|
<tr>
|
|||
|
|
<th>参数</th>
|
|||
|
|
<th>类型</th>
|
|||
|
|
<th>含义</th>
|
|||
|
|
<th>是否必须</th>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> sdk </td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td> 客户端的一些配置 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> sdk.configPath </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td> 配置文件路径 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> auth </td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td> 授权信息 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> auth.productId </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td> dui 上创建产品ID </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> auth.deviceProfile </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td>授权信息</td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> front </td>
|
|||
|
|
<td>json 对象 </td>
|
|||
|
|
<td> 前端信号处理的相关配置 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> front.aecBinPath </td>
|
|||
|
|
<td> string</td>
|
|||
|
|
<td>aec 算法资源路径</td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> front.wakeupBinPath </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td>唤醒算法资源路径</td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> front.beamformingBinPath </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td>beamforming 算法资源路径</td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> vad </td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td>vad 算法模块配置 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> vad.resBinPath </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td>vad算法资源路径 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> vad.pauseTime </td>
|
|||
|
|
<td>int</td>
|
|||
|
|
<td>vad截止检测时长，单位为 ms，默认为 500ms </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> vad.slienceTimeout </td>
|
|||
|
|
<td>int</td>
|
|||
|
|
<td> vad 的静音检测超时时长，单位为s，默认为 6s </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> cloud </td>
|
|||
|
|
<td>json 对象 </td>
|
|||
|
|
<td>云端产品的相关配置 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td>cloud.productId </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td>dui平台上创建的产品ID</td>
|
|||
|
|
<td>必选 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> cloud.aliasKey </td>
|
|||
|
|
<td>string</td>
|
|||
|
|
<td> dui平台上创建的产品ID 发布支持， 取值为 "prod | test"</td>
|
|||
|
|
<td>可选，默认为 prod 分支</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td>recorder </td>
|
|||
|
|
<td>json 对象 </td>
|
|||
|
|
<td>录音机的相关配置 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> recorder.mode </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td> 录音方式，取值为 "internal | external" 分列表示内部录音和外部录音。 </td>
|
|||
|
|
<td> 可选 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> recorder.samplerate </td>
|
|||
|
|
<td> int </td>
|
|||
|
|
<td> 录音采样率 </td>
|
|||
|
|
<td> 必选 </td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> recorder.bits </td>
|
|||
|
|
<td>int</td>
|
|||
|
|
<td>录音采样位数</td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> recorder.channels </td>
|
|||
|
|
<td>int</td>
|
|||
|
|
<td>录音采用通道数</td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> recorder.device </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td>内部录音机的设备名，默认当前系统的 default 音频设备 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> player</td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td>内部播放器的设置 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> player.device </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td>内部播放器的设备名，默认为 default </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts </td>
|
|||
|
|
<td>json 对象</td>
|
|||
|
|
<td>合成音的相关配置</td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.type </td>
|
|||
|
|
<td>string </td>
|
|||
|
|
<td>合成音的类型，支持 "cloud" | "local" 分别表示云端合成和本地合成</td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.voice </td>
|
|||
|
|
<td>string </td>
|
|||
|
|
<td>合成音的音色，如果为本地合成， 仅支持 "zhilingf" </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.zhilingf </td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td>当 tts.type 为 "local"时，会根据 tts.voice 选择对应音色的合成资源路径。</td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.zhilingf.resBinPath </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td>本地合成 zhilingf 的资源路径</td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.zhilingf.dictPath </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td>本地合成 zhilingf 的词典路径</td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.volume </td>
|
|||
|
|
<td> int </td>
|
|||
|
|
<td>合成音的音量 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> tts.speed </td>
|
|||
|
|
<td> int </td>
|
|||
|
|
<td>合成音的速度 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> oneShot </td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td> oneshot 模块配置 </td>
|
|||
|
|
<td>必选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> oneShot.enable </td>
|
|||
|
|
<td> bool </td>
|
|||
|
|
<td> 是否启用oneshot，当前仅支持 false </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> abnormal </td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td> sdk异常情况下对话配置 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> abnormal.netErrorHint </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td> 网络错误下的提示音，需要配置成本地文件，网络不好的情况下云端合成也用不了。 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> abnormal.ttsErrorHint </td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td> 云端tts合成播放错误情况下的的提示音，需要配置成本地文件。 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
<tr>
|
|||
|
|
<td> debug</td>
|
|||
|
|
<td> json 对象 </td>
|
|||
|
|
<td> 保存音频的配置选项 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
<tr>
|
|||
|
|
<td> debug.recAudioDumpFile</td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td> 原始录音保存文件路径 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
<tr>
|
|||
|
|
<td> debug.bfAudioDumpFile</td>
|
|||
|
|
<td> string </td>
|
|||
|
|
<td> beamforming算法输出的音频文件路径 </td>
|
|||
|
|
<td>可选</td>
|
|||
|
|
</tr>
|
|||
|
|
|
|||
|
|
</table>
|
|||
|
|
|
|||
|
|
|
|||
|
|
**唤醒词说明**
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
{
|
|||
|
|
"greetingFile":"path:./res/tts/help.mp3",
|
|||
|
|
"greeting": "我在，有什么可以帮你",
|
|||
|
|
"pinyin": "ni hao xiao chi",
|
|||
|
|
"name": "你好小驰",
|
|||
|
|
"threshold": 0.127
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
|
|||
|
|
greetingFile: 唤醒之后播放的提示音，支持本地录音文件，传入文件路径。
|
|||
|
|
greeting: 唤醒之后的提示文本， sdk内部合成。
|
|||
|
|
pinyin: 唤醒词拼音。
|
|||
|
|
name: 唤醒词的中文。
|
|||
|
|
threshold: 唤醒词阈值。
|
|||
|
|
|
|||
|
|
唤醒提示音播放优先级: 如果配置了 greetingFile 则播放 greetingFile ，
|
|||
|
|
否则播放 greeting 。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
|
|||
|
|
**命令唤醒词说明**
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
{
|
|||
|
|
"pinyin": "jiang di yin liang",
|
|||
|
|
"threshold": 0.100,
|
|||
|
|
"action": "decrease.volume",
|
|||
|
|
"name": "降低音量"
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
pinyin: 唤醒词拼音。
|
|||
|
|
name: 唤醒词的中文。
|
|||
|
|
threshold: 唤醒词阈值。
|
|||
|
|
action: 该命令唤醒词对应的动作，比如这个例子中，sdk回调函数会抛出
|
|||
|
|
command://decrease.volume 消息。
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
|
|||
|
|
**用户事件说明**
|
|||
|
|
|
|||
|
|
用户可以通过dds_client_publish(struct dds_client *ds, int ev, const char *data); 接口给sdk发送事件。
|
|||
|
|
|
|||
|
|
`DDS_CLIENT_USER_EXTERNAL_WAKEUP` 表示外部唤醒事件，用户可以在收到声纹计算结果之后通过这个事件让sdk继续交互。
|
|||
|
|
|
|||
|
|
这个事件传入的数据格式为：
|
|||
|
|
|
|||
|
|
{
|
|||
|
|
"nlg":"xxxx"
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
可以通过nlg字段传入所有播报的内容，播放完成之后就可以做识别了。 如果传入NULL表示立刻开始识别。
|
|||
|
|
|